图片展示

利用本构方程和加工图研究了Al-Cu-Li-Zn合金的热变形特征及其显微组织演变

作者:DSI上海办事处 浏览: 发表时间:2023-10-12 14:43:48

The hot deformation characteristics and the associated microstructural evolution of an Al–Cu–Li–Zn alloy studied by constitutive equations and processing maps

Hua Wang,Dengfeng Yin,Ming-Chun Zhao,Yan Tian,Andrej Atrens

The hot deformation behavior and the associated microstructural evolution (which were the keys to decide mechanical properties) were studied using different deformation conditions for the quaternary Al–Cu–Li–Zn alloy containing 1.0 wt% Zn. The constitutive equations and processing map were established. Dynamic recrystallization (DRX) was related to the orientation of the initial grains. The DRX rarely occurred in the grains with an orientation near<110>Al and mainly occurred in the grains with an orientation near<100>Al. Particle simulated nucleation (PSN) promoted by pre-existed coarse T1 phase particles was the main mechanism of the DRX when the deformation temperature was below 450 °C. In the temperature range of 450–500 °C with a strain rate of 0.01 s−1, the DRX occurred by grain boundary bulging or the increase in the cumulative misorientation, but only small fraction of the DRX grains appeared. This work provided an important guideline for the optimization of deformation techniques and microstructures.

    Keywords

    Al–Cu–Li–Zn alloy
    Processing map
    Constitutive equation
    DRX
    Microstructural evolution
    Hot deformation

    1. Introduction

    Al–Cu–Li alloys are third generation Al–Li alloys, and have a significant advantage in high stiffness to density ratio and large elastic modulus compared with conventional 2xxx and 7xxx series Al alloys. These good properties result from alloying with Li [1,2]. The Al–Cu–Li alloys haves gained extensive attention due to their increased strength-toughness balance, corrosion resistance, cold working ability, etc. Compared to the previous two generations of Al–Li alloys [1,2]. As the commercialized latest generation Al–Li alloys, Al–Cu–Li alloys have a lower Li content and contain a variety of microalloying elements (Mg, Mn, Zr and Zn, etc.). The high strength of the Al–Cu–Li alloys is attributed to the combined precipitation of the main precipitates, including T1 (Al2CuLi), θ' (Al2Cu) and δ′(Al3Li). The type and number density of the strengthening phase particles are significantly influenced by the Cu/Li ratio [3,4]. T1 is considered as the most effective phase for improving strength, whose precipitation requires a higher Cu/Li ratio [5]. Alloying Al with minor microalloying elements, particularly Zn, is an effective method to obtain prominent precipitation hardening. Zhang et al. suggested that Zn alloying promoted the precipitation of the strengthening phases in Al–Cu–Li alloys [6]. Increasing the Zn content to ∼1 wt% greatly improved the strength of the Al–Cu–Li alloy with a high Cu/Li ratio, which was attributed to the enhanced precipitation of the T1 phase [7,8]. Key compositional design principles may promote successful application and commercialization of Al–Cu–Li alloys.

    The hot deformation behavior and the associated microstructural evolution during hot deformation determine the mechanical properties [9,10]. The deformation behavior varies with deformation parameters, such as strain rate, temperature and strain [11,12]. An appropriate constitutive equation is required to establish the relationship between flow stress, strain rate and deformation temperature. This equation can be used to predict the flow behavior. The associated microstructural evolution including dynamic recovery (DRV), dynamic recrystallization (DRX) and dynamic precipitation is also influenced by the deformation conditions [13,14]. The high stacking fault energy (SFE) of Al alloy facilitates dislocation cross-slip and climb, and causes DRV to be dominant in Al alloys. However, the degree of DRX plays an important role in flow softening [15,16]. Therefore, it is important to investigate the hot deformation characteristics and the associated microstructural evolution. Symmetric isothermal compression experiments have been widely used to study the hot deformation behavior. The dynamic softening mechanism associated with deformation conditions can be understood by microstructural investigation [[16][17][18][19][20]]. A processing map constructed from experimental data and a dynamic material model can be used for optimizing hot deformation, which can predict stable domains and instability domains. A processing map has been used to identify the deformation temperature and strain rate corresponding to different microstructural evolution of stable domains and instability domains [21]. Obviously, a processing map can evaluate workability of the bulk materials, can predict deformation mechanisms in different regions using the power dissipation efficiency η, and can also define the unstable regions that should be avoided during hot deformation. This allows the determination of optimized thermal processing parameters. Investigation of the microstructure corresponding to different regions in the hot processing map allows understanding the deformation mechanism under different conditions. Zhang et al. [19] found that the main DRX mechanism at medium temperatures was discontinuous dynamic recrystallization (DDRX), while that at high temperatures was continuous dynamic recrystallization (CDRX). Ke et al. [20] found that microcracks appeared when the AA7020 Al alloy was deformed at a high strain rate, which corresponded to the low η value region or unsafe region of the hot processing map. Lu et al. [13] found that a microshear band was formed in 2195Al–Li alloy when deformed at a low temperature with a high strain rate (300 °C/10 s−1), which was defined as an unsafe region in the processing map. Despite considerable efforts devoted to studying the hot deformation behavior and the associated microstructural evolution of Al–Li alloys [[22][23][24][25][26][27]], to the best of our knowledge, the study of the hot deformation characteristics and the associated microstructural evolution of Al–Cu–Li–Zn alloys with a high Zn content has been quite limited.

    This work studied the hot deformation characteristics and the associated microstructural evolution of a homogenized Al–Cu–Li–Zn alloy containing 1 wt% Zn using constitutive equations and a processing map based on the true stress-strain data measured using a hot compression test. The microstructure of the specimens under various conditions corresponding to different regions of the processing map was characterized, and the softening mechanism, especially the DRX mechanism, was evaluated based on the microstructural evolution during hot deformation using electron backscatter diffraction (EBSD) and transmission electron microscope (TEM) analysis. This work will provide an important guideline for the optimization of deformation techniques and microstructures.

