Ambient- and elevated temperature properties of Sc- and Zr-modified Al–6Ni alloys strengthened by Al₃Ni microfibers and Al₃(Sc, Zr) nanoprecipitates

Abstract

The eutectic Al–6Ni (wt.%) alloy exhibits excellent strength at ambient and elevated temperature, provided by a high volume fraction of Al₃Ni microfibers formed during solidification. Here, Al–6Ni is micro-alloyed with Sc and Zr (with 0.1Sc+0.2Zr, 0.2Sc+0.4Zr and 0.3Sc+0.2Zr, wt.%), creating two additional populations of primary and secondary Al₃(Sc,Zr) precipitates. The fully eutectic microstructure (α-Al + Al₃Ni) observed in Al–6Ni alloy changes, with Sc and Zr addition to hypoeutectic microstructure with primary α-Al grains nucleated on solidification by primary Al₃(Sc,Zr) precipitates. Upon subsequent aging, fully-coherent Al₃(Sc,Zr) nano-precipitates form in the α-Al matrix between Al₃Ni microfibers, providing substantial precipitation strengthening, which is maintained for up to 1 month at 350 °C. Alloy strength - both at ambient temperature and during creep at 300 °C - can be quantitatively described through a superposition of precipitation strengthening by Al₃(Sc,Zr) nanoprecipitates and load-transfer strengthening by Al₃Ni microfibers.

1. Introduction

Dilute, binary aluminum-scandium (Al–Sc) and aluminum-zirconium (Al–Zr) alloys form, upon aging, a high number density of coherent, stable Al₃Sc and metastable Al₃Zr L1₂-nanoprecipitates, respectively, providing a sizeable precipitation strengthening effect at ambient and elevated temperatures [1–8]. Although the volume fraction of these nanoprecipitates is below 1 vol.% (based on the equilibrium phase diagram) [9], they have high thermal stability and excellent coarsening resistance up to 300 °C for Al₃Sc [1,10–14] and 450 °C for Al₃Zr [5]. This is because of the slow diffusivity of Sc and Zr in Al (D_Sc/Al = 2 × 10⁻¹⁷ m²/s, DZr/Al = 1 × 10⁻²⁰ m²/s at 400 °C [4]), which is much lower than those of other elements in conventional age-hardenable alloys, such as Si and Mg in Al–Si–Mg and Cu in Al–Cu alloys (D_Si/Al = 8.46 × 10⁻¹⁵ m²/s, DMg/Al = 1 × 10⁻¹⁴ m²/s, DCu/Al = 2 × 10⁻¹⁵ m²/s at 400 °C [15]). The simultaneous addition of Zr and Sc in Al provides a combination of rapid L1₂-precipitation from Sc and slow L1₂-coarsening from Zr [16,17], achieving better coarsening- and creep resistance than if equivalent amounts of Sc and Zr are added separately. This is due to the precipitation of core-shell L1₂ Al₃(Sc,Zr) nanoprecipitates on aging [9,16–20], whose shells are enriched with Zr [9,17,20–24] which decreases the precipitate coarsening rate by hindering Sc diffusion out of the Sc-enriched cores [9,20,25]. Partial replacement of Sc with Zr also reduces the cost of the alloy, while maintaining good mechanical properties at both room and elevated temperature [26].

Much recent interest has focused on slow-coarsening eutectic systems to raise the high-temperature limit of aluminum casting alloys, in particular those based on the Al-Ce [27,28] and Al–Ni [13,29–32] eutectic compositions. Eutectic Al–Ni alloys exhibit good long-term thermal stability of the eutectic Al₃Ni phase formed during solidification [13,27,29,30,33–38], which has a rod- or fiber shape with sub-micron diameter and high aspect ratio. A fully eutectic microstructure - consisting of colonies of aligned Al₃Ni microfibers embedded in α-Al matrix grains (essentially an in-situ fiber-reinforced metal matrix composites [29]) – is formed under conventional casting conditions [13, 29] at the eutectic Ni concentration of 6.1 wt% (2.9 at.%) [39]. The resulting Al₃Ni microfibers, which can be semi-coherent or incoherent with the Al matrix [32], resist coarsening up to 400–500 °C [27,29]. This good castability and low hot-tear tendency of the alloy, due to the high volume fraction of the Al₃Ni microfibers (~10 vol.%), is useful for conventional casting [35,40] and powder-bed fusion additive manufacturing [41–43].

Dual-scale strengthening of Al alloys – with phases with very different size, shape and volume fractions - can significantly improve mechanical properties at room and elevated temperature [29,33, 44–46]. Examples include high volume fraction of incoherent Al₂O₃ particles and coherent Al₃Sc nanoprecipitates in binary Al–Sc with Al₂O₃ additions [44]; high volume fraction Al₃Ni microfibers formed upon solidification, with Al₃Sc [13,29] or Al₃Zr [33,46] or Al₃(Zr,Er,Si) [45] nanoprecipitates created after aging in Al–Ni-Sc, Al–Ni–Zr and Al–Ni–Zr–Er–Si alloys, respectively. The improvement in mechanical properties results from the combined effects of dislocation pinning by Al₂O₃ particles [44], load transfer and/or microfiber/dislocation interactions for Al₃Ni microfibers, and precipitation strengthening for Al₃Sc [13,29], Al₃Zr [33,46] or Al₃(Zr,Er,Si) [45] nanoprecipitates. Also, the precipitation kinetic of phases during solidification (for Al₃Ni microfibers) and aging (for Al₃Sc and Al₃Zr nanoprecipitates) are unaffected by the other alloying elements, and the solubility of Sc and Zr in Al₃Ni microfibers is very low, just like that of Ni in Al₃Sc and Al₃Zr [29, 45]. Finally, in ternary Al–Ni–Zr alloys [46], Zr segregates at the Al–Al₃Ni interface, slowing the coarsening kinetics of the Al₃Ni microfibers.

The objectives of this work are to study microstructure and strength, at ambient and elevated temperatures, of Al–Ni-Sc-Zr cast alloys containing two strengthening phases: Al₃Ni microfibers forming during solidification and core-shell Al₃(Sc,Zr) nanoprecipitates formed during aging. The precipitates are characterized by SEM and TEM observations and their evolution is followed via electrical conductivity and Vickers microhardness measurements for various isochronal and isothermal aging conditions. Both ambient strength and creep resistance at 300 °C are measured and compared to those displayed by alloys with only one population of strengthening phase: (i) Al₃Ni microfibers in binary eutectic Al–Ni alloys and (ii) Al₃(Sc,Zr) nano-precipitates in ternary Al-Sc-Zr alloys.

