Grain size-dependent Sc microalloying effect on the yield strength-pitting corrosion correlation in Al-Cu alloys

Materials Science & Engineering A 721 (2018) 200–214

Contents lists available at ScienceDirect

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

Grain size-dependent Sc microalloying effect on the yield strength-pitting corrosion correlation in Al-Cu alloys ⁎

⁎

S.H. Wu, P. Zhang, D. Shao, P.M. Cheng, J. Kuang, K. Wu, J.Y. Zhang , G. Liu , J. Sun

T

⁎

State Key Laboratory for Mechanical Behavior of Materials, Xi’an Jiaotong University, Xi’an 710049, PR China

A R T I C L E I N F O

A B S T R A C T

Keywords: Al alloys Grain size effect Interfacial segregation Microalloying Strength and corrosion

Coarse-grained (CG), fine-grained (FG), and ultrafine-grained (UFG) Al-2.5 wt% Cu (Al-Cu) alloys were respectively prepared, with and without 0.3 wt% Sc addition, for comparison. The influences of minor Sc addition on the microstructural evolution, tensile mechanical properties and pitting corrosion resistance were systematically studied at different grain scales. A significant Sc microalloying effect on the precipitation was observed that the minor Sc addition promoted the dispersion of finer θ′-Al2Cu precipitates in the CG alloy and favored the intragranular θ′-Al2Cu precipitation in the FG and UFG alloys, with the smaller grain size leading to a stronger Sc microalloying effect. The Sc addition induced convincing increases in the yield strength at all the three grain scales, and improved (in the CG and UFG Al-Cu alloys) or retained (in the FG Al-Cu alloy) the pitting corrosion resistance at the same time. This indicates that the inverse strength-pitting corrosion correlation as usually observed can be broken by minor Sc addition. The strengthening mechanisms were discussed and the grain sizedependent pitting corrosion resistance mediated by the Sc addition was rationalized in terms of a competition between the positive influence derived from the interfacial Sc segregation and the negative influence come from the deformation-induced dislocations. The present findings provide a possible approach to break the inverse strength-pitting resistance correlation in heat-treatable Al alloys by modifying the precipitate/matrix interfaces through the suitable microalloying atom segregation.

1. Introduction The low density combined with the high strength have made aluminum (Al) alloys the primary material of choice for applications such as in aircraft, where specific strength (strength-to-weight ratio) is a major design consideration [1,2]. Besides the high strength, the excellent corrosion resistance is also a prerequisite for the Al alloys used in practice services [3]. Improving the mechanical properties and simultaneously retaining the excellent corrosion resistance is a relentless pursuit in the Al community to develop advanced Al alloys with enhanced performances. Both the strength and the corrosion resistance of Al alloys are closely dependent on the microstructures. The high strength of Al alloys is predominantly derived from the fine hardening second phase particles that, precipitated from the supersaturated solid solution, act as obstacles to moving dislocations [4]. Other strengthening mechanisms mainly include solid solution strengthening, cluster strengthening, grain boundary strengthening, and dislocation hardening [5–8]. From an electrochemical perspective, the presence of the hardening precipitates causes the microstructure heterogeneous and renders the heat-

⁎

treatable Al alloys susceptible to localized corrosion [9]. Effects of the grain size on the corrosion resistance are relatively complex and some other influencing factors should be considered at the same time [10–13]. Summarizing pertinent literatures mainly on pure metals (including pure Al), Ralston et al. claimed [10] that there exists a relationship between the corrosion rate and the grain size, i.e., the corrosion rate is proportional to the grain size. The underlying mechanism was proposed that the increasing grain boundary density enhances the formation of the protective oxide films on surfaces [10], which hints that fine grain structures are more corrosion resistant. However, this relationship is confined to the case where an oxide/passivity exists. Without the oxide films on surfaces, such as when the current densities are high up to greater than about 10 μA cm−2, the increased grain boundary densities will likely enhance the overall surface reactivity and concomitantly raise the corrosion rate [10,11,13]. In the fine- and ultrafine-grained metals or alloys prepared by the severe plastic deformation, the dislocation density increases accompanied by the refinement in the grain size [7,12,14–16], which promotes the pitting corrosion due to their disorder characteristic [13,14]. Thus a competition exists between the dislocation effect and the grain size effect on

Corresponding authors. E-mail addresses: [email protected] (J.Y. Zhang), [email protected] (G. Liu), [email protected] (J. Sun).

https://doi.org/10.1016/j.msea.2018.02.089 Received 5 December 2017; Received in revised form 23 February 2018; Accepted 25 February 2018 Available online 27 February 2018 0921-5093/ © 2018 Elsevier B.V. All rights reserved.

Materials Science & Engineering A 721 (2018) 200–214

S.H. Wu et al.

thermodynamic/kinetics and lead to the precipitation behaviors much different from those in the CG counterparts [37–41]. Since the precipitates are the predominant factors in determining the pitting corrosion susceptibility, the strength-corrosion correlation is expected to be sensitive to the grain size.

the corrosion susceptibility, especially in the grain-refined pure Al and non-heat-treatable Al alloys. In the heat-treatable Al alloys with precipitates dispersed, however, the corrosion susceptibility was found to be generally dominated by the precipitates [9,17,18]. Pitting is the most common form of localized corrosion in Al alloys, difficult to predict and prevent due to its random occurrences [19]. Pitting corrosion can be the origin of other types of corrosion, including intergranular corrosion, exfoliation corrosion and stress corrosion cracking [20]. It is commonly believed that the approaches to improve the strength of Al alloys through the microstructural modification generally cause detrimental changes to the pitting corrosion resistance [21]. This inverse correlation between the two important engineering properties of the strength and the pitting corrosion resistance substantially limits the practice services of Al alloys. How to break this inverse correlation attracts increasing attentions [22–24]. Pioneer work by Hutchinson et al. [22] was devoted to understanding the effect of the precipitate size and state on the combination of the yield strength and the pitting corrosion susceptibility of an Al-2.5 wt% Cu-1.5 wt% Mg alloy by controlling the aging time. They found that the formation of small coherent atomic solute clusters within the matrix at early aging time did not appear to cause the alloy to become susceptible to the pitting corrosion, but they did increase the yield strength substantially [22]. In addition, the microstructures containing S-phase precipitates with the thickness less than ~ 10 nm were also observed to have similar pitting resistance to the as-quenched solid solution [22,25]. It appears that a critical precipitate size exists in the Al-Cu-Mg alloy, below which the second-phase precipitates cease to act as distinct electrochemical entities. These experimental results highlight the strategy of tailoring the precipitates to break the inverse yield strength-pitting corrosion resistance correlation in Al alloys. Microalloying is an effective approach to tailor the precipitation in Al alloys, in which the addition of minor microalloying elements will remarkably change the precipitate size and distribution [26]. Two microalloying mechanisms have been generally proposed [27–30]: (i) the one is that the microalloying atoms facilitate the heterogeneous nucleation of precipitates, and (ii) the other is that the microalloying atoms segregate to the precipitate/matrix interface, decreasing the interfacial energy and concomitantly reducing the precipitate size. The first mechanism has been experimentally confirmed in the Sn (In or Cd)microalloyed Al–Cu alloys and (Si, Ge)-coadded Al-Cu alloys by using Atom Probe Tomography (APT) and High-resolution TEM [27,28]. While the second mechanism was evident in the Ag-doped Al-Cu-Mg [29] and Al-Cu [30] alloys. The precipitate evolution and resultant hardening/strengthening response related to the microalloying effect have been extensively investigated in Al alloys [31–33]. However, the microalloying effect on the corrosion resistance of Al alloys was much less studied. In particular, the microalloying with the second mechanism modifies not only the precipitate size/distribution but also the precipitate interface, which may impact on the corrosion resistance [34]. The microalloying possibly becomes another strategy to break the inverse yield strength-corrosion resistance correlation, for which systematical researches are required. In this paper, we illustrate that minor Sc addition into a model Al-Cu alloy increases the yield strength and simultaneously improves the pitting corrosion resistance. More interestingly, the Sc microalloying effect displays a significant grain size-dependence within the three length scales studied, i.e., the coarse-grained (CG), fine-grained (FG), and ultrafine-grained (UFG) sizes. The choice of the Sc-microalloyed AlCu alloys with different grain sizes as the studied materials in present study is because: (i) our primary experimental results have shown [35,36] that the Sc microalloying in coarse-grained Al-Cu alloys is mainly the Sc segregation at the θ′-Al2Cu/matrix interfaces, which, enhancing the precipitate hardening, may also affect the corrosion resistance positively; (ii) the refinement of the grain size is in accompany with marked changes in the disorder density and the grain boundary characteristics that may highly influence the precipitation

2. Experimental procedures 2.1. Material preparation and heat treatments Two kinds of alloys with the compositions of Al-2.5 wt% Cu (abbreviated as Al-Cu alloys) and Al-2.5 wt% Cu-0.3 wt% Sc (abbreviated as Al-Cu-Sc alloys) were respectively cast by using 99.99 wt% pure Al, 99.99 wt% pure Cu and Al-2.0 wt% Sc master alloys. All cast ingots were homogenized at 723 K for 5 h. Billets with the size of 100 mm × ∅10 mm were machined from the cast ingots for Equalchannel angular pressing (ECAP) to prepare the FG and UFG samples. After being solution-treated in vacuum for 3 h at 873 K and quenched in cold water, the billets were subjected to 4 and 12 ECAP passes by route Bc [12,15,42] at room temperature, respectively. After ECAP processing, the samples were immediately aged at 398 K for 6 h, 10 h, 20 h, and 30 h, respectively. For comparison, the CG samples were prepared by hot extruding the cast ingots into plates with 14 mm in thickness and 60 mm in width at 723 K. The plates were subsequently solution-treated at 873 K for 3 h, quenched in cold water (defined as CG), and finally aged at 523 K for a series of time. The maximum error of all temperature measurements in the present experiments was ± 1 K.

