Structural Characterisation and Mechanical Properties of

6 downloads 0 Views 1MB Size Report
without fracture, which implies ..... and showed a small plastic deformation zone, with fracture .... serrated yielding at constant stress, as was observed previ-.
Materials Transactions, Vol. 43, No. 8 (2002) pp. 2017 to 2025 Special Issue on Bulk Amorphous, Nano-Crystalline and Nano-Quasicrystalline Alloys IV c 2002 The Japan Institute of Metals

Structural Characterisation and Mechanical Properties of Nanocomposite Al-based Alloys Fernando Audebert1,2 , Frederic Prima1 , Marina Galano1 , Marilena Tomut1 , Paul J. Warren1 , Ian C. Stone1 and Brian Cantor1 1 2

Department of Materials, University of Oxford, Parks Road, OX1 3PH, United Kingdom Facultad de Ingenier´ıa, Universidad de Buenos Aires, Paso Col´on 850, Buenos Aires (1063), Argentina

Nanocomposite Al-based alloys can be obtained with a combination of amorphous, crystalline and quasicrystalline phases. In order to understand the correlation between the nanostructure and the mechanical behaviour, four nanocomposite alloys with different characteristics were studied: two alloys from the Al–Fe–Cr–Ti system consisting of a spherical nanoquasicrystalline phase in an α-Al matrix; one alloy from the Al–Fe–V–Ti system consisting of a mixture of amorphous and α-Al phases; and one alloy from the Al–Mn–Cr–Cu system consisting of nanocrystalline particles embedded in an α-Al matrix. Melt-spun samples were prepared and the structure was characterised by means of X-ray diffraction and transmission electron microscopy. Differential scanning calorimetry was used to study the thermal stability and the transformation processes. Tensile tests, fractographic analysis and Vickers microhardness at room temperature were performed in order to evaluate the mechanical behaviour. A combination of solid solution, particle dispersion and grain refinement strengthening was responsible for the high strength of the alloys. The microstructure of the alloy Al93 Fe3 Cr2 Ti2 (at%) remained acceptably stable up to 703 K, due to the slow coarsening rate of the icosahedral phase. (Received May 16, 2002; Accepted July 16, 2002) Keywords: rapid solidification, aluminium-based alloys, nanostructured alloys, mechanical properties, nanoquasicrystalline particles, icosahedral phases

1. Introduction Since the first non-brittle amorphous Al alloys were obtained in 19881, 2) several processes and new Al alloys have been developed. At present, it is possible to produce a wide variety of nanostructured Al-based alloys containing different combinations of amorphous, crystalline and quasicrystalline phases.3, 4) Nanocrystalline Al-based alloys produced using rapid solidification and powder metallurgy techniques have been developed with improved mechanical properties in comparison to conventional crystalline alloys.5, 6) One important route in the development of new high strength Al alloys has been focused on the concept of nanocomposite alloys.4, 7) One type of nanocomposite that has been studied intensively is amorphous nanocomposites composed of α-Al nanocrystals finely dispersed in an amorphous matrix.8, 9) The most important characteristic of these nanocomposites is their high strength, having up to 50% higher tensile strength than the precursor amorphous alloy.8, 9) In the Al–Ni–Y system, it has been shown that the strength is controlled by the solute concentration in the amorphous matrix.10) During the crystallisation process, which produces the dispersed α-Al nanocrystals, the average chemical composition of the amorphous matrix changes resulting in an increase in the strength. The resulting strengths can be described by a simple rule of mixtures involving the strengthened amorphous matrix and α-Al nanocrystals of ideal strength.11) Another promising new type of nanocomposite that has been studied intensively is nanoquasicrystalline composites, composed of nanoquasicrystals dispersed in an α-Al matrix.12, 13) These alloys have higher microstructural stability and higher ductility than the amorphous nanocompos-

ites.14, 15) At present, three important groups of nanoquasicrystalline composites can be distinguished: (i) Al– (Mn/Cr)–Ln (Ln = lanthanides) based with very high strength;12, 16) (ii) Al–(Mn/Cr)–Cu based with the best ductility;17, 18) and (iii) Al–Fe–(Cr/V)–Ti based with the best stability.19, 20) The deformation and strengthening mechanisms in the nanoquasicrystalline composites are still not well known. It has been proposed that an approximant interface phase may improve interfacial bonding between the Al matrix and the quasicrystalline particles, increasing the nanoquasicrystalline composite strength.12) Another type of nanocomposite is that obtained in the Al– (V/Ti)–Fe systems, composed of a very fine nanogranular amorphous phase embedded in an α-Al matrix. The formation of the nanogranular amorphous phase has been related to the presence of icosahedral clusters in the liquid before solidification.21, 22) The different types of Al-based nanocomposites are all metastable alloys with good mechanical properties and potential for structural application. It is important therefore to develop a fundamental understanding of the microstructure formation, stability and mechanical properties of the different types of alloy. In the present work, four nanocomposite alloys from three different alloy systems were studied. One alloy from the Al–Mn–Cu–Cr system was selected for good ductility,17) another from the Al–Fe–V–Ti system20) was selected for its fine-scale nanogranular amorphous composite structure, and two alloys from the Al–Fe–Cr–Ti system19) were selected as nanoquasicrystalline composites with high thermal stability.