2. Experimental procedures

2.1. Alloy preparation

Seven alloys were prepared by induction-melting in a SiC crucible, using pure Al (99.9% purity) and master alloys (Al–20Ni, Al-2Sc and Al–10Zr): binary Al–6Ni; ternary Al-0.1Sc-0.2Zr, Al-0.2Sc-0.4Z, and Al-0.3Sc-0.2Zr; quaternary Al–6Ni-0.1Sc-0.2Zr, Al–6Ni-0.2Sc-0.4Zr and Al–6Ni-0.3Sc-0.2Zr (all compositions are in wt.% unless otherwise stated). Melt cleanliness was provided by fluxing and degassing via Ar purging, and after the dross removal, the melt was poured at 800 °C into a thick-walled cylindrical copper mold (90 mm outer diameter, 30 mm inner diameter and 70 mm height), which produced a cooling rate of ~10 °C/s. The chemical composition of each alloy, which is an average of three separate measurements performed using spark emission spectrometry, is reported in Table 1. The alloys additionally contain small concentrations of Fe and Si impurities.

Table 1.

Alloy chemical composition as determined by spark emission spectrometry.

Nominal composition, wt. %	experimental composition, wt.% (at.%)						Sc/Zr ratio,	Sc + Zr ratio,
	Ni	Sc	Zr	Si	Fe	Al	wt.% (at.%)	wt.% (at.%)
Al-0.1Sc-0.2Zr	0.02 (0.01)	0.10 (0.06)	0.19 (0.06)	0.13 (0.13)	0.10 (0.05)	Bal.	0.5 (1)	0.29 (0.12)
Al-0.2Sc-0.4Zr	0.04 (0.02)	0.20 (0.12)	0.40 (0.12)	0.13 (0.13)	0.11 (0.05)	“	0.5 (1)	0.60 (0.24)
Al-0.3Sc-0.2Zr	0.02 (0.01)	0.30 (0.18)	0.19 (0.06)	0.13 (0.13)	0.11 (0.05)	“	1.5 (3)	0.49 (0.24)
Al–6Ni	5.91 (2.81)	0* (0)	0* (0)	0.12 (0.12)	0.05	“	-	-
Al–6Ni-0.1Sc-0.2Zr	6.03 (2.87)	0.11 (0.07)	0.19 (0.06)	0.11 (0.11)	0.05 (0.03)	“	0.5 (1)	0.30 (0.13)
Al–6Ni-0.2Sc-0.4Zr	5.99 (2.86)	0.24 (0.15)	0.38 (0.12)	0.11 (0.11)	0.05 (0.03)	“	0.5 (1)	0.62 (0.27)
Al–6Ni-0.3Sc-0.2Zr	5.95 (2.84)	0.32 (0.20)	0.22 (0.07)	0.11 (0.11)	0.05 (0.03)	“	1.5 (3)	0.54 (0.27)

2.2. Microstructure characterization

Macro-/microstructural observations of as-cast alloys were performed on cross-sections perpendicular to the longitudinal direction of cast cylinders and at a height of 15 mm from the bottom of the casting. The sections were prepared by standard metallographic methods, using Poulton’s reagent as an etchant to reveal the grain structure. The macrostructure was imaged by the high-resolution flat-based scanner. The as-cast microstructure was imaged by a field emission scanning electron microscope (FE-SEM, JSM-7800F, JEOL) equipped with an energy-dispersive spectroscopy (EDS) detector. The microstructures of isothermally peak-aged (350 °C/24h) alloys were investigated using transmission electron microscopy (TEM, ARM 200CF, JEOL). The TEM thin foils were obtained by electrochemical polishing of the disk-shaped specimens (3 mm in diameter and ~0.1 mm in thickness) extracted from the mid-center region of the representative castings. Twin-jet electrochemical polishing was performed at 20 V DC in a solution of nitric acid (25%) and methanol (75%), whose temperature was maintained at 25 ± 7 °C during polishing process.

2.3. Aging studies

The 30 mm diameter ingots were sectioned into ~1 mm slices which were further cut into four quarters, used for aging heat treatments and subsequent measurement of Vickers microhardness and electrical conductivity. For isochronal aging, specimens were aged in 3 h steps with 25 °C increments between 200 and 600 °C. Isothermal aging was conducted at 300 or 350 °C up to 1344 h (~2 months). All aging treatments were performed in air and terminated by water quenching. The specimens were polished to 1 μm for Vickers microhardness and electrical conductivity measurements. At least ten microhardness measurements were made on each specimen with a 200 g load and 5 s dwell time, using an Innovatest hardness tester (Nova 330). Electrical conductivity was measured using a Sigmatest 2.069 (Forester Instruments). Four measurements were made at each following frequencies: 60, 120, 240, 480 kHz. Errors are reported in terms of one standard deviation from the mean values for both electrical conductivity and Vickers microhardness.

2.4. Compression creep testing

Cylindrical creep specimens (22 mm in height and 10 mm in diameter) were extracted by electro-discharged machine (EDM) from the casting. These specimens, except for binary Al–6Ni, were aged at 350 °C for 24 h to reach the peak-aged condition before testing. Compressive creep experiments were conducted in air at 300 °C in a creep frame with a three-zone heating element. A compressive load was converted from the tensile load through a stainless steel cage coated with boron-nitride lubricant, using a calibrated dead weight. The platen displacement was measured by linear displacement transducer and continuously recorded by a data acquisition unit. Each specimen was tested until a minimum creep strain rate was achieved and the load was then increased. This was repeated for a series of increasing stresses until the specimen failed or up to 20% strain.

3. Experimental results

3.1. As-cast macro and microstructure

Fig. 1 shows representative ingot cross-sections, exhibiting typical grain structures. Except for Al-0.2Sc-0.4Zr and Al-0.3Sc-0.2Zr which show fine equiaxed grain structure (<1 mm) throughout the cross-section, all Ni-containing alloys as well as Al-0.1Sc-0.2Zr exhibit coarse grains (~1 cm long), with high aspect ratio, and elongated along the radial solidification direction, from the edge of cast cylinder towards its center.

Fig. 1.

Optical macrographs of full horizontal cross sections of as-cast specimens, showing grain structure: (a) Al-0.1Sc-0.2Zr, (b) Al-0.2Sc-0.4Zr, (c) Al-0.3Sc-0.2Zr, (d) Al–6Ni, (e) Al–6Ni-0.1Sc-0.2Zr, (f) Al–6Ni-0.2Sc-0.4Zr, and (g) Al–6Ni-0.3Sc-0.2Zr.