2.2. Microstructural characterization Microstructural analyses were carried out by using scanning electron microscopy/electron backscattered diffraction (SEM/EBSD), transmission electron microscopy (TEM) and high-resolution TEM (HRTEM). SEM-EBSD specimens of the post-ECAP samples were cut from the cross-sectional center of the ECAP billets, and then electropolished by using an electrolyte of 25% nitric acid and 75% methanol at 253 K (− 20 °C) with an operation voltage of 15 V. 3 mm TEM foils were twin-jet electropolished using the same solution. Statistical results on the number density and size of precipitates were obtained from more than 300 measurements. The volume fraction of precipitates was evaluated following the methods in references [43–45]. The precipitate diameter was determined after correction for the truncation effects based on a method by Crompton et al. [46]. The reader can refer to our previous publications [35,36,47] for more measurement details. The dislocation density was measured by performing X-ray diffraction (XRD) experiments. Each sample was tested at least six times to obtain a set of diffraction profiles. The evaluation of these profiles was done following the Multiple Whole Profile (MWP)-fit method developed by Ungár and co-workers [48–50], where simulated profiles are fitted to the recorded profiles. This is done for all reflections simultaneously with ab initio theoretical functions for the strain- and size-induced profile broadening. The reader can refer to reference [51] for experimental details. Three-dimensional atom probe (3DAP) experiments were performed using a Cameca LEAP 3000 HR instrument. Needle-like ATP samples were prepared from blanks with the size of 300 × 300 µm2× 1 cm by combining the mechanical grinding and a two-step electropolishing [52,53], which consists of the coarse polishing using a solution of 10.0 vol% perchloric acid in methanol and the final polishing using a solution of 2.0 vol% perchloric acid in butoxyethanol. 3DAP experiments were performed using a pulse repetition rate of 200 kHz, with a background gauge pressure of < 6.7 × 10−8 Pa (5 × 10−10 Torr). The specimen temperature was controlled at 30 ± 0.3 K.

201

Materials Science & Engineering A 721 (2018) 200–214

S.H. Wu et al.

samples is insensitive to either the Sc addition or the aging treatment in this case. In accompany with the grain refinement, both the high angle grain boundary (HAGB, ≥ 15°) area fraction (fHAGB) and the average grain boundary misorientation (ϑ) highly increase. Statistical results show fHAGB = ~ 62.2%, ϑ = ~ 32.7° in the FG Al-Cu sample and fHAGB = ~ 63.3%, ϑ = ~ 33.4° in the FG Al-CuSc sample, with no apparent differences between the Sc-added and Sc-free samples. The comparison can be seen between Fig. 1(c) and (d). (iii) When the samples were subjected to 12 ECAP passes, the crosssectional grain size was further reduced to several hundreds of nanometers that lies in the length scale defined as the ultrafinegrained (UFG) samples. Similarly, the grain size of the UFG sample is also slightly affected by either the Sc addition or the aging treatment within the ranges studied. Fig. 1(e) and (f) show the representative EBSD images of the UFG Al-Cu and Al-Cu-Sc samples after aging for 20 h. In contrast, apparent texture preferences can be observed in the peak-aged UFG Al-Cu and Al-Cu-Sc alloys. However, it should also be addressed that in present work, the strength and the pitting corrosion resistance in the current Al-Cu binary alloys at different length scales are predominantly determined by the aging precipitates, the grain size and the dislocations, the texture preference imposes relatively little influence on both the two engineering properties [9,10,14,15,17,57]. The average cross-sectional grain sizes of the UFG Al-Cu and Al-Cu-Sc samples are about 535 and 376 nm, respectively. Statistical results show fHAGB = ~ 90.5%, ϑ = ~ 45.3° in the UFG Al-Cu sample (Fig. 1(g)) and fHAGB = ~ 92.2%, ϑ = ~ 46.30 in the UFG Al-Cu-Sc sample (Fig. 1(h)). Experimental data on the grain measurements is summarized in Table 1.

2.3. Measurements of tensile properties Tensile testing was performed to measure the yield strength (σ0) and the tensile strength (σt) of samples before and after artificial aging treatment. The dog-bone-shaped tensile specimens were machined with a gauge size of 1 mm in thickness, 2 mm in width, and 10 mm in length. All tensile tests were performed using a MTS-C43 Tester at a constant strain rate of 2 × 10−4 s−1. The yield strength was determined as the 0.2% offset from the engineering stress-strain curves. At least 3 samples were tested for each condition to guarantee reproducibility. 2.4. Electrochemical characterization Potentiodynamic polarization experiments and potentiostatic polarization experiments were adopted to quantitatively evaluate the pitting corrosion resistance. Both experiments were conducted in 0.1 M NaCl by using a standard three-electrode system, which consists of a saturated calomel reference electrode (SCE) with a Lug-gin capillary, a large Pt mesh counter electrode and the studied alloys with the area of 0.04 cm2 (0.2 cm × 0.2 cm) as the working electrode. Before each experiment, specimens were ground progressively by hand until 2000 grit SiC grinding papers and finished under ethyl alcohol. Specimens were stored in a refrigerator in order to prevent the natural aging. Potentiodynamic polarization experiments were conducted after 30 min at the open circuit potential (OCP) in deaerated electrolytes, then initiated from – 100 mV below the OCP and scanned upward until the total current densities larger than 5.0 × 10−4 A·cm−2 at a scan rate of 1.0 mV·s−1. Potentiostatic polarization experiments were used to produce measureable current transients at potentials below Epit [22]. These experiments were conducted in deaerated electrolytes at a fixed potential of – 660 mVSCE (this value is evaluated by the Epit of pure Al (99.999%) in 0.1 M NaCl, as shown in Fig. 5(c) in latter section), which was aimed at comparing the results based on an iso-potential [22]. The electrochemical signal gathering time was set as 30 min. A custom program was adopted to discriminate and count the metastable pitting events from a background current [54]. Each electrochemical experiment was tested at least 3 times to guarantee reproducibility.

3.1.2. Dislocations The severe plastic deformation induces a great number of dislocations in the processed metals or alloys. However, the dislocation density will be greatly reduced after the aging treatment, due to the dislocation annihilation effect. For example in the present work, the measured dislocation density was reduced from ~ 8.07 × 1014 m−2 before aging to ~ 2.01 × 1014 m−2 after aging for 20 h in the UFG Al-Cu alloy and from ~ 10.98 × 1014 m−2 before aging to ~ 3.92 × 1014 m−2 after aging for 20 h in the UFG Al-Cu-Sc alloy. Other experimental results showed that the dislocation density was ~ 5.28 × 1012 m−2 in the 8 haged CG Al-Cu alloy, ~ 5.45 × 1012 m−2 in the 8 h-aged CG Al-Cu-Sc alloy, ~ 1.86 × 1014 m−2 in the 20 h-aged FG Al-Cu alloy and ~ 1.93 × 1014 m−2 in the 20 h-aged FG Al-Cu-Sc alloy. The comparisons indicate that there is a great difference in the dislocation density between the CG alloys and the UFG ones, while the difference between the FG and UFG alloys is not so obvious, which is consistent with the experimental and model prediction results by Starink et al. [7,16]. The dislocations will impact on the corrosion resistance rather than the strength, as will be discussed later.

3. Results 3.1. Microstructures 3.1.1. Grains (i) The samples prepared by hot-extrusion have grains in pancake shape. Statistically, the average grain sizes in the Al-Cu-Sc samples are ~ 110 µm in width and ~ 800 µm in length. While in their Scfree Al-Cu counterparts, coarser grains are evident that have width of ~ 180 µm and length of ~ 1300 µm. The extruded Al-Cu and AlCu-Sc samples show grains larger than 100 µm, lying in the length scale we defined as the coarse grained (CG) samples. The refinement of the grains caused by the Sc addition is mainly related to the Al3Sc dispersoids inhibiting the grain growth through Zenerdrag action [36,55] and the Sc solutes promoting the grain heterogeneous nucleation [56]. (ii) The samples prepared by 4 ECAP passes have grains with the crosssectional size of several micrometers, which lies in the length scale defined as the fine grained (FG) samples. Since the aging temperature was relatively low of 398 K, the grain sizes of the FG samples were almost unchanged after the aging treatment even up to 20 h. Fig. 1(a) and (b) show the representative EBSD images of the FG Al-Cu and Al-Cu-Sc samples after aging for 20 h. It appears that the peak-aged FG Al-Cu and Al-Cu-Sc alloys present no apparent texture preference. The average cross-sectional grain sizes of the FG Al-Cu and Al-Cu-Sc samples are about 2.3 and 2.8 µm, respectively. This indicates that the grain size of the FG Al-Cu

3.1.3. Al3Sc constituents and dispersoids The Sc addition into Al alloys generally results in the formation of coarse Al3Sc constituents (Fig. 2(a)) during the melting process and fine Al3Sc dispersoids (Fig. 2(b)) during the homogenization/solution treating. The coarse Al3Sc constituents have been known to be cathodic with respect to the surrounding Al matrix, which causes the Sc-added Al alloys having lower breakdown potentials and suffering higher weight loss compared with the Sc-free alloys [58]. The Al3Sc dispersoids, however, were experimentally found to be electrochemically compatible with Al alloys [59], due to their small size (several tens to hundreds of nanometers in diameter) and comparatively slow cathodic kinetics. In present work, a relatively high solution treatment temperature of 873 K was adopted, which dissolved most of the Al3Sc constituents and the Al3Sc dispersoids [35]. Since the solid solution treatment was 202

Materials Science & Engineering A 721 (2018) 200–214

S.H. Wu et al.