    2. Materials and experiments

    2.1. Materials and hot compression test

    A laboratory-made Al–Cu–Li–Zn alloy was used in this study. The alloy ingot was produced by melting Al-50 wt% Cu and Al-4.0 wt% Zr master alloys and pure Li, Mg, Ag, Zn and Al in a resistance furnace using a high-purity graphite crucible and casting into a rectangular water-cooled copper mold protected by an argon atmosphere. The chemical composition of the alloy determined by Inductively Coupled Plasma Optical Emission Spectrometry (ICP-OES) is presented in Table 1. As in our previous work [28], the as-cast alloy ingot was subjected to a two-step homogenization (annealing) treatment at 470 °C for 8 h and 510 °C for 16 h in an air furnace and was furnace cooled to room temperature. Several cylindrical specimens for hot compression tests were produced with a size of 10 mm diameter and 15 mm height from the center part of the annealed rectangular ingot. Hot compression tests were carried out at temperatures in the range of 300 °C–500 °C and at strain rates of 0.001–10 s−1 on a Gleeble-3180 isothermal simulator as in previous studies [[29][30][31]]. During hot compression, graphite slices were placed on the end face of the samples to reduce the effect of friction. The temperature was measured using a thermal couple that was linked to the specimen. The test was completed when the height reduction reached 50%. Thereafter the specimen was immediately water-quenched. The specimens used for microstructure observation were obtained by slicing the hot-compressed specimens along the centerline parallel to the compression direction.

    Table 1. Chemical composition of the Al–Cu–Li–Zn alloy (wt.%).

    CuLiMgAgZrZnFeSiAl
    4.971.310.440.410.161.010.110.043Bal.

    2.2. Microstructure characterization

    The microstructure of the specimens was characterized using a scanning electron microscope (SEM) and a TEM. SEM observation used a TESCAN MIRA4 LMH instrument equipped with One Max 50 EDSEBSD maps were employed to analyze the grain structure of the deformed specimens. Samples for SEM and EBSD were first ground using silicon carbide abrasive papers and then mechanical polished using diamond paste. Samples for EBSD testing were further electropolished in a mixture solution of 10% perchloric acid and 90% ethanol at 20V for 5 s to obtain high quality EBSD maps. EBSD was carried out on an Oxford EBSD detector operated at 20 kV with a step size of 3.5 μm. The data were processed using the AztecCrystal software. In the EBSD data analysis, if not specified, low angle grain boundaries (LAGBs, 2–10°) are represented by grey lines and high angle grain boundaries (HAGBs, >15°) are represented by black lines.

    TEM observation used an atomic resolution Thermo Fisher Scientific TM Talos F200X STEM operated at 200 KV. Specimens for TEM observation were prepared by mechanically grinding the samples to a final thickness of 60–80 μm, then 3 mm thin foil discs were punched out for twin-jet electro-polishing in a solution of 25 vol% nitric acid in methanol at 18 V and at ∼ -25-30 °C.

    3. Results

    3.1. True stress-strain curves

    Fig. 1 shows true stress-strain (σ-ε) curves of the Al–Cu–Li–Zn alloy deformed at different strain rates. The study of the deformation behavior of an alloy during hot expression requires consideration of the friction between specimen and anvils [12,20,32]. The barrel shape coefficient (B) was used to evaluate the impact of friction, expressed as [33]:B=hRM2h0R02" role="presentation" style="color: rgb(0, 0, 0); display: inline-block; line-height: normal; font-size: 14.4px; word-spacing: normal; overflow-wrap: normal; text-wrap: nowrap; float: none; direction: ltr; max-width: none; max-height: none; min-width: 0px; min-height: 0px; position: relative;">=2002where h0 and R0 represent the initial height and diameter of the specimen, respectively, and h and RM are the height and maximum diameter of the specimens after thermal compression. Friction has little influence on flow stress if 1 < B ≦ 1.1, and no correction of the flow stress-strain curves is needed. In contrast, B ≥ 1.1 indicates that friction has an impact on the flow stress and the modification of the stress-strain curves is necessary. Evaluation of the barrel shape coefficients of each specimen indicated that B ≦ 1.1. Therefore, it was not necessary to correct the curves.

    Fig. 1

    Fig. 1

    The true stress-strain curves of the specimens for deformation temperatures in the range of 300–500 °C and at strain rates of 0.001–10 s−1 are depicted in Fig. 2. The flow stress varied with temperature and strain rate. Nevertheless, the characteristic of these curves was essentially similar for all these deformation conditions. During the initial stage of the deformation, the flow stress increased rapidly to reach a peak. Then the flow dropped continuously (for example, the curves at low temperatures) or was constant with increasing strain. The occurrence of these behaviors was mainly influenced by temperature. The degree of flow softening was influenced by temperature and strain rate. However, at a strain rate of 1, there was almost no stress peak during the whole deformation temperature range, indicating weak dynamic softening for these conditions. In addition, the σ-ε curves shown in Fig. 1 exhibited serrated flow behavior under most deformation conditions. The serrated flow varied in magnitude for different deformation conditions and has been attributed to partial dynamic recrystallization [29] or precipitation of the second phase [34].

    Fig. 2

    Fig. 2

    The flow softening during thermal compression was quantified using the relative softening amplitude (γ) defined by Ref. [35]:γ=(σpσf)/σp×100%" role="presentation" style="color: rgb(0, 0, 0); display: inline-block; line-height: normal; font-size: 14.4px; word-spacing: normal; overflow-wrap: normal; text-wrap: nowrap; float: none; direction: ltr; max-width: none; max-height: none; min-width: 0px; min-height: 0px; position: relative;">=()/×100%where σp and σf are the peak stress and the stress at final deformation, respectively. The values of γ are plotted in Fig. 2. The values of γ were greater at the low strain rate (0.001 s−1) and at the high strain rate (10 s−1). The γ value decreased as the strain rate increased from 0.001 to 0.1 s−1, indicating that the dynamic softening decreased as the strain rate increased. The value of γ was only slightly larger than zero at 350 °C and was negative at all other temperatures, indicating almost no softening at strain rate of 1 s−1, which is also evident in Fig. 1 (d). At a strain rate of 0.01 s−1, γ first increased to a maximum at 350 °C, then decreased to a minimum at 450 °C followed by an increase at 500 °C. There was a similar trend at the strain rate of 10 s−1.