The as-cast microstructures of the Ni-containing alloys are shown in Fig. 2(a–d). Binary Al–6Ni (Fig. 2(a)) exhibits a fully eutectic microstructure with colonies of aligned Al₃Ni microfibers within an α-Al matrix, in agreement with previous studies on Al–6Ni alloy [13,29]. By contrast, the Sc/Zr-modified Al–Ni alloys (Fig. 2(b–d)) show hypoeutectic microstructures, with α-Al grains (dendrite- and globular-shaped) present throughout their cross-sections within a binary eutectic α-Al + Al₃Ni matrix, displaying the same Al₃Ni size and morphology as for the binary, fully eutectic Al–Ni alloy (Fig. 2(a)). The fraction of α-Al grains is higher in the Al–6Ni-0.2Sc-0.4Zr and −0.3Sc-0.2Zr alloys (34 ± 2 and 32 ± 3%, respectively, in area fraction) as compared to Al–6Ni-0.1Sc-0.2Zr) in which α-Al grains are rarely seen. Microstructural observations frequently revealed petal-shaped Sc/Zr-rich precipitates within the α-Al grains, as confirmed by EDS mapping analyses of one of such precipitates (Fig. 3). This petal-like morphology is characteristic of L1₂-Al₃Zr [47] and -Al₃Sc [48] primary particles, which have a lattice structure close to that of α-Al, thus being effective at heterogeneous nucleation of α-Al during solidification and refining the grain size of the alloy [48,49]. The formation of these primary particles is undesirable since they reduce the degree of supersaturation of α-Al with Sc and Zr, thus lowering their precipitation-strengthening effect. Moreover, they nucleate α-Al grains, which are free of Al₃Ni microfibers and thus much weaker than to eutectic regions strengthened by a combination of Al₃Ni and Al₃(Sc,Zr) [38].

Fig. 2.

BSE-SEM micrographs of as-cast specimens of (a) Al–6Ni, (b) Al–6Ni-0.1Sc-0.2Zr, showing elongated Al–Al₃Ni eutectic grain with no sign of Zr- and Sc-rich primary particles; top inset shows coarse primary Al dendrites (dark phase) at the center of the specimen. (c) Al–6Ni-0.2Sc-0.4Zr and (d) Al–6Ni-0.3Sc-0.2Zr, showing Al-dendrites (dark), many of which display a nucleating Al₃(Sc,Zr) primary particle (arrows). The bottom insets display high magnification views of the eutectic Al–Al₃Ni region, including eutectic colony boundaries, illustrating that Sc and Zr addition do not significantly alter the fine Al₃Ni fiber morphology.

Fig. 3.

Primary Al₃(Sc,Zr) particle in Al–6Ni-0.2Sc-0.4Zr, with EDS maps showing high concentrations of Zr and Sc, with much lower Si and Ni content. The insets represent the concentration profile of those elements following the EDS line scan, shown as a white line in the SEM micrograph (top left).

The as-cast microstructure of the ternary Al-0.1Sc-0.2Zr, Al-0.2Sc-0.4Zr and Al-0.3Sc-0.2Zr alloys is shown in Supplementary (Figs. S1 (a–c), respectively). In addition to primary α-Al grains that constitute over 99% of the microstructure, Si- and Fe-rich intermetallic phases are present at grain boundaries of all three alloys. Also, primary Al₃(Sc,Zr) particles are occasionally observed in Al-0.2Sc-0.4Zr cross-sections (Fig. S1(d)), consistent with the refined grain structure (Fig. 1(b)) and with reports of primary L1₂-precipitates in cast Al-0.2Sc-0.2Zr alloy [50].

3.2. Isochronal aging

The Vickers microhardness and electrical conductivity (EC) evolution during isochronal aging are presented in Fig. 4(a) and Fig. 4(b), respectively. In the as-cast state, all ternary (Ni-free) alloys show microhardness values higher than for pure Al (~280–370 vs. 200 MPa [51]), reflecting solid-solution strengthening effects of Sc and Zr supersaturated in the α-Al matrix. The as-cast microhardness of the binary Al–6Ni is much higher than those of the three ternary Al-Sc-Zr alloys (~560 vs. ~280–370 MPa), indicating much greater strengthening contribution from Al₃Ni microfibers as compared to Sc- and Zr solid-solution strengthening. The quaternary alloys display the highest as-cast microhardness (~630–700 MPa), reflecting synergistic strengthening contributions from the two sources observed for binary and ternary as-cast alloys.

Fig. 4.

Evolution during isochronal aging of (a) Vickers microhardness and (b) electrical conductivity for the ternary Al-Sc-Zr alloys, and the eutectic binary Al–Ni and quaternary Al–Ni-Sc-Zr alloys. The dashed-line curves for eutectic Al–6Ni-0.1Sc-0.2Zr, Al–6Ni-0.2Sc-0.4Zr and Al–6Ni-0.3Sc-0.2Zr alloys were calculated from Eq. (2), with k = 1.17, 1.12 and 1.42, respectively, from the aging curves of the respective ternary alloys (Al-0.1Sc-0.2Zr, Al-0.2Sc-0.4Zr and Al-0.3Sc-0.2Zr).

The electrical conductivity of Al–6Ni is the highest among all alloys, consistent with its limited solubility in α-Al [39]. The as-cast conductivity significantly decreases when Sc content increases from 0.1 to 0.2–0.3 wt.%Sc (in both ternary and quaternary alloys) or when eutectic Al₃Ni microfibers are added to the ternary alloys. The conductivities of Al-0.2Sc-0.4Zr and Al-0.3Sc-0.2Zr alloys (with or without Ni) are similar, unlike their as-cast microhardness. This may be due to their much finer grain size that results in strong electron scattering at the grain boundaries [52]. Furthermore, the decrease in conductivity due to the addition of 6 wt.%Ni to Al-0.1Sc-0.2Zr alloys is stronger that when doubling the Sc + Zr concentration (to Al-0.2Sc-0.4Zr and Al-0.3Sc-0.2Zr) illustrating the combined effect of size and volume fraction of second phases on electron scattering.