Fig. 1. Representative EBSD orientation maps of the peak-aged FG and UFG Al-Cu (a and e) and Al-Cu-Sc (b and f) alloys, respectively. Statistical results on the high angle grain boundaries (HAGBs) area fraction (fHAGB) and the average grain boundary misorientation (ϑ) are shown in (c and g) and (d and h), respectively.

3.1.4. Grain size-dependent precipitation

identical for all the CG, FG, and UFG samples, the three alloys contained almost the same Al3Sc constituents and Al3Sc dispersoids, in both the size and the number density. For this reason, the influences of the Al3Sc constituents and Al3Sc dispersoids on the pitting corrosion resistance can be simply neglected when discussing the grain size effect later. The grain size-dependent precipitation will be in particular focused.

(i) CG samples with vs without Sc addition: In both the CG Al-Cu and Al-Cu-Sc alloys, the θ′-Al2Cu precipitates are basically dispersed within the grain interior. Fig. 3(a) and (b) present the representative TEM images to show the θ′-Al2Cu precipitates in the CG Al-Cu and Al-Cu-Sc alloys after aging for 8 h. It is clear that the Sc addition refines the size and raises the number density of the θ′203

Materials Science & Engineering A 721 (2018) 200–214

S.H. Wu et al.

severe plastic deformation preparation should be responsible for these unusual phenomena. Firstly, the grain boundaries (GBs) in the FG sample are mainly HAGBs with high energy. From the point of thermodynamic view, the Cu atoms moving to and segregated at GBs can reduce the grain boundary energy and stabilize the FG structure. Secondly, the severe plastic deformation leads to the enhanced defect density [6,15], which can accelerate the Cu atoms diffusing to GBs. In addition, the average diffusion distance for the Cu atoms to GBs is greatly shortened in the FG structure with the grain size of only ~ 2.0–3.0 µm. These meet the kinetics requirements for the Cu atoms to segregate at GBs once exposed to an elevated temperature. Thirdly, the Cu concentration at GBs is high enough to trigger the nucleation of the θ-Al2Cu precipitates, bypassing the metastable precipitates within the traditional precipitation sequence in the CG Al-Cu alloys. As a result, the Cu supersaturation within the grain interior is decreased and the intragranular θ′-Al2Cu precipitation is correspondingly suppressed. However, the Sc addition into the FG Al-Cu sample surprisingly cleared up the intergranular θ-Al2Cu particles. Fig. 3(d) presents a representative TEM image of the FG Al-Cu-Sc sample after aging for 20 h, in which no intergranular θ-Al2Cu particles are detected and all the particles are the θ′-Al2Cu precipitates dispersed within the grain interior. The magic Sc microalloying effect on the precipitation in the FG Al-Cu samples can be rationalized as follows. On the one hand, the Sc atoms can effectively capture the Cu atoms and form the Sc-Cu atom couples due to the strong Sc-Cu interaction [60]. On the other hand, the diffusion rate of the Sc atom in Al matrix is several orders of magnitude slower than the Cu atom [61]. The firm Sc-Cu bonding highly limits the Cu atoms diffusing to the GBs, reducing the Cu concentration at GBs and suppressing the intergranular θ-Al2Cu precipitation. The Sc segregation is also detected at the θ′-Al2Cu/matrix interfaces in the aged FG Al-Cu-Sc sample, as typically shown in Fig. 4(c) and (d). The interfacial Sc concentration is higher than that in the CG Al-Cu-Sc sample. (iii) UFG samples with vs without Sc addition: the precipitation is more subversive in the UFG Al-Cu sample. No intragranular θ′-Al2Cu precipitates were found in the UFG Al-Cu sample after aging for 20 h. Instead, all the precipitates were the intergranular θ-Al2Cu precipitates (Fig. 3(e)). Similar experimental results have been reported in UFG Al-Cu alloys [37,62]. Smaller grains and more HAGBs with high energy in the UFG structure are responsible for this full intergranular precipitation. While in the aged UFG Al-CuSc sample, almost all the intergranular θ-Al2Cu particles disappeared and a full intragranular θ′-Al2Cu precipitation is observed (Fig. 3(f)). This means that the Sc effect is still significant even at the UFG length scale. The interfacial Sc segregation is

Table 1 Measurements on the microstructures of the peak-aged (CG, FG, UFG) Al-Cu and Al-Cu-Sc alloys. Alloys/conditions

Grain size (μm)

Approx. HAGBs (%)

Misorientation (deg)

Al-Cu

180.00 ± 15.00 2.78 ± 0.19 0.54 ± 0.08 110.00 ± 10.00 2.27 ± 0.28 0.38 ± 0.04

– 62.2 ± 3.2 90.5 ± 4.6 – 63.3 ± 3.4 92.2 ± 4.8

– 32.7 ± 1.7 45.3 ± 2.4 – 33.4 ± 1.6 46.3 ± 2.3

Al-Cu-Sc

CG FG UFG CG FG UFG

Al2Cu precipitates. Quantitatively, the average radius (rp) and thickness (hp) were experimentally measured to be ~ 490 nm and ~ 19 nm, and the number density (ρp) was ~ 0.15 × 1019 m−3 in the Al-Cu alloy after aging for 8 h. In the same aged Al-Cu-Sc alloy, by contrast, the average radius and thickness were reduced to ~ 295 nm and ~ 7 nm, and the number density increased to ~ 0.48 × 1019 m−3. The representative 3DAP image and relative atom concentration analyses crossing the precipitate/matrix interfaces are given in Fig. 4(a) and (b), respectively, for the CG AlCu-Sc alloy. A strong Sc segregation is found at the precipitate/ matrix interfaces, with the Sc concentration about 10 times greater than that in the matrix. This interfacial Sc segregation reduces the interfacial energy strikingly and hence inhibits the precipitate growth. The Sc microalloying mechanisms in the CG Al-Cu-Sc alloy include: (a) reducing the precipitate size by decreasing the interfacial energy (driving force for growth) and slowing down the Cu diffusion; (b) facilitating the precipitate nucleation by promoting the Cu-vacancy bonding through the strong Cu/Sc/vacancy interaction [35]. (ii) FG samples with vs without Sc addition: The FG Al-Cu sample displayed the precipitation behaviors much different from that in the CG one. Besides the θ′-Al2Cu precipitates within the grain interior, a quantity of second-phase particles located at the grain boundaries are also observed (Fig. 3(c)). These intergranular particles were proved to be the equilibrium θ-Al2Cu precipitates rather than the θ′-Al2Cu precipitates [57]. In the FG Al-Cu sample aging for 20 h, the volume fraction of the intergranular θ-Al2Cu precipitates was ~ 0.38 vol%, which was of the same order of magnitude as the volume fraction of the intragranular θ′-Al2Cu precipitates (~ 0.73 vol%). The formation of the equilibrium θAl2Cu precipitates at so low a temperature of 398 K disobeys what have been well known in CG Al-Cu model alloys. The microstructural characteristics of the FG sample associated with its

Fig. 2. Representative TEM images showing the coarse Al3Sc constituents (a) and fine Al3Sc dispersoids (b) in the Al-Cu-Sc alloy.

204

Materials Science & Engineering A 721 (2018) 200–214

S.H. Wu et al.