    3.2. Establishment of constitutive model

    The deformation conditions of temperature and strain rate control the flow behavior of the alloy. The constitutive equations were established based on the experimental data to connect the key variables. They can allow optimization of thermal mechanical processing parameters and the prediction of flow behavior. The following three equations are commonly used to determine the constitutive equation parameters:ε˙=A1σn1exp(QRT)" role="presentation" style="display: inline-block; line-height: normal; font-size: 14.4px; word-spacing: normal; overflow-wrap: normal; text-wrap: nowrap; float: none; direction: ltr; max-width: none; max-height: none; min-width: 0px; min-height: 0px; position: relative;">˙=11exp()ε˙=A2exp(βσ)exp(QRT)" role="presentation" style="display: inline-block; line-height: normal; font-size: 14.4px; word-spacing: normal; overflow-wrap: normal; text-wrap: nowrap; float: none; direction: ltr; max-width: none; max-height: none; min-width: 0px; min-height: 0px; position: relative;">˙=2exp()exp()ε˙=A(sinh(ασ))nexp(QRT)" role="presentation" style="display: inline-block; line-height: normal; font-size: 14.4px; word-spacing: normal; overflow-wrap: normal; text-wrap: nowrap; float: none; direction: ltr; max-width: none; max-height: none; min-width: 0px; min-height: 0px; position: relative;">˙=(sinh())exp()where ε˙" role="presentation" style="display: inline-block; line-height: normal; font-size: 14.4px; word-spacing: normal; overflow-wrap: normal; text-wrap: nowrap; float: none; direction: ltr; max-width: none; max-height: none; min-width: 0px; min-height: 0px; position: relative;">˙ is the strain rate (s−1); A1, A2, A, n, n1 and β are material constants, Q is the deformation activation energy (J/mol); T is the temperature (K); R is the gas constant (8.314 Jmol−1K1); α is stress multiplier. The relationship between ˙, σ and T at a low stress level, a high stress level and all stress levels can be described by Eqs. (3)(4)(5), respectively. Taking natural logarithm on both sides of Eqs. (3)(4)(5) transforms these expressions as follows:lnε˙=lnA1+n1lnσQRT" role="presentation" style="display: inline-block; line-height: normal; font-size: 14.4px; word-spacing: normal; overflow-wrap: normal; text-wrap: nowrap; float: none; direction: ltr; max-width: none; max-height: none; min-width: 0px; min-height: 0px; position: relative;">ln˙=ln1+1lnlnε˙=lnA2+βσQRT" role="presentation" style="display: inline-block; line-height: normal; font-size: 14.4px; word-spacing: normal; overflow-wrap: normal; text-wrap: nowrap; float: none; direction: ltr; max-width: none; max-height: none; min-width: 0px; min-height: 0px; position: relative;">ln˙=ln2+lnε˙=lnA+nln(sinh(ασ))QRT" role="presentation" style="display: inline-block; line-height: normal; font-size: 14.4px; word-spacing: normal; overflow-wrap: normal; text-wrap: nowrap; float: none; direction: ltr; max-width: none; max-height: none; min-width: 0px; min-height: 0px; position: relative;">ln˙=ln+ln(sinh())

    The relationships of lnε˙lnσ" role="presentation" style="display: inline-block; line-height: normal; font-size: 14.4px; word-spacing: normal; overflow-wrap: normal; text-wrap: nowrap; float: none; direction: ltr; max-width: none; max-height: none; min-width: 0px; min-height: 0px; position: relative;">ln˙lnlnε˙σ" role="presentation" style="display: inline-block; line-height: normal; font-size: 14.4px; word-spacing: normal; overflow-wrap: normal; text-wrap: nowrap; float: none; direction: ltr; max-width: none; max-height: none; min-width: 0px; min-height: 0px; position: relative;">ln˙ and lnε˙ln[sinh(ασ)]" role="presentation" style="display: inline-block; line-height: normal; font-size: 14.4px; word-spacing: normal; overflow-wrap: normal; text-wrap: nowrap; float: none; direction: ltr; max-width: none; max-height: none; min-width: 0px; min-height: 0px; position: relative;">ln˙ln[sinh()] can be easily obtained from the above equations and they were plotted and linear fitted in Fig. 3. In these equations σ is the peak stress. The parameters of n1, β and n can be evaluated from [lnε˙lnσ]T" role="presentation" style="color: rgb(0, 0, 0); display: inline-block; line-height: normal; font-size: 14.4px; word-spacing: normal; overflow-wrap: normal; text-wrap: nowrap; float: none; direction: ltr; max-width: none; max-height: none; min-width: 0px; min-height: 0px; position: relative;">[ln˙ln][lnε˙σ]T" role="presentation" style="color: rgb(0, 0, 0); display: inline-block; line-height: normal; font-size: 14.4px; word-spacing: normal; overflow-wrap: normal; text-wrap: nowrap; float: none; direction: ltr; max-width: none; max-height: none; min-width: 0px; min-height: 0px; position: relative;">[ln˙] and [lnε˙[sinh(ασ)]]T" role="presentation" style="color: rgb(0, 0, 0); display: inline-block; line-height: normal; font-size: 14.4px; word-spacing: normal; overflow-wrap: normal; text-wrap: nowrap; float: none; direction: ltr; max-width: none; max-height: none; min-width: 0px; min-height: 0px; position: relative;">[ln˙[sinh()]] at constant temperature. The calculated values were 7.1250, 0.1080 MPa−1 and 5.0453, respectively. α was determined to be 0.0152 using α = β/n1. The activation energy Q was determined from the following expression:Q=Rnln[sinh(ασ)](1/T)" role="presentation" style="color: rgb(0, 0, 0); display: inline-block; line-height: normal; font-size: 14.4px; word-spacing: normal; overflow-wrap: normal; text-wrap: nowrap; float: none; direction: ltr; max-width: none; max-height: none; min-width: 0px; min-height: 0px; position: relative;">=ln[sinh()](1/)ln[sinh(ασ)](1/T)" role="presentation" style="color: rgb(0, 0, 0); display: inline-block; line-height: normal; font-size: 14.4px; word-spacing: normal; overflow-wrap: normal; text-wrap: nowrap; float: none; direction: ltr; max-width: none; max-height: none; min-width: 0px; min-height: 0px; position: relative;">ln[sinh()](1/) was obtained from the average slope of the ln[sinh(ασ)]-1000/T plots in Fig. 3(d). The value of Q for the present alloy under these hot deformation conditions was 167 kJ/mol, which is only slightly higher than that of pure Al (142 kJ/mol) and lower than for other Al–Cu–Li alloys, such as 2195 alloy (253 kJ/mol) [36], 2196 alloy (181 kJ/mol) [37], 2099 alloy (203 kJ/mol) [38] and 2060 alloy (205 kJ/mol) [32].

    Fig. 3
    1. Download : Download high-res image (481KB)
    2. Download : Download full-size image
    Fig. 3

    Comparison with the previous studies indicated that the activation energy not only depends on the hot deformation parameters such as deformation temperature, strain rate and strain [37,39], but also depends on the heat treatment process, initial microstructure, alloy composition and the method of alloy preparation [19,26,35,36,40]. The deformation activation energy is an important parameter, which reflects the ease of deformation. The physical meaning is that a lower deformation activity energy means easier hot deformation.