For the binary Al–6Ni alloy, the microhardness remains unchanged during isochronal aging up to 425 °C, above which a steady decrease occurs up to 600 °C, the highest aging temperature in this study, due to the coarsening of Al₃Ni microfibers. However, this coarsening barely affects the conductivity of the alloys, implying that very little Ni enters in solid solution in the matrix up to 600 °C. By contrast, all ternary and quaternary alloys exhibit a substantial increase in hardness with isochronal aging, reaching a peak hardness at ~325–350 °C indicative of aging-induced Al₃(Sc,Zr) precipitation, which is also supported by a steady increase in conductivity. Higher concentrations of Zr and/or Sc in the alloy lead to more pronounced hardness increases upon peak aging, relative to the as-cast state, for both ternary and quaternary alloys as compared to Al-0.1Sc-0.2Zr and Al–6Ni-0.1Sc-0.2Zr (~350–400 vs. ~ 225–250 MPa). Moreover, while the peak hardness for these Sc- and/or Zr-richer alloys remains near constant between 300 and 425 °C (beyond which alloy hardness drops steadily), the two alloys with the lowest Sc + Zr levels (Al-0.1Sc-0.2Zr and Al–6Ni-0.1Sc-0.2Zr) experience substantial hardness loss at aging temperatures of as low as 375 °C. The conductivity of all Sc/Zr-containing alloys increases with aging up to ~500 °C, consistent with reduction of Sc- and Zr- in solid solution associated with Al₃(Sc,Zr) precipitation, but it declines at higher temperature, as expected as Sc and Zr return in solid solution in the matrix, as Al₃(Sc,Zr) precipitates begin to dissolve.

3.3. Isothermal aging

Fig. 5(a–d) show the evolution of microhardness and electrical conductivity of all alloys as function of isothermal aging time at 300 and 350 °C. The hardness of binary Al–6Ni alloy remains constant for up to one month at 300 and 350 °C, decreasing slightly, from 550 to 490 MPa, after an additional month of aging at 350 °C. The conductivity of these alloys remains accordingly unchanged with aging. These results confirm the excellent coarsening resistance of Al₃Ni microfibers implied by the isochronal aging results (Fig. 4(a)).

Fig. 5.

Evolution of (a,b) Vickers microhardness and (c,d) electrical conductivity during isothermal aging at: (a,c) 300 °C and (b,d) 350 °C, for the ternary Al-Sc-Zr alloys and the eutectic binary Al–Ni and quaternary Al–Ni-Sc-Zr alloys. The dashed-line curves for Al–6Ni-0.1Sc-0.2Zr, Al–6Ni-0.2Sc-0.4Zr and Al–6Ni-0.3Sc-0.2Zr were calculated from Eq. (2), with k = 1.23, 1.21 and 1.47, at 300 °C and k = 1.25, 1.40 and 1.51 at 350 °C, from the aging curves of the respective ternary alloys (Al-0.1Sc-0.2Zr, Al-0.2Sc-0.4Zr and Al-0.3Sc-0.2Zr).

All ternary Al-Sc-Zr alloys display noticeable increase in their hardness with aging at both temperatures. Consistent with isochronal aging curves, both ternary and quaternary alloys with the lowest Sc + Zr level harden less than the Sc + Zr-richer alloys. As expected, all three ternary alloys reach peak-aging at 300 °C much more slowly than at 350 °C (~1 month vs. ~ 1 day). As a result, the overaging drop in hardness of these alloys occurs much earlier at 350 °C (~2 weeks) than at 300 °C (>2 months). The alloy conductivity increases continuously with aging time at both temperatures consistent with precipitate growing and coarsening, but not dissolving, for these aging temperatures.

Quaternary Al–Ni-Sc-Zr alloys exhibit trends of hardness and conductivity evolution similar to those of ternary alloys with the same Sc and Zr levels, but the whole aging curves for quaternary alloys are shifted to higher hardness values by ~300 MPa, indicating substantial strengthening contribution of Al₃Ni microfibers superposed to that from L1₂-nanoprecipitates. Aging curves of Sc/Zr-containing alloys, particularly those with higher Sc + Zr levels aged at 300 °C, exhibit two peaks, the first after ~4 h and the second after ~1 month of aging, which can be attributed to the precipitation of faster-diffusing Sc and slower-diffusing Zr, respectively [25,53,54].

3.4. TEM investigation of aged Al–Ni-Sc-Zr alloys

Fig. 6 shows typical microstructures of quaternary Al–6Ni-xSc-xZr alloys peak-aged isothermally at 350 °C for 24 h. The Al₃Ni precipitates exhibit the same cross-sectional diameters as those in as-cast states (Fig. 6(a) vs. Fig. 2(a)), consistent with lack of coarsening. The selected area diffraction pattern (SADP), shown as inset in Fig. 6(a), revealed diffraction spots for L1₂-Al₃(Sc,Zr) precipitates in addition to those originated from α-Al and Al₃Ni. High-magnification TEM micrographs in Fig. 6(b and c) show fine Al₃(Sc,Zr) precipitates (~2 nm in radius) distributed evenly within the α-Al matrix. Atomic resolution TEM micrograph in Fig. 6(d) (and the corresponding fast-Fourier transform (FFT) in inset) show cube-on-cube relationship and the coherent interfaces between α-Al and Al₃(Sc,Zr). Al₃(Sc,Zr) precipitates are also frequently observed at the interface between α-Al and Al₃Ni (Fig. 6(e)); these interfacial precipitates also exhibit a cube-on-cube relationship with α-Al (inset 3 in Fig. 6(e)), similar to those nucleated homogeneously within the α-Al. Further TEM investigations revealed the same characteristics (size, distribution and morphology) for both Al₃(Sc,Zr) and Al₃Ni in various quaternary alloys, and these microstructural details are therefore not presented here.

Fig. 6.

(a) TEM micrograph, acquired along [011]_Al, showing Al₃Ni microfibers with their longitudinal direction aligned with beam, and the corresponding SADP (inset) showing the diffractions spots for the α-Al, Al₃Ni and L1₂-Al₃(Sc,Zr) phases; (b) scanning TEM micrograph, acquired in high-angle annular dark-field mode, showing Al₃Ni microfiber cross-sections within the α-Al matrix, and (c) TEM micrograph, acquired in dark-field mode, showing uniform distribution of Al₃(Sc,Zr) nanoprecipitates between Al₃Ni microfibers; (d) atomic resolution TEM micrograph and corresponding FFT (shown as inset), acquired along [011]_Al, showing the coherent interface and cube-on-cube relationship existing between α-Al and Al₃(Sc,Zr); (e) atomic resolution TEM micrograph showing Al₃(Sc,Zr) which precipitated at an α-Al-Al₃Ni interface, and the corresponding FFT patterns representing the whole area of (e) (inset 1) and some selected areas in (e) (insets 2–4), confirming the constituent phases and the presence of cube-on-cube relationship between α-Al and Al₃(Sc,Zr), which precipitated at the α-Al-Al₃Ni interface.