Fig. 3. Representative TEM images showing the size and distribution of the intragranular θ′ precipitates and the intergranualr θ precipitates in the peak-aged CG, FG and UFG Al-Cu (a, c, e) and Al-Cu-Sc (b, d, f) alloys, respectively.

mentioned before. Typical potentiodynamic polarization curves of the CG Al-Cu, CG Al-Cu-Sc alloys aging for 8 h and pure Al are given in Fig. 5(c). The pitting potential (Epit) obtained from such polarization curves can be partially representative of the pitting resistance of materials. Fig. 5(d) quantitatively demonstrates that the Sc addition leads to a striking improvement in Epit. It has been suggested that the parameter of Epit only represents the potential above which the formation of stable pits can be sustained [63]. To comprehensively reveal the pitting corrosion resistance, current transient experiments were adopted to quantify the metastable pitting events [18,22,64], which are composed of the nucleation and rapid repassivation of pits. A metastable pitting event can be distinguished as a small current spike in the background current [63,65], as typically shown in the zoom-in of Fig. 5(e). Fig. 5(e) shows a representative current transient curve of the CG Al-Cu-Sc aging for 8 h. In order to distinguish an individual metastable pitting event from the background current, small working electrodes (0.2 cm × 0.2 cm) were used. As revealed from Fig. 5(f), the Sc addition yields a lower rate of metastable pitting events in the CG Al-Cu-Sc sample than that in the CG Al-Cu one. Therefore, the metastable pitting events and Epit evolutions present the same trend, i.e., the lower metastable pitting events is in accompany with the higher Epit, which is consistent with the notion that the nucleation of metastable pits is crucial in the propagation of stable pits [65]. The origin of the Sc microalloying effect on the electrochemical response in the CG samples is closely related to the Sc-dependent precipitate evolution. It has been claimed that the smaller dimensional size of precipitates (such as the thickness of plate-like precipitates) usually plays a critical role in determining the pitting resistance, i.e., the smaller precipitate size leads to the higher pitting

further intensified in the UFG Al-Cu-Sc sample, and the representative 3DAP image and corresponding concentration analyses are shown in Fig. 4(e) and (f). For comparison, the measured size, volume fraction and number density of the intragranular θ′-Al2Cu precipitates (rpintra /hpintra , fintra , and N pintra ) and the intergranular θ-Al2Cu precipitates (rpinter /hpinter , finter , and N pinter ) of the peak-aged CG, FG, UFG Al-Cu and Al-Cu-Sc samples are summarized in Table 2. One can see that the intragranular θ′-Al2Cu precipitates in the CG alloys have greater average sizes but lower number densities than those in the FG and UFG alloys. Coarser intergranular θ-Al2Cu precipitates are evident in the FG Al-Cu ( finter ~ 0.38 vol%) and UFG Al-Cu ( finter ~ 1.65 vol%) alloys. The average size of the intragranular θ′-Al2Cu precipitates is larger and the number density is smaller in the UFG Al-Cu-Sc alloy than that in the FG Al-Cu-Sc alloy.

3.2. Strength and electrochemical response 3.2.1. CG Al-Cu samples with and without Sc addition Fig. 5(a) and (b) show the yield strength and the tensile strength of the CG Al-Cu and Al-Cu-Sc alloys, respectively, as a function of the aging time. Generally, the strength increases monotonically with the aging time. The Sc addition induces a remarkable increase in the strength. For example, the CG Al-Cu-Sc alloy aging for 8 h displays the yield strength increased by 100 MPa and the tensile strength by 46 MPa when comparing with its Sc-free counterpart. The Sc-induced increase in the strength is attributed to the Sc-promoted precipitation, i.e., the reduced precipitate size and the enhanced number density as 205

Materials Science & Engineering A 721 (2018) 200–214

S.H. Wu et al.

Fig. 4. 3DAP maps typically showing the distribution of Cu and Sc atoms across a sectioned θ'-Al2Cu precipitate in the CG Al-Cu-Sc alloy aging at 523 K for 8 h (a), FG Al-Cu-Sc (c), and UFG Al-Cu-Sc (e) alloys aging at 398 K for 20 h, respectively (orange = Cu atoms, blue = Sc atoms, dimensions: 40 × 30 × 30 nm3, 20 × 20 × 40 nm3 and 26 × 26 × 38 nm3, respectively). Only Cu and Sc atoms are presented in (a, c, e) to clearly reveal the Sc segregation at the θ'/matrix interface. (b), (d) and (f) are the corresponding 1-D concentration analyses along the precipitate/matrix interface marked in (a), (c) and (e), respectively.

Table 2 Measurements on the precipitate parameters of the intergranular θ-Al2Cu precipitates, the intragranular θ′-Al2Cu precipitates and the dislocation densities of the peak-aged (CG, FG, UFG) Al-Cu and Al-Cu-Sc alloys. Alloys

Length scales

rpinter(nm)

finter(vol%)

Npinter(1019m−3)

rpintra(nm)

hpintra (nm)

fintra(%)

Npintra(1019m−3)

Dislocation density(1014 m−2)

Al-Cu

CG FG UFG

– 32.6 ± 4.5 45.3 ± 4.7

– 0.38 ± 0.12 1.65 ± 0.19

– 2.84 ± 0.37 5.29 ± 0.60

490.0 ± 30.0 16.8 ± 2.4 –

19.00 ± 0.50 1.75 ± 0.10 –

2.05 ± 0.15 0.73 ± 0.08 –

0.15 ± 0.01 104.00 ± 17.00 –

0.05 ± 0.01 1.86 ± 0.35 2.01 ± 0.36

Al-Cu-Sc

CG FG UFG

– – –

– – –

– – –

295.0 ± 10.0 18.9 ± 2.2 38.6 ± 4.5

7.00 ± 0.50 1.68 ± 0.09 4.52 ± 0.40

2.02 ± 0.18 1.62 ± 0.13 1.66 ± 0.21

0.48 ± 0.03 215.00 ± 25.00 79.00 ± 9.00

0.06 ± 0.01 1.93 ± 0.34 3.92 ± 0.37

Note: (1) rpinter, finter, and Npinter are the average size, volume fraction, and number density of the intergranular θ-Al2Cu precipitates, respectively. (2) rpintra, hpintra, fintra, and Npintra are the average size, thickness, volume fraction, and number density of the intragranular θ′-Al2Cu precipitates, respectively.

206

Materials Science & Engineering A 721 (2018) 200–214

S.H. Wu et al.

Fig. 5. Yield strength (a), tensile strength (b), pitting potential (d) and metastable pitting events (f) as a function of the aging time for the CG Al-Cu and Al-Cu-Sc alloys aging at 523 K. (c) Potentiodynamic polarization curves of the CG Al-Cu, Al-Cu-Sc alloys (aging at 523 K for 8 h) and pure Al for comparison; (e) a typical current transient record for the CG Al-Cu-Sc alloy aging at 523 K for 8 h, the sample was held potentiostatically at − 660 mVSCE during the current transient collection, the magnification of a single metastable pitting event is provided in the inset.

yield strength is increased by about 110 MPa and the tensile strength by about 105 MPa after the Sc addition. The enhanced strength is ascribed to the prominent Sc microalloying effect that turns the partial intergranular θ-Al2Cu precipitation in the FG Al-Cu alloy to the full intragranular θ′-Al2Cu precipitation in the FG Al-Cu-Sc alloy. Fig. 6(c) and (d) present the evolutions of the Epit and metastable pitting events with the aging time, respectively, of the FG Al-Cu alloys with and without the Sc addition. The Sc addition causes no apparent changes in the pitting potential within the studied aging time range. However, an obvious increase in the metastable pitting events is detected in the FG Al-Cu-Sc alloy after aging for 10–20 h. After aging for either below 10 h or beyond 20 h, the Sc-induced change in the metastable pitting events is not evident. In summary, the Sc addition in the FG Al-Cu alloy, although enhancing the strength, has few effects on the

corrosion resistance [22,24,34]. Here in present CG samples, the thickness of the θ′-Al2Cu precipitates is reduced with the Sc addition, which is responsible for the improvement in the pitting corrosion resistance. Of interest to note is that the Sc microalloying in the CG Al-Cu sample enhances the yield strength and simultaneously improves the pitting corrosion resistance, breaking the inverse yield strength-pitting corrosion correlation. 3.2.2. FG Al-Cu samples with and without Sc addition Fig. 6(a) and (b) show the aging time-dependent yield strength and tensile strength of the FG Al-Cu and Al-Cu-Sc alloys, respectively. The peak aging time of both the two samples was ~ 20 h. The Sc-induced increase in the strength is slightly higher in the FG alloys than that in the CG ones. Quantitatively in the peak-aged FG Al-Cu-Sc alloy, the 207

Materials Science & Engineering A 721 (2018) 200–214

S.H. Wu et al.

Fig. 6. Yield strength (a), tensile strength (b), pitting potential (c) and metastable pitting events (d) as a function of the aging time for the FG Al-Cu and Al-Cu-Sc alloys aging at 398 K.

simultaneously in the CG and UFG Al-Cu alloys. Besides, the microalloying effect is highly grain size-dependent. These indicate that the effects of the precipitates on the two properties can be decoupled, through the Sc microalloying. On the other hand, a coupling effect of the precipitates and the grain size on the strength-pitting resistance combination is highlighted. In this section, the single Sc microalloying effect as well as the coupling effect of the Sc microalloying and the grain refinement on the breaking of the traditional strength-corrosion resistance correlation will be respectively discussed in terms of the underlying mechanisms.

pitting corrosion susceptibility. The inverse strength-pitting corrosion resistance correlation is also broken in the FG alloys to some extent. 3.2.3. UFG Al-Cu samples with and without Sc addition The Sc effect in the UFG alloys is more significant than that in the FG alloys. The UFG Al-Cu alloy displays the strength monotonically decreasing with the aging time, which was related to the gradual dislocation annihilation during the aging treatment [6]. The Sc addition, however, results in the greatly enhanced strength, as shown in Fig. 7(a) and (b). This means that the strengthening derived from the intragranular precipitates prevails over the softening induced by the dislocation annihilation. In particular, the peak-aged UFG Al-Cu-Sc alloy (aging for 20 h) also shows a much improved pitting potential (Fig. 7(c)), indicative of an enhancement in the pitting corrosion resistance. Meanwhile, the metastable pitting events are decreased but still slightly higher than that in the UFG Al-Cu alloys (Fig. 7(d)). This discrepancy between the pitting potential evolution and the metastable pitting evens evolution indicates that their dominant influencing factors should be different. Fig. 8 summarizes the correlations between Epit and the yield strength, the metastable pitting events and the yield strength of the peak-aged CG (aging for 8 h), FG (aging for 20 h), and UFG Al-Cu alloys (aging for 20 h) with and without the Sc addition, respectively. The inverse strength-pitting resistance correlation is clearly broken by the Sc microalloying in the CG and UFG alloys. In the FG alloys, the Sc addition can increase the strength and keep the pitting corrosion resistance slightly changed. A notable grain size dependence is evident that requires comprehensive discussions.