    Based on these experiments, the constitutive model for the Al–Cu–Li–Zn alloy can be expressed as:ε˙=3.182×1011[sinh(0.0152σ)]4.568exp(167.16RT)" role="presentation" style="display: inline-block; line-height: normal; font-size: 14.4px; word-spacing: normal; overflow-wrap: normal; text-wrap: nowrap; float: none; direction: ltr; max-width: none; max-height: none; min-width: 0px; min-height: 0px; position: relative;">˙=3.182×1011[sinh(0.0152)]4.568exp(167.16)

    3.3. Hot processing map

    The hot processing map (i) can reflect the hot deformation mechanism, (ii) is characteristic of microstructure evolution with variation of the deformation parameters and (iii) delineates the unstable region and safe region for thermal processing, which provide an important guiding role in optimizing the hot working process. The hot processing map based on a dynamic material model (DMM) have been widely used to determine the optimal hot deformation process parameters for Al alloys, Mg alloys and other materials [[41][42][43]]. This approach regards the material as an energy dissipator. The total energy (P) dissipated during hot working consists of two parts: heat generation and microstructural change such as dynamic recovery, dynamic recrystallization and dynamic precipitation. The total energy can be presented as follows [44]:P=0˙¯εσ¯dε¯˙+0¯σε¯˙dσ¯=G+J" role="presentation" style="color: rgb(0, 0, 0); display: inline-block; line-height: normal; font-size: 14.4px; word-spacing: normal; overflow-wrap: normal; text-wrap: nowrap; float: none; direction: ltr; max-width: none; max-height: none; min-width: 0px; min-height: 0px; position: relative;">=0˙¯ε¯¯˙+0¯σ¯˙¯=+where σ¯" role="presentation" style="color: rgb(0, 0, 0); display: inline-block; line-height: normal; font-size: 14.4px; word-spacing: normal; overflow-wrap: normal; text-wrap: nowrap; float: none; direction: ltr; max-width: none; max-height: none; min-width: 0px; min-height: 0px; position: relative;">¯ is the effective stress and ε¯˙" role="presentation" style="color: rgb(0, 0, 0); display: inline-block; line-height: normal; font-size: 14.4px; word-spacing: normal; overflow-wrap: normal; text-wrap: nowrap; float: none; direction: ltr; max-width: none; max-height: none; min-width: 0px; min-height: 0px; position: relative;">¯˙ is the effective strain rate. The integral parts of Eq. (11) are designated G and J. A key parameter, the strain rate sensitivity m, represents the relative partitioning of power between heat generation and microstructural change, and is given by:m=dJ/dG=(logσ)/(logε˙)" role="presentation" style="display: inline-block; line-height: normal; font-size: 14.4px; word-spacing: normal; overflow-wrap: normal; text-wrap: nowrap; float: none; direction: ltr; max-width: none; max-height: none; min-width: 0px; min-height: 0px; position: relative;">=/=(log)/(log˙)

    The value of m can be calculated by fitting logσlogε˙" role="presentation" style="display: inline-block; line-height: normal; font-size: 14.4px; word-spacing: normal; overflow-wrap: normal; text-wrap: nowrap; float: none; direction: ltr; max-width: none; max-height: none; min-width: 0px; min-height: 0px; position: relative;">loglog˙ with a cubic spline function.

    To represent the proportion of energy dissipated for microstructure evolution (J), the parameter of η was defined as the power dissipation efficiency and was evaluated from the following equation:η=2mm+1" role="presentation" style="display: inline-block; line-height: normal; font-size: 14.4px; word-spacing: normal; overflow-wrap: normal; text-wrap: nowrap; float: none; direction: ltr; max-width: none; max-height: none; min-width: 0px; min-height: 0px; position: relative;">=2+1

    The calculated m values lead to the corresponding η values determined by Eq. (13).

    Fig. 4 shows the 3D power dissipation iso-efficiency contour maps for strain rates 0.001–10 s−1 and the temperature range 300–500 °C at the strain of 0.2, 0.4 and 0.6. The efficiency of power dissipation increases as the strain increases. For a given strain, η increases with increasing temperature and decreasing strain rate. A maximum efficiency of dissipation occurs at 500 °C and 0.01 s−1 and reaches 41% at the strain of 0.6. In general, materials deformed in the region with high η values have better workability [36]. However, it is not necessarily safe for hot working in the high power dissipation efficiency regions as damage mechanism like cracking may also occur in these regions [44]. Therefore, it is not sufficient to determine the conditions (e.g., strain rate and deformation temperature) suitable for thermal processing based on the maximum power dissipation principle from a microstructural perspective. An additional instability map is also necessary. An instability criterion, based on the principle of maximum entropy production, is used to identify the domain of flow instability, and is expressed as:ξ(ε˙)=lg[m/(m+1)]lgε˙+m0" role="presentation" style="display: inline-block; line-height: normal; font-size: 14.4px; word-spacing: normal; overflow-wrap: normal; text-wrap: nowrap; float: none; direction: ltr; max-width: none; max-height: none; min-width: 0px; min-height: 0px; position: relative;">(˙)=lg[/(+1)]lg˙+0

    Fig. 4
    1. Download : Download high-res image (438KB)
    2. Download : Download full-size image
    Fig. 4

    This criterion can be used to establish the instability map. The processing map is obtained by superposing the power dissipation map and the instability map. This can clearly identify the unsafe region (shaded area). Fig. 5 shows the processing maps of the studied alloy as contour maps, which correspond to strains of 0.2, 0.4 and 0.6. The number on the contour lines represent the power dissipation efficiency and the shaded domain represents flow instability such as flow localization or adiabatic shear bands [25]. Fig. 5 indicates that the instability region varies with the strain magnitude, but is mainly distributed in the high strain rate interval and low temperature interval of the hot processing map. Processing in these regions should be avoided in practical applications.

    Fig. 5
    1. Download : Download high-res image (369KB)
    2. Download : Download full-size image
    Fig. 5

    The processing map provides information of “safe” and flow instability domains for process design. In actual production, these domains correspond to different microstructural mechanisms. The microstructure of several hot compressed samples deformed at various conditions that correspond to different areas in the processing map (as shown in Fig. 5(c)) were chosen to study the related microstructural mechanisms.

    3.4. Microstructure evolution during hot compression

    Fig. 6 shows back scattered electron (BSE) micrograph and element mapping of the homogenized alloy. The coarse intermetallic compounds were distributed along the grain boundaries. In addition, there were many plate-like precipitates with specific orientation and some short rod-like phases in the matrix, which were formed during the slow cooling. Energy dispersive X-ray spectroscopy (EDS) mapping showed that the plate-like precipitates were mainly Cu-rich and the short rod like particles were mainly (Cu, Mg)-rich. The coarse constituent particles along grain boundaries were the Al–Cu phase, in addition to some Fe-rich impurity which are difficult to eliminate by heat treatment [45]. The chemical composition of some of these coarse phases was analyzed using EDS point scan, as listed in Table 2. These plate-like precipitates and short rod-like particles were identified to be the T1 phase and the S(Al2CuMg) phase [24,26,36]. The nanoscale T1 phase is precipitated during low temperature aging. The formation of coarse T1 phase is due to the slow cooling rate during furnace cooling after homogenization which enables T1 to precipitate and coarsen. It has been reported that the presence of coarse T1 phase improved workability of the Al–Cu–Li alloy and triggered particle stimulated nucleation (PSN) by high misorientation generated near coarse T1 [26]. In contrast, PSN triggered by the coarse T1 phase was attributed to the gradual rotation of the intensively strained matrix subdivided by the T1 phase during deformation, resulting in increased misorientation with adjacent regions, after which DRX occurs [27].