3.5. Creep properties

Creep data are presented in Fig. 7(a–c) as double-logarithm plots of minimum strain rate vs. applied stress at 300 °C of our Al–Ni and Al–Ni-Sc-Zr alloys, together with literature data for Al-Ni [55] and Al–Ni-Sc alloys [29]; all Sc/Zr-containing alloys are peak-aged, while the eutectic Al–6Ni alloys are in the as-cast state. Fig. 7(a) shows creep data for the seven alloys in the present study, with best fit lines providing the apparent creep exponent (Fig. S2). It is apparent that creep resistance increases from alloys with nanoprecipitates, to the alloy with microfibers, to alloys with both types of reinforcement. Fig. 7(b) presents alloys with a single population of strengthening phase: our three ternary Al-Sc-Zr alloys (with Al₃(Sc,Zr) nanoprecipitates), our Al–6Ni (with Al₃Ni microfibers), and binary Al-0.2Sc and Al-0.4Sc alloys (with Al₃Sc nanoprecipitates, formed on aging at 300 °C for 24 h), from our previous work [29]. The creep behavior of our as-cast Al–6Ni alloy is similar to that of Al–6Ni from our previous study [29], which was prepared by a similar casting method, although it was subjected to a prior heat-treatment (300 °C, 24 h): this confirms that exposure at 300 °C for 24 h does not significantly influence the creep properties of the binary Al–6Ni. Fig. 7(b) shows that, at strain rate above 2 × 10⁻⁸ s⁻¹, the creep-resistance of precipitation strengthened alloys from the literature [29] and the present study can be ranked from the highest to lowest as: Al-0.4Sc [29], Al-0.3Sc-0.2Zr, Al-0.2Sc-0.4Zr, Al-0.2Sc [29], Al-0.1Sc-0.2Zr, directly correlated with the Sc concentration. All these precipitation-strengthened alloys exhibit high apparent stress exponents (14–71), indicating the presence of threshold stress (σ_th) below which creep rate is not experimentally measurable, originating from interactions between dislocations and second phases [56,57]. The strain rate $\dot{\epsilon }$ can be described by modified power-law equation [56]:

$$\dot{\epsilon }=A{\left(\sigma -{\sigma }_{th}\right)}^n\exp \left(-\frac{Q}{RT}\right)$$ (1)

where A is a dimensionless constant (containing the diffusion coefficient, the Burgers’ vector and the shear modulus), σ the applied stress, n the matrix stress exponent, Q the matrix activation energy, R the ideal gas constant, and T the absolute temperature. Based on Eq. (1), the threshold stress is calculated by plotting ${\dot{\epsilon }}^{\frac{1}{n}}$ vs. σ, and extrapolating to $\dot{\epsilon }$ by using a weighted least-squares linear regression. The matrix stress exponent is taken as n = 4.4 for pure Al [58]. The creep threshold stress of all alloys presented in Fig. 7(b) are reported in Table 2. The threshold stress of Al-0.1Sc-0.2Zr (28 MPa) is slightly higher than that of Al-0.2Sc-0.4Zr (26 MPa), even though Sc and Zr contents are lower. However, at strain rates above 1 × 10⁻⁸ s⁻¹, Al-0.2Sc-0.4Zr has better creep-resistance than Al-0.1Sc-0.2Zr. Al-0.3Sc-0.2Zr, with the highest Sc content, exhibits the highest creep threshold stress (34 MPa) than the other two, leaner ternary alloys, which also exceed that of binary Al–6Ni.

Fig. 7.

Double-logarithmic plots of minimum strain rate vs. applied compressive stress at 300 °C for the ternary Al-Sc-Zr alloys and the eutectic binary Al–Ni and quaternary Al–Ni-Sc-Zr alloys, peak-aged at 350 °C for 24 h for (a) all alloys in the present study with the best-fit curves with a threshold stress taken into account; (b) ternary Al-Sc-Zr and binary eutectic Al–Ni, both with a single population of precipitates, compared with alloys from Ref. [29] aged at 300 °C for 24 h and (c) quaternary Al–Ni-Sc-Zr alloys, with two populations of precipitates, compared with alloys from Ref. [29]. (For interpretation of the references of the color in this figure legend, the reader is referred to the Web version of this article.)

Table 2.

Nanoprecipitate radius and volume fraction, alloy creep parameters at 300 °C (apparent stress exponent and threshold stresses), for alloys from literature and the present study, with various aging treatments.

Alloy (wt.%)	Aging treatment	Precipitate radius (nm)	Volume fraction (%)	Apparent stress exponent at 300 °C (−)	Threshold stress at 300 °C (MPa)
Al-0.15Sc-0.16Zr [9]	350 °C/17h	2.7	0.69	–	18
Al–6Ni [29]	300 °C/24h	–	–	13	26
Al-0.2Sc [29]	“	–	–	42	27
Al-0.4Sc [29]	“	–	–	100	38
Al–6Ni-0.2Sc [29]	“	2.0±0.6	0.51±0.07	28	57
Al–6Ni-0.4Sc [29]	“	1.9±0.5	0.87±0.07	32	63
Al-6.41Ni [45]	As-cast	–	–	18	45
Al-6.33Ni-0.33Zr [45]	425 °C/3h	–	–	17–21	31–46
Al–6Ni	As-cast	-	-	14	29
Al-0.1Sc-0.2Zr	350° C/24h	-	-	60	28
Al-0.2Sc-0.4Zr	“	-	-	26	26
Al-0.3Sc-0.2Zr	“	-	-	71	34
Al–6Ni-0.1Sc-0.2Zr	“	-	-	18	41
Al–6Ni-0.2Sc-0.4Zr	“	~4–5	-	16	37
Al–6Ni-0.3Sc-0.2Zr	“	-	-	19	42

Fig. 7(c) presents the creep data of dual population strengthened alloys; our quaternary Al–Ni-Sc-Zr strengthened by Al₃Ni microfibers and core-shell Al₃(Sc,Zr) nanoprecipitates, and ternary Al–Ni-Sc alloys, from our previous study, with Al₃Ni microfibers and Al₃Sc nanoprecipitates [29]. The creep-resistant of dual population strengthening alloys can be ranked from the best toward the worst as: Al–6Ni-0.4Sc [29]>Al–6Ni-0.2Sc [29]>Al–6Ni-0.3Sc-0.2Zr > Al–6Ni-0.1Sc-0.2Zr > Al–6Ni-0.2Sc-0.4Zr. It is apparent that Al–Ni-Sc alloys exhibit higher creep resistance than Al–Ni-Sc-Zr alloys. The latter alloys exhibit primary Al₃(Sc,Zr) precipitates formed on solidification, which are too coarse to provide strengthening but reduce Sc and Zr available for subsequent formation of secondary nanoprecipitates on aging. However, all Al–Ni-Sc-Zr alloys display high apparent stress exponents, indicating the presence of threshold stress similar to those present in Al–Ni-Sc. The creep threshold stresses, calculated from Eq. (1) and reported in Table 2, are similar for Al–6Ni-0.1Sc-0.2Zr (41 MPa) and Al–6Ni-0.3Sc-0.2Zr (42 MPa) and somewhat lower for Al–6Ni-0.2Sc-0.4Zr (37 MPa).