4.1. Single Sc microalloying effect in the CG alloys 4.1.1. Effect on the strength In the aged CG Al-Cu alloys, the θ′-Al2Cu precipitates are predominantly dispersed within the grain interior. The strengthening response is closely dependent on the size and distribution of the precipitates. It is generally believed [2,66] that smaller precipitates with a denser distribution will produce the higher yield strength. In present work, the Sc addition into the CG Al-Cu alloy promotes the nucleation of the θ′-Al2Cu precipitates by facilitating the Cu-vacancy binding through the strong Cu/Sc/vacancy interaction [35]. This is partially responsible for the increased precipitate number density compared with the Sc-free Al-Cu alloy (see Table 2). In addition, the atomic length scale examinations have demonstrated the Sc atom segregation at the precipitate/matrix interfaces. The interfacial Sc segregation reduces the interfacial energy and decreases the driving force for the precipitate growth [67,68]. According to Lifshit-Slyozov-Wagner (LSW) model modified by Boyd et al. [69], the variation of the mean radius rp with the aging time t of a dispersion of plate-like precipitates is given by

4. Discussions In the aforementioned section, we present the experimental results that the Sc microalloying breaks the usually observed inverse strengthpitting resistance correlation and increases both the two properties

(rp)3 − (rp0)3 = 208

kγ DCu c0 Vm2 (t − t0), RT

(1)

Materials Science & Engineering A 721 (2018) 200–214

S.H. Wu et al.

Fig. 7. Yield strength (a), tensile strength (b), pitting potential (c) and metastable pitting events (d) as a function of the aging time for the UFG Al-Cu and Al-Cu-Sc alloys aging at 398 K.

where rp0 is the mean precipitate radius at the aging time t0 , γ is the precipitate/matrix interfacial energy, k is a constant, c0 is the equilibrium concentration of Cu in Al matrix at the aging temperature T, DCu is the diffusion coefficient of Cu solutes, Vm is the molar volume of the precipitate, and R has its usual meaning. The above expression elucidates that the reduction in the interfacial energy, such as in present work caused by the Sc segregation, will strongly inhibit the precipitate growth and refine the precipitates. Besides, the restraint of the precipitate growth leads to more nucleation of new precipitates, further increasing the number density. The strengthening response can also be correlated to the interfacial energy in a simple model. The minimum inter-particle spacing (λmin ) between the plate-like precipitates is dependent on the interfacial energy and the onset driving force (Δ Gv ) [29,31]:

λmin = (Λγ )/ΔGv,

(2)

where Λ is a constant. Substituting the above expression into the Orowan equation leads directly to the relationship between the strengthening increment from the plate-like precipitates (Δσp ) and the precipitate/matrix interfacial energy:

Δσp∝Δτp ∝ (μb)/ λmin ∝ (μb  ΔGv )/ γ ,

(3)

where μ is the shear modulus of the Al matrix and b is the Burgers vector. This simple equation indicates that the strengthening increment is inversely proportional to the interfacial energy. As the interfacial Sc segregation can greatly reduce the interfacial energy, which is doomed to increase the strength. 4.1.2. Effect on the pitting corrosion resistance In Al alloys, the pitting susceptibility results from the local galvanic cells and the variation in the passive film stability between the soluterich precipitates and the matrix. In particular, the Cu-rich precipitates

Fig. 8. The correlations between the pitting potential and the yield strength (a), the metastable pitting events and the yield strength (b) in the peak-aged (CG, FG, UFG) Al-Cu and Al-Cu-Sc alloys.

209

Materials Science & Engineering A 721 (2018) 200–214

S.H. Wu et al.

4.2. Coupling effect of Sc microalloying and grain refinement

act as local cathodes with respect to the Al matrix that facilitates the oxygen reduction reactions and finally drive the anodic dissolution of the surrounding matrix [9,22]. The precipitate-induced heterogeneities that impede the dislocation motion and provide the strength are often the same ones that drive the local galvanic corrosion [22]. This means that precipitates with a higher number density should induce a lower pitting corrosion resistance, i.e. higher metastable pitting events. In present work, however, the CG Al-Cu-Sc alloy, with more θ′-Al2Cu precipitates, displayed the pitting corrosion resistance superior over its Sc-free counterpart. Other dominant factors seem to be existed, including the precipitate size and the interfacial condition. Hutchinson et al.[22] found a critical precipitate thickness lying between ~ 3 nm and 8 nm in an Al-Cu-Mg alloy, below which the precipitates or atomic solute clusters did not appear to cause the alloy susceptible to the pitting corrosion. The viewpoint of a critical precipitate size was subsequently supported by other researchers [25,70]. In the present 8 h-aged CG Al-Cu alloy, the Sc addition reduced the thickness of θ′-Al2Cu precipitates from ~ 19 nm down to ~ 7 nm. This is in broad agreement with Hutchinson et al.’s conclusion that the ultrafine precipitates are much less harmful to the pitting corrosion resistance. Besides the precipitate refinement induced by the Sc microalloying, the Sc segregation at the precipitate/matrix interfaces itself may impact on the pitting corrosion resistance, because the segregation driven reduction in the interfacial energy can depress the interfacial activity and also decrease the electrochemical difference between the precipitate and the matrix. To further understand the intrinsic effect of the interfacial Sc segregation on the pitting corrosion resistance, we performed additional experiments. The CG Al-Cu-Sc alloy was aged at 473 K and 573 K for 8 h, respectively, which were compared with the previous alloy aging at 523 K for 8 h. The average thickness of precipitates in the 473 K-aged Al-Cu-Sc alloy (~ 10 nm) was slightly greater than that in the 523 K-aged one (~ 7 nm), while the average thickness of precipitates in the 573 K-aged Al-Cu-Sc alloy (~ 15 nm) was close to that of the 523 K-aged Al-Cu alloy (~ 19 nm) (see Fig. 9(a)). The electrochemical results as compared in Fig. 9(b) clearly show that the 523 Kaged Al-Cu-Sc alloy has the pitting corrosion resistance obviously superior to the 473 K-aged Al-Cu-Sc alloy, and the 573 K-aged Al-Cu-Sc alloy is less corrosion resistant than the 523 K-aged Al-Cu alloy. The underlying mechanism, as revealed by the 3DAP examinations (representatively shown in Fig. 4(a) and Fig. 9(c, e)), is that the Sc concentrations segregated at the interfaces are much different among the three Al-Cu-Sc alloys aging at different temperatures. As seen in Fig. 4(b) and Fig. 9(d, f), the interfacial Sc concentrations are quantitatively measured to be ~ 0.80 at%, ~ 0.28 at%, and ~ 0.21 at% in the 523 K-, 573 K-, and 473 K-aged Al-Cu-Sc alloys, respectively. The much lower interfacial Sc concentration in the 573 K-aged alloy is attributed to the abundant formation of Sc clusters at the relative high temperature that makes the free Sc atoms for the interfacial segregation insufficient. In the 473 K-aged alloy, however, the Sc diffusivity is suppressed due to the relatively low aging temperature, limiting the Sc atoms diffusing to the precipitate/matrix interfaces. In a mediate aging temperature between the above two, the Sc diffusivity could be sufficient and the free Sc atoms available for the interfacial segregation could also be ensured, which is possibly responsible for the strongest interfacial Sc segregation observed in the present 523 K-aged alloy. The present results demonstrate that the pitting corrosion resistance of the Al-Cu-based alloys depends not only on the precipitate thickness, but also on the interface structure (or the interfacial energy) of the precipitates. Actually, similar findings can be also found in Hutchinson et al.’s results that the formation of the small coherent solute clusters was unharmful to the pitting resistance. The effects of the precipitate interface on the strengthening and the corrosion susceptibility should be further studied in our following work.