    Fig. 6
    1. Download : Download high-res image (1MB)
    2. Download : Download full-size image
    Fig. 6

    Table 2. EDS compositions of Point 1, Point 2 and Point 3 in Fig. 6 (wt.%).

    Empty CellAlCuMgZnZrAgFe
    151480.060.60.20.40.2
    250480.040.40.20.50.4
    346521.30.80.040.40.1

    Fig. 7 displays BSE-SEM images of the specimens hot compressed under various conditions. In the low temperature range (300–350 °C), there were a large number of T1 precipitates in the grain interiors and residual coarse second phase aggregates at grain boundaries after deformation, as shown in Fig. 7(a–c). The strain rate increased from 0.01 s−1 to 1 s−1 did not show a significant influence on the presence of the second phase at 300 °C. When the deformation temperature rose to 400 °C, the T1 precipitates become discontinuous and the coarse intermetallic particles become smaller, indicating that they gradually dissolved into the matrix. When the temperature was 500 °C, the T1 was completely dissolved into the matrix and there was only a small fraction of the residual Al–Cu phase and the Fe-rich particles, which was confirmed by the EDS results of points A, B and C listed in Table 3. Compared to the coarse particles in Fig. 6, the solute content in these particles decreased with decreasing particle size (a decrease in the solute content dissolved into the matrix, e.g., point C in Fig. 7(f)), corresponded to an increase in the Al content in the particle).

    Fig. 7
    1. Download : Download high-res image (422KB)
    2. Download : Download full-size image
    Fig. 7

    Table 3. EDS compositions of Point A, Point B and Point C in Fig. 7(f) (wt.%).

    Empty CellAlCuMgZnZrAgFe
    A51461.10.80.90.50.2
    B61270.10.50.10.610
    C72250.50.60.0420.4

    Fig. 8 shows the IPF maps of specimens hot compressed at 350–500 °C. All specimens exhibited elongated deformed grains, but there was a clear difference in the microstructure between the specimens deformed below 400 °C and those deformed above 400 °C. Fig. 8(a) shows a large amount of HAGBs, as well as some combinations of HAGBs and LAGBs, which were in the form of many small grains and incomplete grains with partial HAGBs and partial LAGBs within the original grains. The appearance of incomplete grain boundaries normally treated indicates continuous dynamic recrystallization (CDRX) [46]. This kind of microstructure in Fig. 8(a) also occurred in the specimen hot compressed at 400 °C, but with a sharply reduced fraction, as shown in Fig. 8(b). Furthermore, apparent grain subdivisions, i.e., different parts of a grain, show different orientations except for those with<110>. When the deformation temperature was higher than 400 °C, there were rarely HAGBs inside the original grains. In addition, there were a great number of LAGBs in the deformed matrix and the newly formed small DRX grains at pre-existing grain boundaries for the specimen deformed at 450 °C, as well as occasionally serrated or bulging original boundaries. Nucleation of DRX grains through original grain boundary bulging followed by the growth of DRX grains is an important feature of discontinuous dynamic recrystallization (DDRX) [46,47]. When the temperature further increased to 500 °C, the sub grains become well-profiled and there were more DRX grains at the grain boundaries, as indicated by blue ellipse. In addition, most of the new grains had an orientation close to<101>Al and<001>Al.

    Fig. 8
    1. Download : Download high-res image (2MB)
    2. Download : Download full-size image
    Fig. 8

    Fig. 9 shows that, in the temperature range 350–450 °C, the volume fraction of the DRX grains (identified by the green color) decreased from 6% at 350 °C to 1.2% at 450 °C, then slightly increased to 1.9% at 500 °C. Grain orientation spread (GOS) is an important parameter to distinguish the DRX grains from the initial grains. GOS is the average difference in orientation between each point and all measurements within a single grain [48]. The magnitude of the GOS value reflects the lattice distortion and dislocation density, can be directly obtained from the EBSD data, and is almost unaffected by strain and step size [49]. Generally, a higher the GOS value indicates more serious lattice distortion and a higher dislocation density. Therefore, newly formed DRX grains can be distinguished from the deformed matrix by a critical GOS value. Grains with a GOS value less than 2° were considered as the DRX grains [50,51]. The small fraction of the DRX grains suggests that the DRV and the DRX cooperated in the dynamic softening and that the DRV dominated during hot compression under these deformation conditions.

    Fig. 9
    1. Download : Download high-res image (1MB)
    2. Download : Download full-size image
    Fig. 9

    Further analysis of the orientation gradient in the areas identified by the white boxes in Fig. 8(c) and (d) was carried out by misorientation analysis as shown in Fig. 10. The cumulative (point-to-origin) misorientation along Line 1, 2 and 3 exceeded the critical value of HAGBs even though the subgrain boundaries were characterized by low-angle misorientations (point-to-point), suggesting that the misorientation was well developed and CDRX played a major role at 500 °C. Furthermore, the large cumulative misorientation near the original grain boundary can accumulate enough misorientation for the boundaries of growing nuclei to become HAGBs.

    Fig. 10
    1. Download : Download high-res image (1MB)
    2. Download : Download full-size image
    Fig. 10

    Fig. 11 shows the grain boundary evolution for the specimens for different temperatures at a strain rate of 0.01s−1. Except for the HAGBs and LAGBs, the blue lines represent the grain boundaries with misorientation angle between 10 and 15°, which are defined as medium angle grain boundaries (MAGBs). The microstructure evolution during hot compression was significantly influenced by the deformation temperature. The biggest difference between the microstructure of the four specimens was the distribution and the amount of MAGBs. At 350 °C, there was a dense distribution of MAGBs inside the deformed grains that decreased somewhat as the temperature rose to 400 °C. When the temperature was 450 °C and above, the fraction of MAGBs greatly decreased and was mainly distributed near the original grain boundaries rather than within the grains. MAGBs are believed to be the prerequisite for the nucleation of CDRX and the progress of CDRX with temperature can be tracked by the fraction of MAGBs [48,52].