4. Discussion

4.1. Microstructure evolution

For binary Al–6Ni, a fully eutectic α-Al + Al₃Ni microstructure is obtained upon solidification (Fig. 2(a)), consistent with the eutectic reaction expected to occur at 6.1 wt. % Ni according to the binary phase diagram [39]. Adding Sc (0.1%) and Zr (0.2%) to Al–6Ni still generates near-eutectic microstructure, with rare instances of primary α-Al grains observed in the microstructure; further increasing the Sc and Zr concentrations however produces hypoeutectic microstructure, with primary α-Al grains (dendritic- and globular-shaped) constituting roughly one third of the microstructure, embedded within a binary α-Al + Al₃Ni eutectic matrix with the same characteristics as those in binary Al–Ni alloy (Fig. 2). The higher Sc and Zr additions activated the precipitation of primary Al₃(Sc,Zr) during early stage of solidification, and these precipitates nucleated primary α-Al grains upon further cooling (Fig. 2 (c) and (d)). It is known that primary Al₃(Sc,Zr) leads to strong grain refinement in aluminum alloys [48,49], as seen for our ternary Al-0.2Sc-0.4Zr (Fig. 1(b)) and Al-0.3Sc-0.2Zr (Fig. 1(c)) alloys. By contrast, directional growth of eutectic colonies from the ingot edge towards its center (Fig. 1(d–g)) suggests the absence of heterogeneous nucleation sites for eutectic Al₃Ni precipitates in the melt of quaternary alloys; this produces greater degree of melt undercooling, which becomes sufficient to nucleate metastable α-Al on preexisting Al₃(Sc,Zr).

The rare primary α-Al grains observed in Al–6Ni-0.1Sc-0.2Zr indicate that some primary Al₃(Sc,Zr) still formed during solidification of this alloy, but this phase does not form in ternary Al-0.1Sc-0.2Zr as evidenced by its coarse, unrefined grain structure (Fig. 1(a)). This can be explained by the suppressed formation temperature of α-Al after the addition of ~6% Ni to the alloy, which increases the temperature interval at which primary Al₃(Sc,Zr) tends to precipitate. Despite the precipitation of some primary Al₃(Sc,Zr) on solidification, large amount of Zr and Sc still become supersaturated in the α-Al matrices of as-cast quaternary alloys, which then form fine (~2 nm in radius) and uniformly distributed Al₃(Sc,Zr) nanoprecipitates upon subsequent aging (Fig. 6). The uniform Al₃(Sc,Zr) distribution within the α-Al matrix and between Al₃Ni microfibers observed for our quaternary alloys is in contrast to L1₂-strengthened, dilute Al alloys which show segregation of Sc (a eutectic former) and Zr (a peritectic former) within and between dendrites [59–61]. In eutectic Al–Ni alloys, the narrow temperature interval at which α-Al forms during solidification leads to a much more uniform distribution of Sc and Zr solutes within α-Al, unlike the large solidification interval of dilute Al-Sc-Zr alloys which show a segregated dendritic structure with large local variation in Sc and Zr concentrations, and after aging, in Al₃(Sc,Zr) nanoprecipitate number density.

The addition of 6 wt.% Ni to the ternary Al-Sc-Zr alloys does not affect the precipitation kinetics of Al₃(Sc,Zr) during both isochronal and isothermal aging (Fig. 4(a) and Fig. 5(a–b)): incubation times, peak-aging times and time for onset of overaging are all similar. This agrees well with our previous study [29] where 6 wt.% Ni addition to binary Al-0.2Sc and Al-0.4Sc alloys also had no influence on the kinetics of Al₃Sc precipitation, with no segregation of Ni in the Al₃Sc precipitates, even though the diffusion coefficient of Ni (D_Ni/Al = 3 × 10⁻¹² m²/s) is much faster than Sc (D_Sc/Al = 2 × 10⁻¹⁷ m²/s) in Al. It is also consistent with a study of Al-6.3Ni-0.33Zr showing only a weak core-shell segregation of Ni in Al₃Zr nanoprecipitates [55]. It is therefore anticipated that Ni, which precipitates first as Al₃Ni during solidification without scavenging Sc or Zr, has no effect on the subsequent formation, during aging, of Al₃(Sc,Zr) nanoprecipitates.

Nevertheless, TEM observations (e.g., Fig. 6(e)) revealed Al₃(Sc,Zr) nanoprecipitates forming at α-Al− Al₃Ni interfaces. This might be due to local Sc- and Zr-enrichment from segregation at the α-Al− Al₃Ni interface which may reduce the interfacial energy, as Al₃Ni is either semi-coherent or incoherent with α-Al [32]. Earlier atom probe tomography studies on Al–Ni–Zr alloys revealed minor enrichment of Zr at α-Al− Al₃Ni interfaces, which is thought to reduce the coarsening resistance of Al₃Ni [62]. Given that Al₃Ni experiences non-negligible coarsening at ~400 °C [63], this strategy, i.e., Al₃(Sc,Zr) precipitation at α-Al− Al₃Ni interfaces, might reduce coarsening kinetics of Al₃Ni precipitates (this hypothesis will be addressed in future research).

4.2. Ambient-temperature strength

It is apparent from the isochronal and isothermal aging evolutions (Figs. 4 and 5) that both Al₃Ni microfibers and Al₃(Sc,Zr) nanoprecipitates contribute to strengthening in all quaternary alloys. The overall strength increment (Δτ_t) of an alloy with two or more strengthening mechanisms can be described by an empirical equation [44,64, 65]:

$$\Delta {\tau }_t={\left(\sum _i(\Delta {\tau }_i^k)\right)}^{\frac{1}{k}}$$ (2)

where Dτ_i is the strength increment for a specific mechanism and the exponent k is between 1 (linear sum) and 2 (Pythagorean sum). A k value of unity corresponds to the case where the strengthening contributions are independent [66], while a larger value reflects cooperative strengthening, where contributions are less than additive [67]. These exponents were calculated using hardness increments (difference between measured hardness and hardness of high purity Al, 200 MPa [51]) from the binary alloy (Al–6Ni) and the ternary alloys (Al-Sc-Zr) and compared with the hardness increments from quaternary alloys. We used as criterion a least square difference between summation (as per Eq. (2)) of the binary Al–6Ni and ternary Al-Sc-Zr alloys. The calculated k values from isochronal and isothermal aging results are reported in Table 3. Predicted hardness values over the full aging span, following Eq. (2) with a single optimized k value, are shown as dashed lines in isochronal (Fig. 4(a)) and isothermal (Fig. 5(a) and (b)) aging curves.