4.2.1. FG alloys The interfacial Sc segregation in the FG Al-Cu-Sc alloy (with the interfacial Sc concentration of ~ 3.5 at% as shown in Fig. 4(d)) is stronger than that in the CG Al-Cu-Sc alloy (with the interfacial Sc concentration of ~ 0.8 at% as shown in Fig. 4(b)). The intensified interfacial Sc segregation in accompany with the grain refinement is possibly related to two reasons: the one is that the increased dislocations promote the Sc atoms diffusing to the precipitate/matrix interfaces, and the other is that the Sc clusters are difficult to form in the fine grains due to the destruction of the massive defects and hence the free Sc atoms are accordingly boosted for the interfacial segregation. The refined intragranular precipitates with the stronger interfacial Sc segregation is mainly responsible for the greater strength increment in the Sc-microalloyed FG alloy (the yield strength difference between the FG Al-Cu-Sc and Al-Cu alloys) than that in the Sc-microalloyed CG alloy (the yield strength difference between the CG Al-Cu-Sc and Al-Cu alloys), which can be clearly seen in Fig. 8(a). In the CG Al-Cu-Sc alloy, we have elucidated that the interfacial Sc segregation should ameliorate the pitting corrosion susceptibility. While in present work, the FG Al-Cu-Sc alloy presents the stronger interfacial Sc segregation, however, the pitting corrosion resistance is not improved but even slightly inferior to the FG Al-Cu alloy. This indicates that the pitting corrosion resistance of the FG Al-Cu-based alloys should be dominated by other factors. The one possible factor is the grain size or the grain boundaries. The increasing area fraction of the grain boundaries with the grain refinement generally results in more active anodic reactions, causing the surfaces to passivate more readily [10]. From this viewpoint, slower corrosion rates should be achieved by merely decreasing the grain size. Besides, there were also reports that the electrochemical driving force for the initiation and propagation of pits was independent of the grain size, which meant that the grain size had few influences on the pitting resistance [71]. Based on these discussions, the factor of the grain size for negatively affecting the pitting corrosion resistance can be ruled out. Another possible factor may be the dislocations that greatly increase with the grain refinement [7,15]. The dislocations can be considered as the active electrochemical heterogeneities susceptible to corrosion, possibly owing to the high stored strain energy within the structure, which may also be the preferential sites for the dissolution of the precipitates during corrosion [72–74]. There have been direct experimental evidence showing that the increasing dislocation density deteriorated the corrosion resistance, lowering the pitting resistance and increasing the corrosion rate [13,14]. Considering that the dislocation density reached a level of ~ 1014 m−2 in the aged FG Al-Cu and Al-CuSc alloys, about two orders of magnitude higher than that in the CG AlCu and Al-Cu-Sc alloys, the detrimental influence of the dislocations on the pitting resistance can not be neglected. As shown in Fig. 10(a) and (b), the precipitates in the FG Al-Cu-Sc alloy are fully the intragranular θ′-Al2Cu precipitates, which can more efficiently hinder and stabilize the dislocations within the grain interior than the intergranular θ-Al2Cu precipitates in the FG Al-Cu alloy, leading to the entangled dislocations around precipitates. Viewing from the electrochemical perspective, the coalitions of precipitates and dislocations are distinct from other microstructural features. The precipitate entangled with the dislocations enlarges the heterogeneities within small local zones, which will lower the pitting resistance [14,73,74]. In the FG Al-Cu alloy with a comparatively lower density of intragranular precipitates, much fewer precipitate-dislocation entangling heterogeneities were formed and the local electrochemical heterogeneities were accordingly reduced and weakened. Therefore, the detrimental influence of the dislocations on the pitting resistance can be alleviated. In summary, the slight change of the pitting resistance in the peak-aged FG Al-Cu-Sc alloy compared with the FG Al-Cu counterpart should be ascribed to the detrimental influence of the dislocations that is comparable with the positive 210

Materials Science & Engineering A 721 (2018) 200–214

S.H. Wu et al.

Fig. 9. Dependence of precipitate thickness (a), pitting potential and metastable pitting events (b) on the aging temperatures and the compositions. 3DAP maps (c and e) typically showing the distribution of Cu and Sc atoms across a sectioned θ' precipitate in the CG Al-Cu-Sc alloy aging at 473 K and 573 K for 8 h, respectively (orange = Cu atoms, blue = Sc atoms, dimensions: 10 × 10 × 40 nm3 and 30 × 30 × 60 nm3, respectively). (d) and (f) show the corresponding proxigrams of Al, Cu and Sc across the θ′/matrix interface in (c) and (e), respectively.

shown in Fig. 4(d)). The dislocation density in the peak-aged UFG AlCu-Sc alloy is greater than that in the FG one, but they are of the same order of magnitude. The Sc addition-induced great improvement in the pitting potential of the peak-aged UFG Al-Cu-Sc alloy (compared with the UFG Al-Cu one) indicates that the positive influence of the interfacial Sc segregation on the pitting corrosion resistance dominates over the negative influence of the dislocations. Of special interest to note is that, although the pitting potential is significantly improved in the UFG Al-Cu-Sc alloy, the metastable pitting events of the UFG Al-Cu-Sc alloy are slightly changed in comparison with that in the UFG Al-Cu alloy. This is because that the two parameters, both representing the pitting resistance, are actually sensitive to different microstructure features [18,22,24,25,34]. The pitting potential is mainly controlled by the local microgalvanic cells caused by the electrochemical discrepancy between the heterogeneities (such as the precipitates and the dislocations, et al.) and the Al matrix, which is reflected on the destruction of the stability in the passive film

influence of the interfacial Sc segregation and hence counteracts the latter.

4.2.2. UFG alloys As mentioned before, the full intergranular θ-Al2Cu precipitation in the UFG Al-Cu alloy was transformed into the full intragranular θ′Al2Cu precipitation in the UFG Al-Cu-Sc alloy. The most significant Sc microalloying effect on the microstructural evolution partially explains why the greatest Sc-induced strength increment (the yield strength difference between the Al-Cu-Sc and Al-Cu alloys) was found in the UFG alloys (see Fig. 8(a)). It was also revealed that the θ′-Al2Cu precipitates in the UFG Al-Cu-Sc alloy have a larger average size while a lower number density than that in the FG Al-Cu-Sc alloy. These match the 3DAP results that the interfacial Sc segregation in the UFG Al-Cu-Sc alloy (with the interfacial Sc concentration of ~ 13.0 at% in the peakaged condition, as shown in Fig. 4(f)) is even stronger than that in the FG Al-Cu-Sc alloy (with the interfacial Sc concentration of ~ 3.5 at%, as 211

Materials Science & Engineering A 721 (2018) 200–214

S.H. Wu et al.

Fig. 10. Representative TEM images showing the precipitates entangled with dislocations in the aged FG Al-Cu-Sc alloy.

the main mechanism for the Sc microalloying effect. The interfacial Sc segregation also exhibits a notable grain size dependence, with the stronger segregation found in the smaller grain scale. The interfacial Sc segregation could reduce the interfacial energy and improve the pitting corrosion resistance. The dislocations are another microstructural feature evolving with the grain size. The FG and UFG alloys have the dislocation densities much greater than the CG ones. The dislocations, especially those entangled with the precipitates, are found to be detrimental to the pitting corrosion resistance. (3) The Sc addition induced apparent increments in the strength at all the three grain scales. In the CG Al-Cu, the minor Sc addition improves the pitting corrosion resistance, due to the refinement in the precipitates and the interfacial Sc segregation. In the FG Al-Cu, the Sc addition changes the pitting corrosion resistance slightly, because the negative effect of the dislocations on the pitting corrosion resistance is comparable with the positive effect of the interfacial Sc segregation. While in the UFG Al-Cu, the more intensified interfacial Sc segregation, by prevailing over the dislocation influence, results in apparent improvements in the pitting corrosion resistance. At all the three grain scales studied here, the inverse strength-pitting resistance correlation in the present Al-Cu alloys is substantially broken by modifying the precipitate/matrix interfaces through the interfacial Sc segregation.

[22,24,25]. While the metastable pitting events depend more on the size and number density of the precipitates, including the intergranular and intragranular ones. Generally, a higher number density leads to more metastable pitting events. However, it was also claimed that a critical thickness was existed for the intragranular precipitates (~ 2.5 nm in Al-Mg-Si alloy) [34], below which the thickness of precipitates dominated the metastable pitting events and above which the number density of precipitates was the main determining factor. A critical thickness of ~ 3–8 nm was similarly detected in Al-Cu-Mg alloy [22]. Besides, the intergranular precipitates with a larger average size will impact on the pitting events. The average thickness of the intragranular precipitates in the present UFG Al-Cu-Sc alloy is ~ 4.5 nm. It is reasonable to regard the number density of intragranular precipitates as the dominant factor in the UFG Al-Cu-Sc alloy. The UFG AlCu-Sc alloy contains intragranular precipitates with the number density more than ten times greater than that of intergranualr precipitates in the UFG Al-Cu alloy. On the other hand, the intergranular precipitates in the UFG Al-Cu alloy are much denser and larger than that in the UFG Al-Cu-Sc alloy. The two kinds of precipitates may produce comparable influences on the metastable pitting events, which can then rationalize the slight change in the metastable pitting events in the UFG Al-Cu alloy before and after the Sc addition. Finally, it should be emphasized that some investigations have been dedicated to the Sc addition to improve the corrosion resistance of Al alloys [58,59,75–77]. However, almost all the previous reports utilized the Al3Sc precipitates or dispersoids to improve the electrochemical stability. Here in present work, however, the Sc effect on the electrochemical stability is realized via the microalloying at the atomic length scale. The modification of the precipitate/matrix interfaces by the segregation of suitable microalloying atoms may be an alternative approach to break the inverse strength-pitting resistance correlation, besides controlling the precipitate size to the ultrathin thickness [22,24,25,70].

Acknowledgements This work was supported by the National Natural Science Foundation of China (Grant Nos. 51621063, 51625103, 51722104, 51571157 and 51790482), the National Key Research and Development Program of China (2017YFB0702301) and the 111 Project of China (B06025). This work is also supported by the International Joint Laboratory for Micro/Nano Manufacturing and Measurement Technologies. GL thanks the support from the Major Program of NSFC. The helpful comments of the reviewer are sincerely acknowledged.