    Fig. 11
    1. Download : Download high-res image (2MB)
    2. Download : Download full-size image
    Fig. 11

    The fractions of each type of grain boundaries in each sample are present in Table 4. The fraction of HAGBs first decreased as the temperature increased to 450 °C and then increased as the temperature reached 500 °C, which was consistent with the trend of the recrystallization fraction determined using GOS. However, the recrystallization fraction of 1.9% at 500 °C should be compared with the fraction of HAGBs that was greater than the recrystallization fraction of 6% at 300 °C. The increased fraction of HAGBs at 500 °C with a small fraction of DRX grains may be related to the coarser subgrains developed at the higher deformation temperature, which can also account for the appearance of fine unrecrystallized grains as shown in Fig. 9(d).

    Table 4. Grain boundary distribution of the specimens hot compressed at different temperatures.

    Deformation conditionsGrain boundary (%)
    2–1010–15>15
    350 °C/0.01s−1631819
    400 °C/0.01s−1741215
    450 °C/0.01s−1814.315
    500 °C/0.01s−1717.322

    Note that the formation of HAGBs within the original grains seems to be closely related to their orientation, as can be seen in Fig. 8(a) and (b). In other word, most of the HAGBs formed during deformation were distributed in the grains with the<001>orientation, while there were few in the grains with the<101>orientation.

    Fig. 12 shows the bright-field (BF) STEM characterization of the microstructure of the specimen after deformation at 300 °C with different strain rates. There were many precipitate bands and pre-existing coarse T1 phase particles as well as some constitutive particles distributed in the matrix. Cell structures formed in the deformed areas, some of which had been transformed to sub grains with thin dislocation boundaries, as shown in Fig. 12(b). There were also dislocation tangles between the bands. Fig. 12(c) shows the enlarged precipitate band. The corresponding EDS-mapping shows that the band was a Cu-rich phase accompanied by Mg and Zn co-segregation. These fine precipitates were identified as dynamically precipitated T1 phase [13,39], which precipitated in a wide temperature range and gradually dissolved into matrix until 500 °C [53]. Hence, the T1 band reduced as the temperature increased. As known, Mg and Zn contained in the Al–Cu–Li series alloys enter the T1 structure. These T1 bands can act as barriers impeding movement of dislocations and subgrain boundaries, as shown in Fig. 12(b) and (d). A strain rate increase to 1s−1 caused (i) a significant increase in the dislocation density so that dislocation tangles became more obvious, and (ii) a significant reduction in the precipitate bands. The substructure became finer and showed a gradual transition from a cellular structure at a low strain rate to a lamellar structure oriented along the sample flow direction, as presented in Fig. 12(e).

    Fig. 12
    1. Download : Download high-res image (844KB)
    2. Download : Download full-size image
    Fig. 12

    When the deformation temperature was 350 °C, there were still many pre-existing T1 phase precipitates distributed in the matrix, which was consistent with the SEM observation in Fig. 7(c). The original large grains were greatly subdivided by these coarse particles. Subgrains formed with straight boundaries were well developed. The enlarged view of the black box area in Fig. 13(b) shows that dislocations accumulated intensively around the coarse T1 phase, interacted with each other and with Al3Zr particles, to form dislocation tangles. In addition, dislocation subgrain boundaries and dislocations in this area were pinned by Al3Zr particles, indicating the strong effect of these small particles on pinning individual dislocation and hindering the movement of subgrain boundaries.

    Fig. 13
    1. Download : Download high-res image (336KB)
    2. Download : Download full-size image
    Fig. 13

    The number of coarse particles continued to decrease when the sample was hot compressed at 400 °C/0.01 s−1, as shown in Fig. 14. Nevertheless, they still played an important role in the formation of subgrains. Fig. 14(b) shows that several subgrains were formed around the residual second phase. A higher magnification of the area framed by the black box in Fig. 14(a) showed that the movement of the dislocations was impeded by the T1 precipitate band, as mentioned before. The subgrain boundaries indicated by the red arrows were composed of multi-directional dislocations, implying that different dislocation slip systems were activated in this area. Increasing temperature enhanced the movement of dislocations. The rearrangement and annihilation of dislocations led to a decrease in the density of dislocations within the original grains and to a formation of sub-grain boundaries. The pinning of dislocations by Al3Zr particles weakened significantly in this case, however the pinning effect on (sub) grain boundaries was till effective, as shown in Fig. 14(d).

    Fig. 14
    1. Download : Download high-res image (701KB)
    2. Download : Download full-size image
    Fig. 14

    Deformation at 450 °C dissolved more second phase particles into the matrix. There was migration of subgrain boundaries and the sub-grain coalescence characterized by the gradual disappearance of sub-boundaries, as shown in Fig. 15(a). At 500 °C, there was clear disappearance of sub-grain boundaries, implying the easy merging of sub-grains, as shown in Fig. 15(c). Conventional DRX (also called DDRX) occurred by grain boundary bulging as shown in Fig. 15(b) and (c), which corresponded to the EBSD image. In contrast, the generation of the small grain shown in Fig. 15(c) (indicated by red arrow) was more likely formed through geometric recrystallization (GDRX), which usually occurs in materials with a large deformation.

    Fig. 15
    1. Download : Download high-res image (532KB)
    2. Download : Download full-size image
    Fig. 15

    4. Discussion

    4.1. Flow stress behavior

    During thermal processing of a metal, work hardening and dynamic softening occur simultaneously. During the initial stage of the deformation, dislocations accumulate inside the grains and the dislocation density increases rapidly, leading to an increase in the flow stress with increasing strain. With a further increase in strain, the shape of the σ-ε curve varies with the degree of dynamic softening. In the temperature range of 300–400 °C, except for the case where the strain rate was 1 s−1, the σ-ε curves showed a single broad peak, indicating that the rate of dynamic softening was rapid compared to the accumulation of dislocations by work hardening. The single peak flow stress behavior occurs frequently in coarse-grained Al alloys [32,33,38]. A previous study suggested that the single peak flow stress behavior of coarse grain materials arises from highly unsynchronized local recrystallization in the form of “necklace” or “cascade” recrystallization [54]. The “necklace” structure in aluminum alloys is usually treated as a sign of DDRX [14]. However, the TEM observations in this work indicated that the single peak flow behavior occurred also with the PSN mechanism. When the specimens were deformed at 450–500 °C, the curves exhibited a distinct steady state after an insignificant peak, which was attributed to the combined effect of DRV and DRX, as shown in Fig. 9(c) and (d).