Table 3.

Values of exponent k in Eq. (2), illustrating the additive strengthening effects of Al₃Ni microfibers and L1₂- Al₃(Sc,Zr) nanoprecipitate, for the present study and for similar alloys from literature.

Alloys	Isochronal aging	Isothermal aging
		300 °C	350 °C	400 °C
Al–6Ni-0.1Sc-0.2Zr	1.17	1.23	1.25	–
Al–6Ni-0.2Sc-0.4Zr	1.12	1.21	1.40	–
Al–6Ni-0.3Sc-0.2Zr	1.42	1.47	1.51	–
Al–6Ni-0.2Sc [29]	1.39	1.23	1.21	–
Al–6Ni-0.4Sc [29]	1.45	1.39	1.41	–
Al-6.31Ni-0.33Zr [45]	–	–	–	1.26

As shown in Table 3 for the three alloys aged under three different conditions, the k values lie in the range of 1.12–1.51, providing a good fit of hardness throughout the whole range of temperatures (200–600 °C, Fig. 4) and times (0.5 h–2 months, Fig. 5) for isochronal and isothermal aging, respectively. Our k values are also close to those reported for similar dual-population strengthened alloys: (i) k = 1.39–1.45 for Al–Ni-Sc strengthened by Al₃Sc nanoprecipitates (mean radius r = 2 nm) and Al₃Ni microfibers (~100 nm in diameter and ~10 μm in length) [29], (ii) k = 1.26 for Al–Ni–Zr strengthened by Al₃Zr nanoprecipitates (r ~2 nm) and Al₃Ni microfibers (~100 nm in diameter) [45] and (iii) k = 1.30 for Al-Sc-Zr strengthened by Al₃(Sc,Zr) (r ~2 nm) with additions of Al₂O₃ particle (r = 150 nm) [44]. The cumulative effects of strengthening via Al₃Ni microfibers and Al₃(Sc,Zr) nanoprecipitates is in contrast with the independent formation of these two phases: Sc and Zr do not affect the Al₃Ni microfibers eutectic formation (provided no primary Al₃(Sc,Zr) particles are formed), while Ni, with its near-zero solubility in Al, does not impact Al₃(Sc,Zr) bulk nanoprecipitation on subsequent aging (except for some precipitation at the interface, Fig. 6(e)).

4.3. Creep behavior

4.3.1. Single-population strengthened alloys

Both L1₂-strengthened Al-Sc-Zr alloys and Al₃Ni-strengthened Al–Ni alloys exhibit a creep threshold stress. A creep threshold stress is commonly found in precipitation strengthened Al-Sc [29,68], Al-Zr [45, 69] and Al-Sc-Zr [9,69] alloys, whose main mechanism is bypass of nanoprecipitate by dislocation climb [70,71]. This threshold stress is dependent on (i) the inter-precipitate distance (given by volume fraction and radius of precipitates), (ii) the uniformity of precipitates within the alloy; (iii) and the mismatch stresses between precipitates and matrix, which depend on the composition of the precipitate. In Fig. 7(b), it is apparent that the creep resistance and the threshold stress of Al-0.2Sc [29], Al-0.1Sc-0.2Zr and Al-0.2Sc-0.4Zr are all quite similar, despite Zr additions varying between 0 and 0.4 wt.% and total L1₂ former concentration between 0.2 and 0.6 wt.%. Zr additions to Al–Sc alloys affect precipitate radius, number density, spatial uniformity and mismatch (and thus, the creep threshold stress which depends on these parameters) in a complex manner. First, as the Al–Zr is a peritectic system and the Al–Sc system is eutectic, Zr and Sc segregate differently upon solidification, leading to Sc-rich and Zr-rich regions in the alloys, and a subsequent precipitation during aging, an inhomogeneous distribution of precipitates [49,69]. Second, Zr is a very slow diffuser in Al and cannot be homogenized, unlike the more rapidly diffusing Sc. Third, this difference in Zr and Sc diffusion coefficients leads to a Sc-rich core/Zr-rich shell structure in the Al₃(Sc,Zr) nanoprecipitates, whose mismatch with the α-Al matrix is dependent on their composition: mismatch reduced in Zr-rich and increased in Sc-rich precipitates [26]. Finally, the three above alloys may have different Si content in the Al matrix, dependent on their Sc and Zr content, as some Si is scavenged by the primary Al₃(Sc,Zr) particles formed on solidification (Fig. 3): Si in solid solution affects the number density and radii of Al₃(Sc,Zr) nanoprecipitates formed on aging [22,60,61,72], and thus the threshold stress.

The threshold stresses at 300 °C for Al–6Ni in the present study (29 MPa) and for Al–6Ni from a previous study (26 MPa) [29]- fabricated with a similar casting method with a 300 °C in aging treatment - are close. However, their creep resistance is much lower than reported for a Al-6.33Ni alloy which had comparably much finer microstructure [45], with a threshold stress of 45 MPa (as reported in Table 2). Readers are referred to Ref. [45] for the detailed discussion of the causes of differences in the creep threshold stresses of Al–6Ni with lower solidification rate [29] and Al-6.33Ni with higher solidification rate [45]. The creep threshold stress of Al–6Ni is higher than quaternary Al-0.1Sc-0.2Zr and Al-0.2Sc-0.4Zr alloys possibly owing to larger volume fractions of Al₃Ni microfibers, providing a higher degree of strengthener than that of a very low fraction of L1₂ nanoprecipitates, although L1₂ precipitates have higher effectiveness to hinder dislocation motion.

4.3.2. Dual-population strengthened alloys

As quaternary alloys are strengthened by both Al₃(Sc,Zr) nanoprecipitates (blocking dislocations) and Al₃Ni microfibers (via load transfer), the creep properties can be described by combining the creep threshold stress with the load transfer coefficient through modifying Eq. (1) as described by Refs. [73–75] as:

$$\dot{\epsilon }=A{\left((1-\alpha )\left(\sigma -{\sigma }_{th}\right)\right)}^n\exp \left(-\frac{Q}{RT}\right)$$ (3)

where α is the load transfer coefficient which lies between zero (when there is no load transfer) and unity (when the load is completely transferred from matrix to the non-creeping Al₃Ni microfiber). The load transfer coefficient can then be calculated by Ref. [75]:

$$\alpha =1-{\left(\frac{{\dot{\epsilon }}_2}{{\dot{\epsilon }}_1}\right)}^{\frac{1}{n}}$$ (4)

where ${\dot{\epsilon }}_2$ and ${\dot{\epsilon }}_1$ are the minimum creep strain rate of the dual-scale strengthened quaternary Al–Ni-Sc-Zr alloy (strengthened by Al₃Ni microfibers and Al₃(Sc,Zr) nanoprecipitates) and of the ternary Al-Sc-Zr alloy (strengthened by only Al₃(Sc,Zr) nanoprecipitates), respectively.