2018

F. Audebert et al. Table 1 Chemical composition of the alloys studied. Ti-1

Composition (at%) Nominal Measured

Al 91 90.48 ± 0.09

Fe 5 4.94 ± 0.11

Cr 2 2.16 ± 0.05

Ti 2 2.39 ± 0.06

Si — 0.04 ± 0.01

Cr 2 2.14 ± 0.07

Ti 2 2.12 ± 0.07

Si — 0.05 ± 0.01

V 3 3.28 ± 0.18

Ti 2 2.17 ± 0.1

Si — 0.03 ± 0.01

Cu 2 1.94 ± 0.07

Cr 1 1.02 ± 0.05

Si — 0.04 ± 0.01

Ti-2 Composition (at%) Nominal Measured

Al 93 92.37 ± 0.10

Fe 3 3.32 ± 0.04 V-1

Composition (at%) Nominal Measured

Al 92 91.42 ± 1.65

Fe 3 3.10 ± 0.18

Composition (at%) Nominal Measured

Al 94 94.39 ± 0.12

Mn 3 2.61 ± 0.09

(20/32) i

(18/29) i

As Melt-Spun

Intensity (a.u.)

Ti-1 Ti-2

36

38

(200) Al

V-1 (111) Al

Table 1 shows the nominal chemical composition, the measured composition determined by WDX and the code number of the rapidly solidified alloys. Master alloys were prepared in an arc furnace under He atmosphere using pure elements, Al(99.99%), Cr(99.99%), Fe(99.98%), Mn(99.98%), Cu(99.99%), V(99.7%) and Ti(99.7%). Rapidly solidified samples were obtained by melt spinning onto a rotating copper wheel (40 m/s) in a reduced He atmosphere (200 KPa) using a quartz nozzle, and with a closed-loop controlled ejection temperature of 1200◦ C. As can be seen in Table 1, a small amount of Si contamination from the quartz nozzle was detected. The microstructures of ribbon samples were characterised by X-ray diffraction (XRD) in a θ -2θ diffractometer (Philips 1810) using Cu–Kα radiation, and by transmission electron microscopy (TEM) using a Philips CM-20 microscope. The thin foils for TEM were prepared by electropolishing in 20 vol% nitric acid in methanol at 233 K and a voltage of ∼ 15 V. The structural stability and the transformation processes were monitored by differential scanning calorimetry (DSC) in a TA Instruments model 2200 thermal analyser fitted with a 2010 DSC, using a heating rate of 40 K/min under a dynamic Ar atmosphere. The structural evolution was analysed by means of XRD and TEM on samples heat treated for 45 min at selected temperatures representative of the end of each transformation step as determined by DSC. In order to study the mechanical behaviour, tensile tests on the melt spun samples were carried out using a microtensile machine (Oxford Instruments). The tests were performed on ribbons with an initial length of 10 mm at a strain rate of 10−4 s−1 at room temperature. Fractographic analysis was carried out by scanning electron microscopy (SEM) in a JEOL JSM-840F, and Vickers microhardness was measured in a Leitz Miniload instrument with a load of 100 g. The microhardness was measured at room temperature on samples in both as-spun and heat treated states.

i: icosahedral phase

-Al-(Cr,Fe)

2. Experimental

(103)Al3Ti

Mn-1

40

42

2

44

Mn-1

46

48

50

Fig. 1 XRD diffractograms from the as melt-spun alloys.

3. Results 3.1 As melt-spun alloys All melt-spun samples were obtained with a thickness of approximately 30 µm and with the exception of the Ti-1 alloy, could be hand bent up to 180◦ without fracture, which implies good ductility. 3.1.1 XRD analysis XRD traces from the as melt-spun samples are shown in Fig. 1. The presence of the α-Al phase was confirmed in all four alloys. In the XRD traces from the Ti-1 and Ti-2 melt spun alloys, in addition to the α-Al phase were present small peaks at 2θ ∼ 22.9◦ , 41.3◦ , 43.5◦ , 61.6◦ and 73.7◦ , which could be indexed as an icosahedral phase (i-phase) using the indexation scheme for icosahedral quasiperiodic crystals proposed by Cahn et al.23) The Ti-2 melt spun alloy exhibited two additional small peaks which could be assigned to the tetrahedral t-Al3 Ti and distorted θ -Al–(Cr, Fe) phases.24, 25) The XRD traces from the V-1 melt spun alloy had a very broad, low intensity peak, corresponding to an amorphous phase, between the two main peaks of the α-Al phase. The α-