    The degree of saturation of the alloy matrix before hot deformation affects the deformation activation energy (Q). Q indicates how easily dislocations overcome obstacles during plastic deformation [55]. The value of Q strongly depends on material characteristics, such as stacking faulty energy (SFE), alloy composition, initial microstructure etc. It is well known that materials with high SFE (e.g. aluminum alloys) have lower Q values as they are inclined to DRV, whereas DRX occurs in materials with low SFE. The Q value also varies with solute content. Alloying elements that produce strong solution strengthening or precipitation strengthening can hinder dislocation movement, and thereby can increase the Q value [56]. For the present Al–Li alloy, the slow cooling rate after homogenization leads to a large quantity of second phase precipitation and serious coarsening of precipitates. The limited dissolution of the second phase during thermal holding at low temperatures before hot compression resulted in a relatively small solute content in the matrix compared to the alloy cooled rapidly. Therefore, the smaller effect of solute drag on dislocations during deformation was responsible for the low activity energy, which corresponded to the lower flow stress. In a comparative study of Al–Li alloys, the Q value of the alloy with fully precipitated T1 phase distributed in the matrix after homogenization was lower than that of a solution-treated alloy with no coarse T1 phase particles in the matrix when they were hot deformed [26].

    A relatively high degree of stress softening below 450 °C occurred in the σ-ε curves shown in Fig. 1, which was also reflected by the relative softening shown in Fig. 2. At a low deformation temperature, the slow diffusion-controlled restoration process leads to an increase in the storage energy in grains as well as the occurrence of DRX stimulated by PSN. The combined effect of these two factors can give rise to an increase in the fraction of DRX, leading to a relatively high degree of stress softening. For example, at a strain rate of 0.01/s, the proportion of DRX grains reached a maximum as the temperature increased to 350 °C due to the combined effect of stored energy and the PSN mechanisms, and hence the flow softening increased. When the temperature was further increased, the weakening of the PSN mechanism due to the gradual dissolution of the coarse T1 phase and the reduction of the stored energy due to the enhanced DRV process leads to a significant reduction in the proportion of DRX grains. Only a small fraction of DRX grains occurred by grain boundary bulging and by the increased accumulated misorientation, therefore there was a decrease in flow softening.

    When the strain rate was 1 s−1 and the deformation temperature was between 300 and 350 °C, the flow stress increased after the work hardening stage of the deformation process, indicating only dynamic recovery. In this case, the reduction of dislocations due to dynamic recovery was not enough to counteract the increase of dislocations due to work hardening. However, when the strain rate was further increased to 10 s−1, there was a flow softening behavior similar to that at a strain rate of 0.01 s−1. It can be attributed to the increased accumulated dislocation density due to insufficient time for DRV, which makes it easier to generate DRX. Coupled with the effect of the PSN mechanism, the result was an increase in flow softening at low temperatures. At higher temperatures, the flow softening decreased, which would be the same as at a strain rate of 0.01 s−1. Previous studies have shown that DRX can occur in aluminum alloys at high strain rates [57,58].

    Serrated flow in σ-ε curves has been found in the hot deformation of some metals [47,50], but there is little discussion on this behavior. The present study considered serrated flow to be a Portevin-Le Chatelier (PLC) effect, since the σ-ε curves displayed the most pronounced serrated flow at 300 °C with a strain rate of 1 s−1, where the precipitation of the second phase was limited and there was no recrystallization as indicated by the shape of the σ-ε curve and the microstructural observations. Therefore, the serrated flow was clearly not consistent with the situations mentioned in other literature related to partial recrystallization or the precipitation of second phase particles [29,34]. The PLC effect is attributed to dynamic strain aging (DSA), i.e. the dynamic interaction between movable dislocations and solute atoms leading to the repeated pinning and de-pinning of moving dislocations by a solute atmosphere [59]. This serrated flow behavior tends to occur in the alloys that exhibit negative strain rate sensitivity [44,60]. Fig. 1 shows that the strength of the PLC effect increased with increasing strain rate, but became weaker when the strain rate further increased to 10 s−1. This should be related to the effect of strain rate on dynamic precipitation during hot compression. A higher strain rate allowed less dynamic precipitation, so that more solute atoms remained in the matrix and consequently caused greater pinning of dislocations; and hence the PLC becomes more pronounced with increasing strain rate at constant deformation temperature. The influence of strain rate on dynamic precipitation was shown in Fig. 12. This was also confirmed in Ref. [13]. When the stain rate was too high, such as 10 s−1, the applied effective stress on the moving dislocations was greater, resulting in a faster average rate of dislocation movement [61], so that the diffusion rate of solute atoms under this condition was much lower than the rate of dislocation movement, and thus the effect of dynamic strain aging was weak, leading to a small stress fluctuation.

    4.2. DRX mechanism during hot compression

    EBSD and TEM microstructure investigation revealed that DRV was the main dynamic softening mechanism of the present alloy during hot compression. However, the difference in softening degree at various temperatures was closely related to the fraction of DRX grains, and the DRX mechanism also shifted with temperature.

    At a lower temperature (300 °C), T1 precipitate bands were formed inside the grains. A network of LAGBs developed near the precipitate bands by dislocations rearrangement through cross-slip and climb (Fig. 12(b)). The continuous absorption of dislocations in LAGBs was favorable for the formation of new grains. DRX could also occurred by the PSN mechanism, as there were a lot of pre-existing coarse T1 phase, although there was little obvious DRX grains due to the limited observation area in the TEM characterization. After homogenization, both the coarse plate-like T1 phase and short-rod like S phase existed in the matrix. However, it worth to note that the coarse T1 phase was hard to deform during hot deformation at lower temperatures and promoted PSN induced DRX [24,26,27], while the short rod-like S phase were easier to deform and break, so that it was difficult to provide conditions to trigger the PSN mechanism [24]. Thus, the PSN mechanism refers to the PSN generated by the T1 phase.

    As strain increased to 1 s−1, the relatively high strain rate inevitably formed a higher dislocation density, and interaction between dislocations formed dislocation tangles. The high dislocation density and dislocation tangles made further dislocation motion difficult, leading to a higher flow stress and strain and flow instability. Due to the short deformation time, the rate of dislocations reduction caused by DRV was slower than the rate of dislocations increase caused by deformation. Therefore, there was no peak in the flow stress. However, the increased dislocation density was still not sufficient to reach the critical value for DRX to occur under this condition.

    The activity of dislocations increases with increasing temperature. The redistribution and partial disappearance of dislocations leads to flattered subgrain boundaries. TEM images indicated that DRX grains formed within the original grains, closely related to the PSN mechanism during hot compression at 350 °C. The greatly subdivided matrix was intensively strained due to the barrier effect of the pre-existing coarse T1 phase on grain boundary migration and dislocation movement. We agree that progressively the matrix rotated around the coarse T1 phase to accommodate the strain [27], as the lattice rotates towards<110>in the fcc crystal during compression [62], and a significant number of DRX grains showed an orientation of<110>, as shown in Fig. 8(a). Lattice rotation increased the misorientation with adjacent regions and led to the formation of HAGBs and DRX grains. However, the fine Al3Zr dispersoids contributed a contrary effect on DRX behavior, that is, enhanced the DRX resistance by pinning dislocation motion and subgrain boundaries [63]. Therefore, the formation of the DRX grains also required the driving pressure to originate from accumulated stored energy around large particles to overcome the Zener pinning arising from the Al3Zr particles [64].