Fig. 8(a),(c) and (e) display double-logarithmic plots of minimum creep strain rate vs. effective stress (σ-σ_th) with the slope, n = 4.4, reflecting dislocation creep for Ni-containing and Ni-free Al-0.1Sc-0.2Zr, Al-0.2Sc-0.4Zr and Al-0.3Sc-0.2Zr alloys, respectively. The load transfer effect between Ni-free and Ni-containing alloys is visible via the shift along the x-axis. Fig. 8(b), (d) and (f) shows similar plot, but using the corrected effective stress (1-α)(σ-σ_th) for x-axis, using load transfer coefficient α = 0.80, 0.53 and 0.90, respectively calculated from Eq. (4). The high value of α indicates that load transfer is significant in these alloys. The range of α in the present study (0.53–0.90) is similar to the range (0.42–0.81) for dual-population strengthened Al–Ni-Sc alloy strengthened displaying both Al₃Ni microfibers and Al₃Sc nanoprecipitates [29].

Fig. 8.

Double logarithm plots of minimum creep strain rate vs. effective compressive stress (σ-σ_th) for dislocation creep of Ni-containing and Ni-free (a) Al-0.1Sc-0.2Zr, (c) Al-0.2Sc-0.4Zr (e) Al-0.3Sc-0.2Zr alloys; double logarithm plots of minimum creep strain rate vs. load-transfer-corrected effective stress (1-α)(σ-σ_th) with n = 4.4 for dislocation creep of Ni-containing and Ni-free (b) Al-0.1Sc-0.2Zr with load transfer coefficient α = 0.80, (d) Al-0.2Sc-0.4Zr with load transfer coefficient α = 0.53 (f) Al-0.3Sc-0.2Zr with load transfer coefficient α = 0.90. The hollow triangular symbol in (e) and (f) was not utilized in the calculation. (For interpretation of the references of the color in this figure legend, the reader is referred to the Web version of this article.)

The macroscopic load-transfer coefficients are likely affected by the fast-creeping Al₃Ni-free dendrites nucleated by primary Al₃(Sc,Zr) particles (present in the two solute-richer alloys) [76], with these dendrites also possibly affecting the orientation of Al₃Ni microfibers in adjacent eutectic regions. Also, the dendrite shape (and possibly their orientation) are somewhat different between the two solute-richer alloys (Fig. 2): Al–6Ni-0.2Sc-0.4Zr shows equiaxed Al-dendrites while Al–6Ni-0.3Sc-0.2Zr exhibits more globular-shaped Al-dendrites without clearly-defined dendritic arms (Fig. 2(c and d)). Finally, the number density and distribution of Al₃(Zr,Sc) nanoprecipitates within the primary Al grains is expected to vary depending on the local concentrations of Sc, Zr and Si. This creates a complex, hierarchical load-transfer situation, first between Al₃Ni-free Al dendrites and the Al–Al₃Ni eutectic matrix (as studied previously in Al dendritic alloys [76]), and second, in the latter eutectic region, between Al₃Ni microfibers and the Al matrix strengthened by Al₃(Zr,Sc) nanoprecipitates.

5. Summary

This study investigated ambient- and elevated-temperature mechanical properties of Sc- and Zr-modified eutectic Al–Ni cast alloys: Al–6Ni micro-alloyed with 0.1Sc+0.2Zr, 0.2Sc+0.4Zr and 0.3Sc+0.2Zr (all wt.%). The alloys exhibit two populations of strengthening phases: Al₃Ni microfibers formed during solidification and core-shell Al₃(Sc,Zr) nanoprecipitates precipitated during aging. A comparison is made with control alloys exhibiting a single population of strengthening phases, i. e., Al₃Ni microfibers in eutectic Al–6Ni or core-shell Al₃(Sc,Zr) nanoprecipitates in Al-(0.1–0.3)Sc-(0.2–0.4)Zr alloys. The key results are:

1. While a fully eutectic microstructure (α-Al + Al₃Ni) is obtained upon solidification of Al–6Ni, adding Sc and Zr, particularly at higher levels, produces hypoeutectic microstructure, with one third of the volume solidifying into primary α-Al grains nucleated by primary Al₃(Sc,Zr) precipitates.

2. Upon isochronal aging, the peak hardness of both Ni-free and Ni-containing Al-0.1Sc-0.2Zr is reached at 300 °C, with overaging hardness drops occurring above 325 °C. For the two Sc-and Zr-richer alloys, peak hardness is achieved at 350 °C and is maintained up to 400–450 °C, for both Ni-free and Ni-containing alloys.

3. Upon isothermal aging, all alloys exhibit double-peak hardness due to Al₃Sc precipitation and subsequent Zr precipitation, forming a Zr-rich shell on the precipitates.

4. Isothermally peak-aged (350 °C/24h) quaternary Al–6Ni-xSc-xZr alloys contain coherent Al₃(Sc,Zr) nanoprecipitates (~2 nm in radius) distributed uniformly within the α-Al matrix and between Al₃Ni microfibers. Al₃(Sc,Zr) precipitation also occurs at α-Al− Al₃Ni interfaces, providing a potential strategy for slowing the coarsening of Al₃Ni precipitates.

5. The evolution of hardness (and thus strength at ambient temperature) during isochronal and isothermal aging of the quaternary Al–Ni-Sc-Zr alloys can be modeled by superposition of the aging behaviors of the respective ternary Al-Sc-Zr and binary Al–Ni alloys, indicating that the effects of Al₃Ni microfibers and Al₃(Sc,Zr) nanoprecipitates are cumulative.

6. For creep at 300 °C, the single-population strengthened alloys (Al–6Ni, Al-0.1Sc-0.2Zr, Al-0.2Sc-0.4Zr and Al-0.3Sc-0.2Zr) exhibit a threshold stress with values of 26–34 MPa. Strengthening occurs by load transfer to Al₃Ni microfibers and Al₃(Sc,Zr) precipitation strengthening, respectively

7. The dual-population strengthened alloys (Al–6Ni-0.1Sc-0.2Zr, Al–6Ni-0.2Sc-0.4Zr and Al–6Ni-0.3Sc-0.2Zr) exhibit higher threshold stresses, reflecting contribution to creep resistance from both Al₃(Sc,Zr) nanoprecipitates and Al₃Ni microfibers, which can be quantitatively described by combining precipitation strengthening and load transfer mechanisms.

Data availability

The raw/processed data required to reproduce these findings cannot be shared at this time as the data also forms part of an ongoing study.