Structural Characterisation and Mechanical Properties of Nanocomposite Al-based Alloys

Al peaks for the V-1 melt spun alloy were broader than those for the other melt spun alloys, perhaps indicating a high level of crystal distortion caused by residual stresses. The peaks corresponding to the α-Al phase in the XRD traces from the Mn-1 melt spun alloy showed a small displacement towards the high angle side. In addition, two small peaks which could not be indexed were present. Several phases, such as icosahedral and approximant phases, have reflections in these positions.26, 27) 3.1.2 TEM observation Figure 2 shows bright field (BF) TEM images of each alloy in the as melt spun state. The Ti-1 melt spun alloy had a uniform distribution of small near-spherical quasicrystalline particles 10–20 nm in diameter embedded in an α-Al matrix with grains ∼ 0.5–2 µm in size (Fig. 2(a)). The Ti-2 melt spun alloy (Fig. 2(b)), had a lower volume fraction distribution of coarser near-spherical quasicrystalline particles 50–80 nm in diameter embedded in an α-Al matrix. Some non-spherical small particles were observed but could not be identified, possibly the intermetallic compounds detected by XRD (Fig. 1). The inset of Fig. 2(b) shows a nano beam electron diffraction pattern, taken from a near-spherical particle, that denotes a

2019

five-fold symmetry characteristic of an icosahedral phase, in agreement with the XRD profiles (Fig. 1). The V-1 melt spun alloy (Fig. 2(c)) had the finest scale microstructure. A large, dark-imaging α-Al grain (∼ 0.7– 1.5 µm) is shown with a high fraction of 5–10 nm embedded particles. In agreement with the XRD traces, corresponding selected area electron diffraction patterns showed a continuous diffraction halo corresponding to the nanogranular amorphous phase and diffraction spots corresponding to the α-Al matrix (see inset of Fig. 2(c)). The Mn-1 melt spun alloy (Fig. 2(d)) had the most conventional microstructure of ∼ 1–2 µm sized α-Al grains with 20–40 nm intermetallic particles decorating the grain boundaries and a low fraction of 10–20 nm precipitates within the α-Al grains. The presence of the i-phase could not be verified in this alloy by TEM observation. 3.2 Structural stability 3.2.1 DSC runs DSC traces from the melt spun Ti-2, V-1 and Mn-1 alloys are shown in Fig. 3. Each alloy studied showed two exothermic transitions corresponding to two transformation steps: the first one very sluggish and broad with relatively low heat release, and the second one strong and sharp with relatively high heat release. The weak peak of the first transition was centred around 595 K and 600 K for the Mn-1 and V-1 alloys respectively. The Ti-2 alloy showed a first transition centred at a higher temperature of 670–690 K. The second transition showed a much stronger heat release, and had a different shape for each alloy. The Ti-2 alloy showed a second transition apparently composed of only one peak at approximately 773 K, some 35–40 K higher than those of the other two alloys. The V-1 alloy showed a second transition composed of two sharp overlapping peaks at 729 K and 743 K, whereas the Mn-1 alloy showed a second transition with a broad peak at ∼ 739 K. 3.2.2 XRD analysis of heat treated alloys The temperatures of the heat treatments are indicated in Fig. 3 and were chosen in order to analyse the structure developed in each transformation step as determined by DSC. Figure 4 shows XRD traces from the Ti-2 alloy, as melt spun

V-1

EXO

DSC Run (40 K/min)

Heat Flow (a.u.)

Ti-2

Fig. 2 Bright field TEM images of the as melt-spun alloys: (a) Ti-1, (b) Ti-2, with the inset of a nano beam electron diffraction pattern from a near-spherical particle (beam diameter ∼ 15 nm), (c) V-1, with the inset of the selected area electron diffraction pattern, (d) Mn-1.

T1(Ti-2)

Mn-1 T1(V-1) T2

T1(Mn-1)

400

450

500

550 600 650 700 Temperature, T / K

750

800

Fig. 3 DSC traces from the as melt-spun alloys heated at 40 K/min.

F. Audebert et al. i, i-phase

i

Step-1 (Annealed at 703 K)

i

38

2

44

46

48

50

-Fe + -V

-Fe

-Fe -V + -Fe -V + -Fe

Step-2 (Annealed at 823 K)

-Fe -V

-V -V -V -Fe

Intensity (a.u.)