5. Conclusions

References

(1) The Al-Cu and Al-Cu-Sc alloys with different grain size length scales, i.e., CG, FG and UFG, were prepared by the hot-extruding and the ECAP processing, respectively. The Sc addition displays a strong grain size-dependent microalloying effect on the precipitation, i.e., promoting the intragranular precipitation of finer θ′-Al2Cu precipitates in the CG Al-Cu alloy, turning the partial intergranular θ-Al2Cu precipitation to the full intragranular θ′-Al2Cu precipitation in the FG Al-Cu alloy, and completely inhibiting the full intergranular θ precipitation in the UFG Al-Cu alloy. (2) The Sc segregation at the θ′-Al2Cu precipitate/matrix interfaces is

[1] I.J. Polmear, Light Alloys: Metallurgy Of The Light Metals, third ed., Eward Arnold, London, 1995. [2] E. Hornbogen, E.A.S. Jr, Theory assisted design of high strength low alloy aluminum, Acta Metall. Mater. 41 (1993) 1–16. [3] J.R. Davis, Corrosion of Aluminum and Aluminum Alloys, ASM International, United States, 1999. [4] J.W. Martin, 1–The microstructure of precipitation-hardened alloys, in: Precipitation Hardening:second Edition, Elsevier, Netherlands, 1998, pp. 1–78. [5] Ø. Grong, H.R. Shercliff, Microstructural modelling in metals processing, Prog. Mater. Sci. 47 (2002) 163–282.

212

Materials Science & Engineering A 721 (2018) 200–214

S.H. Wu et al.

[38] A. Deschamps, F.D. Geuser, Z. Horita, S. Lee, G. Renou, Precipitation kinetics in a severely plastically deformed 7075 aluminium alloy, Acta Mater. 66 (2014) 105–117. [39] K. Ma, H. Wen, T. Hu, T.D. Topping, D. Isheim, D.N. Seidman, E.J. Lavernia, J.M. Schoenung, Mechanical behavior and strengthening mechanisms in ultrafine grain precipitation-strengthened aluminum alloy, Acta Mater. 62 (2014) 141–155. [40] T. Hu, K. Ma, T.D. Topping, J.M. Schoenung, E.J. Lavernia, Precipitation phenomena in an ultrafine-grained Al alloy, Acta Mater. 61 (2013) 2163–2178. [41] P.V. Liddicoat, X.Z. Liao, Y. Zhao, Y. Zhu, M.Y. Murashkin, E.J. Lavernia, R.Z. Valiev, S.P. Ringer, Nanostructural hierarchy increases the strength of aluminium alloys, Nat. Commun. 1 (2010) 63. [42] G. Sha, Y.B. Wang, X.Z. Liao, Z.C. Duan, S.P. Ringer, T.G. Langdon, Influence of equal-channel angular pressing on precipitation in an Al–Zn–Mg–Cu alloy, Acta Mater. 57 (2009) 3123–3132. [43] G. Liu, G.J. Zhang, X.D. Ding, J. Sun, K.H. Chen, Dependence of fracture toughness on multiscale second phase particles in high strength Al alloys, Mater. Sci. Technol. 19 (2003) 887–896. [44] D.L. Gilmore, E.A. Starke, Trace element effects on precipitation processes and mechanical properties in an Al-Cu-Li alloy, Metall. Mater. Trans. A 28 (1997) 1399–1415. [45] P.M. Kelly, A. Jostsons, R.G. Blake, J.G. Napier, The determination of foil thickness by scanning transmission electron microscopy, Phys. Status Solidi 31 (1975) 771–780. [46] J.M.G. Crompton, R.M. Waghorne, G.B. Brook, The estimation of size distribution and density of precipitates from electron micrographs of thin foils, Br. J. Appl. Phys. 17 (1966) 1301. [47] C. Yang, P. Zhang, D. Shao, R.H. Wang, L.F. Cao, J.Y. Zhang, G. Liu, B.A. Chen, J. Sun, The influence of Sc solute partitioning on the microalloying effect and mechanical properties of Al-Cu alloys with minor Sc addition, Acta Mater. 119 (2016) 68–79. [48] T. Ungár, J. Gubicza, G. Ribárik, A. Borbély, Crystallite size distribution and dislocation structure determined by diffraction profile analysis: principles and practical application to cubic and hexagonal crystals, J. Appl. Crystallogr. 34 (2001) 298–310. [49] G. Ribárik, T. Ungár, J. Gubicza, MWP-fit: a program for multiple whole-profile fitting of diffraction peak profiles by ab initio theoretical functions, J. Appl. Crystallogr. 34 (2001) 669–676. [50] J. Gubicza, L. Balogh, R.J. Hellmig, Y. Estrin, T. Ungár, Dislocation structure and crystallite size in severely deformed copper by X-ray peak profile analysis, Mater. Sci. Eng. A 400 (2005) 334–338. [51] J. May, M. Dinkel, D. Amberger, H.W. Höppel, M. Göken, Mechanical properties, dislocation density and grain structure of ultrafine-grained aluminum and aluminum-magnesium alloys, Metall. Mater. Trans. A 38 (2007) 1941–1945. [52] M.K. Miller, Atom Probe Tomography: Analysis at the Atomic Level, John Wiley & Sons, Inc., Netherlands, 2015. [53] B.W. Krakauer, D.N. Seidman, Systematic procedures for atom-probe field-ion microscopy studies of grain boundary segregation, Rev. Sci. Instrum. 63 (1992) 4071–4079. [54] M.K. Cavanaugh, N. Birbilis, R. Buchheit, A quantitative study on the effects of environment and microstructure on pit initiation in Al alloys, ECS Trans. 16 (2009) 1–11. [55] M.J. Jones, F.J. Humphreys, Interaction of recrystallization and precipitation: the effect of Al3Sc on the recrystallization behaviour of deformed aluminium, Acta Mater. 51 (2003) 2149–2159. [56] E.A. Marquis, D.N. Seidman, Nanoscale structural evolution of Al3Sc precipitates in Al(Sc) alloys, Acta Mater. 49 (2001) 1909–1919. [57] L. Jiang, J.K. Li, P.M. Cheng, G. Liu, R.H. Wang, B.A. Chen, J.Y. Zhang, J. Sun, M.X. Yang, G. Yang, Microalloying ultrafine grained Al alloys with enhanced ductility, Sci. Rep. 4 (2014) 3605. [58] J. Wloka, S. Virtanen, Influence of scandium on the pitting behaviour of Al–Zn–Mg–Cu alloys, Acta Mater. 55 (2007) 6666–6672. [59] M.K. Cavanaugh, N. Birbilis, R.G. Buchheit, F. Bovard, Investigating localized corrosion susceptibility arising from Sc containing intermetallic Al3Sc in high strength Al alloys, Scr. Mater. 56 (2007) 995–998. [60] A.K. Niessen, F.Rd Boer, R. Boom, P.Fd Châtel, W.C.M. Mattens, A.R. Miedema, Model predictions for the enthalpy of formation of transition metal alloys II, Calphad 7 (1983) 51–70. [61] J. Røyset, N. Ryum, Scandium in aluminium alloys, Int. Mater. Rev. 50 (2005) 19–44. [62] M. Murayama, Z. Horita, K. Hono, Microstructure of two-phase Al–1.7at% Cu alloy deformed by equal-channel angular pressing, Acta Mater. 49 (2001) 21–29. [63] R. Newman, Pitting corrosion of metals, Electrochem. Soc. Interface 19 (2010) 33–38. [64] Y. Kim, R.G. Buchheit, A characterization of the inhibiting effect of Cu on metastable pitting in dilute Al–Cu solid solution alloys, Electrochim. Acta 52 (2007) 2437–2446. [65] S.T. Pride, J.R. Scully, J.L. Hudson, Metastable pitting of aluminum and criteria for the transition to stable pit growth, J. Electrochem. Soc. 141 (1994) 3028–3040. [66] G. Liu, G.J. Zhang, X.D. Ding, J. Sun, K.H. Chen, Modeling the strengthening response to aging process of heat-treatable aluminum alloys containing plate/disc- or rod/needle-shaped precipitates, Mater. Sci. Eng. A 344 (2003) 113–124. [67] R. Kirchheim, Reducing grain boundary, dislocation line and vacancy formation energies by solute segregation. I. Theoretical background, Acta Mater. 55 (2007) 5129–5138. [68] E.A. Marquis, D.N. Seidman, M. Asta, C. Woodward, V. Ozoliņs, Mg segregation at Al/Al3Sc heterophase interfaces on an atomic scale: experiments and computations,