    The transformation of LAGBs to HAGBs is affected by the accumulated misorientation or progressive lattice rotation [65], and also depend on the original grain orientations, which were reflected by the EBSD results shown in Fig. 8(a) which showed that there were limited HAGBs in the original grains with<110>orientation. There do exists stable orientations, in which the increase in misorientation is limited and insufficient to reach the transition value between LAGBs and HAGBs even at large strains [66]. This phenomenon should be related to either thermodynamic or kinetic factors. Research has shown that the dislocation density generated within grains differs with their initial orientations during deformation, and showed a positive correlation with the maximum Schmid factor [67]. The grains with<100>and<110>orientation of the Al–Li alloy have the highest and lowest maximum Schmid factor [27], which means that the grains with<100>orientation have a higher accumulated dislocation density, consequently a higher driving force for DRX and for transition from LAGBs to HAGBs. Besides, the new grains had an orientation close to<101>Al and<001>Al. The<101>fiber texture was reported as a typical texture in fcc materials after compression deformation [68,69], whereas the<100>fiber texture that occurred in a Al–Cu–Li alloy experiencing large strains was attributed to the maximum energy release of the system when the new grains were oriented in this direction [27]. However, the DRX grains of the present alloy in this study were limited, further work need to be done to supply more information about this phenomenon.

    The above discussion of the PSN-DRX mechanisms was still applicable at 400 °C. The reduced fraction of DRX grains was mainly due to more coarse particles dissolving into the matrix and their size became smaller. In addition, the T1 phase by dynamic precipitation acted as barrier to impede dislocation motion and subgrain boundary migration, which facilitated the formation of LAGBs. Attention was also paid to the dislocation boundaries shown in Fig. 14(c). During deformation, the inhomogeneity of deformation leads to changes in local stress. Slip systems in different regions of a grain were activated and result in multi-directional dislocations. Misorientation is generated between the areas separated by these dislocations (refer to subgrains) when they meet each other and increased as the strain increased [70], which led to a different orientation.

    The rearrangement of dislocations plays an important role in the evolution of substructures. The fine Al3Zr dispersoids partially dissolved into matrix as the temperature was further increased, resulting in a weakened pinning effect on individual dislocation and boundaries [71], which allowed subgrain boundaries to migrate. The coalescence of subgrains facilitated the formation of HAGBs and could be the nuclei of the DRX grains. The enhanced DRV at higher temperatures reduced the driving force for DRX, so that a low fraction of DRX grains was produced at 450–500 °C.

    The TEM observations confirmed that DRX grains were nucleated at grain boundaries by the boundary bulging mechanism, which is also designated as DDRX, as show in Fig. 15(b). It is generally accepted that partial grain boundary sliding and or shearing during early hot deformation leads to inhomogeneous local strains and serrated boundaries [72]. A different number of dislocation walls was formed at both sides of the serrated grain boundaries, as show in Fig. 15(b). Driven by the stored energy caused by the dislocation walls inside both adjacent grains, grain boundaries began to move toward the grain with a higher dislocation density and front dislocations were absorbed to release energy. However, when the grain was located at a triple junction (Fig. 15(d)) where there can be a large local strain, the DRX grain may form by geometric dynamic recrystallization (GDRX). GDRX is commonly considered to occur in the deformation of Al alloys at large strains at elevated temperatures, and several models have been proposed to illustrate the process of GDRX [65]. When the critical strain is reached, the subgrains formed at the narrow edge of the original grain can pinch off and new grain is formed [73]. DRV is enhanced at high temperatures, while CDRX is considered as a strong process of DRV characterized by continuous absorption of dislocations, which can lead to the formation of HAGBs and new grains [51], as a consequence, a little higher fraction of DRX grains formed at 500 °C.

    5. Conclusions

    • (1)

      An Arrhenius type constitutive equation for the Al–Cu–Li–Zn alloy hot deformed at 300–500 °C with strain rate 0.001–10s−1 was established as: ε˙=3.182×1011[sinh(0.0152σ)]4.568exp(167.16RT)" role="presentation" style="box-sizing: border-box; margin: 0px; padding: 0px; display: inline-block; line-height: normal; font-size: 14.4px; word-spacing: normal; overflow-wrap: normal; text-wrap: nowrap; float: none; direction: ltr; max-width: none; max-height: none; min-width: 0px; min-height: 0px; border: 0px; position: relative;">˙=3.182×1011[sinh(0.0152)]4.568exp(167.16). The processing maps of the alloy at different strains were also established. The domain suitable for hot deformation was established as the domain of 430–500 °C, with strain rates 0.001–0.1 s−1, and where the power efficiency exceeds 30%. The deformation instability domains were mainly located in the region where the strain rate was greater than 1 s−1. Hot working should be avoided in this region.

    • (2)

      The occurrence of dynamic recrystallization had a close relationship with the orientation of the original grains. DRX rarely occurred in grains with an orientation near<110>Al, while greater DRX occurred in grains oriented near<100>Al, attributed to the difference in maximum Schimid factor of grains with these two orientations, which is associated with the dislocation density.

    • (3)

      The PSN mechanism promoted by the coarse T1 phase was the main mechanism of dynamic recrystallization at deformation temperatures below 450 °C. In the temperature range of 450–500 °C, there was the gradual dissolution of the pre-existing coarse T1 phase, so that DRX mechanism transformed to DDRX by the grain boundary bulging and CDRX by the increased accumulated misorientation.

    Data availability statement

    All data included in this study are available upon request by contact with the corresponding author.

    Declaration of competing interest

    All authors agree this submission and declare no interest conflict.

    Acknowledgments

    This work is supported by Guangdong Provincial Science and Technology Project (No. 0214B090903012) and Xuzhou Science and Technology Project (No. KC20006), China.

    References

    利用本构方程和加工图研究了Al-Cu-Li-Zn合金的热变形特征及其显微组织演变
    The hot deformation characteristics and the associated microstructural evolution of an Al–Cu–Li–Zn alloy studied by constitutive equations and processing maps
    长按图片保存/分享
    0

    DSI上海办联系热线

    021-63201985、19215832933

     

         Copyright ©db 2023, All Rights Reserved

    沪公网安备31011502005771号          

         沪ICP备08010803号  

    添加微信好友,详细了解产品
    使用企业微信
    “扫一扫”加入群聊
    复制成功
    添加微信好友,详细了解产品
    我知道了
    沪ICP备08010803号