[6] Y. Chen, N. Gao, G. Sha, S.P. Ringer, M.J. Starink, Microstructural evolution, strengthening and thermal stability of an ultrafine-grained Al–Cu–Mg alloy, Acta Mater. 109 (2016) 202–212. [7] Y. Chen, N. Gao, G. Sha, S.P. Ringer, M.J. Starink, Strengthening of an Al-Cu-Mg alloy processed by high-pressure torsion due to clusters, defects and defect-cluster complexes, Mater. Sci. Eng. A 627 (2015) 10–20. [8] J. Zhang, N. Gao, M.J. Starink, Al–Mg–Cu based alloys and pure Al processed by high pressure torsion: the influence of alloying additions on strengthening, Mater. Sci. Eng. A 527 (2010) 3472–3479. [9] N. Birbilis, R.G. Buchheit, Electrochemical characteristics of intermetallic phases in aluminum alloys–an experimental survey and discussion, J. Electrochem. Soc. 152 (2005) 467–472. [10] K.D. Ralston, N. Birbilis, C.H.J. Davies, Revealing the relationship between grain size and corrosion rate of metals, Scr. Mater. 63 (2010) 1201–1204. [11] T.S. Mahmoud, Effect of friction stir processing on electrical conductivity and corrosion resistance of AA6063-T6 Al alloy, J. Mech. Eng. Sci. 222 (2008) 1117–1123. [12] R.Z. Valiev, R.K. Islamgaliev, I.V. Alexandrov, Bulk nanostructured materials from severe plastic deformation, Prog. Mater. Sci. 45 (1999) 103–189. [13] K.S. Ghosh, N. Gao, M.J. Starink, Characterisation of high pressure torsion processed 7150 Al–Zn–Mg–Cu alloy, Mater. Sci. Eng. A 552 (2012) 164–171. [14] J.G. Brunner, J. May, H.W. Höppel, M. Göken, S. Virtanen, Localized corrosion of ultrafine-grained Al–Mg model alloys, Electrochim. Acta 55 (2010) 1966–1970. [15] M.J. Starink, X.G. Qiao, J. Zhang, N. Gao, Predicting grain refinement by cold severe plastic deformation in alloys using volume averaged dislocation generation, Acta Mater. 57 (2009) 5796–5811. [16] X.G. Qiao, N. Gao, M.J. Starink, A model of grain refinement and strengthening of Al alloys due to cold severe plastic deformation, Philos. Mag. 92 (2012) 446–470. [17] N. Birbilis, M.K. Cavanaugh, L. Kovarik, R.G. Buchheit, Nano-scale dissolution phenomena in Al–Cu–Mg alloys, Electrochem. Commun. 10 (2008) 32–37. [18] R.K. Gupta, N.L. Sukiman, M.K. Cavanaugh, B.R.W. Hinton, C.R. Hutchinson, N. Birbilis, Metastable pitting characteristics of aluminium alloys measured using current transients during potentiostatic polarisation, Electrochim. Acta 49 (2004) 245–254. [19] W. Tian, S. Li, B. Wang, J. Liu, M. Yu, Pitting corrosion of naturally aged AA 7075 aluminum alloys with bimodal grain size, Corros. Sci. 113 (2016) 1–16. [20] W. Tian, S. Li, X. Chen, J. Liu, M. Yu, Intergranular corrosion of spark plasma sintering assembled bimodal grain sized AA7075 aluminum alloys, Corros. Sci. 107 (2016) 211–224. [21] M.F. Ashby, On the engineering properties of materials, Acta Metall. 37 (1989) 1273–1293. [22] K.D. Ralston, N. Birbilis, M. Weyland, C.R. Hutchinson, The effect of precipitate size on the yield strength-pitting corrosion correlation in Al–Cu–Mg alloys, Acta Mater. 58 (2010) 5941–5948. [23] Z. Wang, P. Chen, H. Li, B. Fang, R. Song, Z. Zheng, The intergranular corrosion susceptibility of 2024 Al alloy during re–ageing after solution treating and cold–rolling, Corros. Sci. 114 (2016) 156–168. [24] K.D. Ralston, N. Birbilis, M.K. Cavanaugh, M. Weyland, B.C. Muddle, R.K.W. Marceau, Role of nanostructure in pitting of Al–Cu–Mg alloys, Electrochim. Acta 55 (2010) 7834–7842. [25] J. Wang, B. Zhang, B. Wu, X.L. Ma, Size-dependent role of S phase in pitting initiation of 2024 Al alloy, Corros. Sci. 105 (2016) 183–189. [26] S.P. Ringer, K. Hono, Microstructural evolution and age hardening in aluminium alloys: atom probe field-ion microscopy and transmission electron microscopy studies, Mater. Charact. 44 (2000) 101–131. [27] S.P. Ringer, K. Hono, T. Sakurai, The effect of trace additions of Sn on precipitation in Al-Cu alloys: an atom probe field ion microscopy study, Metall. Mater. Trans. A 26 (1995) 2207–2217. [28] D. Mitlin, V. Radmilovic, U. Dahmen, J.W.M. Jr, On the influence of Si-Ge additions on the aging response of Al-Cu, Metall. Mater. Trans. A 34 (2003) 735–742. [29] C.R. Hutchinson, X. Fan, S.J. Pennycook, G.J. Shiflet, On the origin of the high coarsening resistance of Ω plates in Al–Cu–Mg–Ag alloys, Acta Mater. 49 (2001) 2827–2841. [30] J.M. Rosalie, L. Bourgeois, Silver segregation to θ′ (Al2Cu)-Al interfaces in Al-Cu-Ag alloys, Acta Mater. 60 (2012) 6033–6041. [31] J.D.C. Teixeira, D.G. Cram, L. Bourgeois, T.J. Bastow, A.J. Hill, C.R. Hutchinson, On the strengthening response of aluminum alloys containing shear-resistant plateshaped precipitates, Acta Mater. 56 (2008) 6109–6122. [32] J.F. Nie, B.C. Muddle, Strengthening of an Al–Cu–Sn alloy by deformation-resistant precipitate plates, Acta Mater. 56 (2008) 3490–3501. [33] H.K. Hardy, The aging characteristics of some ternary aluminium-copper-magnesium alloys with copper: magnesium weight ratios of 7:1 and 2.2:1, J. Inst. Met. 83 (1954) 17–34. [34] S.K. Kairy, P.A. Rometsch, K. Diao, J.F. Nie, C.H.J. Davies, N. Birbilis, Exploring the electrochemistry of 6xxx series aluminium alloys as a function of Si to Mg ratio, Cu content, ageing conditions and microstructure, Electrochim. Acta 190 (2016) 92–103. [35] B.A. Chen, L. Pan, R.H. Wang, G. Liu, P.M. Cheng, L. Xiao, J. Sun, Effect of solution treatment on precipitation behaviors and age hardening response of Al–Cu alloys with Sc addition, Mater. Sci. Eng. A 530 (2011) 607–617. [36] B.A. Chen, G. Liu, R.H. Wang, J.Y. Zhang, L. Jiang, J.J. Song, J. Sun, Effect of interfacial solute segregation on ductile fracture of Al-Cu-Sc alloys, Acta Mater. 61 (2013) 1676–1690. [37] Y. Huang, J.D. Robson, P.B. Prangnell, The formation of nanograin structures and accelerated room-temperature theta precipitation in a severely deformed Al–4 wt% Cu alloy, Acta Mater. 58 (2010) 1643–1657.

213

Materials Science & Engineering A 721 (2018) 200–214

S.H. Wu et al.

Mag. 95 (2015) 2365–2375. [74] R. Parvizi, A.E. Hughes, M.Y. Tan, R.K.W. Marceau, M. Forsyth, P. Cizek, A.M. Glenn, Probing corrosion initiation at interfacial nanostructures of AA2024T3, Corros. Sci. 116 (2016) 98–109. [75] Y. Deng, R. Ye, G. Xu, J. Yang, Q. Pan, B. Peng, X. Cao, Y. Duan, Y. Wang, L. Lu, Corrosion behaviour and mechanism of new aerospace Al–Zn–Mg alloy friction stir welded joints and the effects of secondary Al3ScxZr1− x nanoparticles, Corros. Sci. 90 (2015) 359–374. [76] Y. Deng, Z. Yin, K. Zhao, J. Duan, J. Hu, Z. He, Effects of Sc and Zr microalloying additions and aging time at 120 °C on the corrosion behaviour of an Al–Zn–Mg alloy, Corros. Sci. 65 (2012) 288–298. [77] Z. Ahmad, A. Ul-Hamid, B.J. Abdul-Aleem, The corrosion behavior of scandium alloyed Al 5052 in neutral sodium chloride solution, Corros. Sci. 43 (2001) 1227–1243.

Phys. Rev. Lett. 91 (2003) 036101. [69] J.D. Boyd, R.B. Nicholson, A calorimetric determination of precipitate interfacial energies in two Al-Cu alloys, Acta Metall. 19 (1971) 1101–1109. [70] R.K. Gupta, A. Deschamps, M.K. Cavanaugh, S.P. Lynch, N. Birbilis, Relating the early evolution of microstructure with the electrochemical response and mechanical performance of a Cu-rich and Cu-Lean 7xxx aluminum alloy, J. Electrochem. Soc. 159 (2012) C492–C502. [71] K.D. Ralston, D. Fabijanic, N. Birbilis, Effect of grain size on corrosion of high purity aluminium, Electrochim. Acta 56 (2011) 1729–1736. [72] C. Luo, X. Zhou, G.E. Thompson, A.E. Hughes, Observations of intergranular corrosion in AA2024-T351: the influence of grain stored energy, Corros. Sci. 61 (2012) 35–44. [73] Y.T. Zhou, Y.J. Wang, S.J. Zheng, B. Zhang, X.L. Ma, Strain-induced preferential dissolution at the dislocation emergences in MnS: an atomic scale study, Philos.

214