V-1

38

(200) Al

(111) Al

Step-1 (Annealed at 628 K)

36

40

42 2

44

-Cr -Cr

Step-1 (Annealed at 603 K)

As Melt-Spun

46

48

50

Fig. 5 X-ray diffractograms from the V-1 alloy, as melt-spun and after heat treatment at 683 K and 823 K.

and after heat treatment for 45 min at 703 K and 823 K. After heat treatment at 703 K, i.e. at the end of the first transformation step, the structure remained similar to the as melt spun alloy, still consisting of a mixture of i-phase, θ -Alx (Fe, Cr) y and t-Al3 Ti embedded in an α-Al matrix. The peak corresponding to (103)t-Al3 Ti at 2θ = 39.2◦ increased in intensity. Moreover, three new peaks were observed, two of which were close to peaks corresponding to the icosahedral phase. These peaks could be indexed as the distorted θ -Al13 (Cr, Fe)2–4 phase, which has been observed in Al-Fe-Cr alloys obtained by rapid quenching.24, 25, 28) This structure is a metastable state which forms prior to transformation to the stable θ -Al13 Cr2 and θ Al13 Fe4 phases. After higher temperature heat treatment at 823 K, α-Al, θ -Al13 Cr2 , θ -Al13 Fe4 and t-Al3 Ti phases were all identified by XRD. Figure 5 shows XRD traces from the V1 alloy, as melt spun and after heat treatment for 45 min at 628 K and 823 K. After heat treatment at 628 K, there was little difference compared with the melt spun XRD trace, both showing a mixture of α-Al and amorphous phases. Small shoulders on the (111) and (200) α-Al peaks after heat treatment at 628 K (see arrows in Fig. 5) indicated the nucleation of some additional crystalline phases. After higher temperature heat treatment at 823 K, α-Al, t-Al3 Ti, θ -Al13 Fe4 and θ -Al45 V7 phases were all

36

38

(200)Al

(111)Al

(200) Al

42

As Melt-Spun

Fig. 4 X-ray diffractograms from the Ti-2 alloy, as melt-spun and after heat treatment at 703 K and 823 K. Ti 3 -V: -Al45V7 -Fe: -Al13Fe4

Step-2 (Annealed at 823 K)

i

i

40

Mn-1

-Cr

-Fe

Step-2 (Annealed at 823 K)

Intensity (a.u.)

-Fe

-Cr

-Cr

-Fe, Al13Fe4

-Cr -Cr

-Cr

, -Al-(Fe,Cr) -Cr, Al 13Cr2

36

: Al11Cu5Mn3 -Cr: Al13Cr2

Ti-2

Ti

(111) Al

Intensity (a.u.)

3

-Cr -Cr

2020

40

42

2

44

As Melt-Spun

46

48

50

Fig. 6 X-ray diffractograms from the Mn-1 alloy, as melt-spun and after heat treatment at 603 K and 823 K.

identified by XRD. Figure 6 shows XRD traces from the Mn-1 alloy, as melt spun and after heat treatment for 45 min at 603 K and 823 K. The unidentified peaks present in the as melt spun alloy (see arrow in Fig. 6) disappeared after heat treatment at 603 K, confirming that the peaks corresponded to a metastable phase. In addition there was some evidence of the initial stages of formation of θ -Al13 Cr2 during the first exothermic transition as shown in Fig. 6. After higher temperature heat treatment at 823 K, α-Al, θ -Al13 Cr2 and O–Al11 Cu5 Mn3 phases were all identified by XRD. 3.2.3 TEM of heat treated alloys TEM observation of the microstructure of the Ti-2 alloy after heat treatment for 45 min at 703 K showed little change with respect to the melt spun microstructure, as can be seen by comparing Figs. 7(a) and 2(b). There was little evidence of the formation of intermetallic compounds, in agreement with the XRD traces in Fig. 4. The near-spherical quasicrystalline particles remained without apparent grain coarsening as shown in Fig. 7(a). After heat treatment at the higher temperature of 823 K a high volume fraction of intermetallic phases of different sizes was observed as shown in Fig. 8(a). TEM observation of the microstructure of the V-1 alloy after heat treatment for 45 min at 628 K again showed no change with respect to the as melt spun microstructure, as can be seen by comparing Figs. 7(b) and 2(c). A continuous diffraction halo was still present which confirmed the continuing presence of the amorphous phase as shown in the inset of Fig. 7(b). After heat treatment at the higher temperature of 823 K, the microstructure was similar to that of the Ti-2 alloy heat treated at the same temperature again consisting of a high volume fraction of intermetallic phases of different sizes, as can be seen in Figs. 8(a) and (b). TEM observation of the microstructure of the Mn-1 alloy after heat treatment for 45 min at 603 K showed significant particle coarsening, as can be seen by comparing Figs. 7(c) and 2(d). After heat treatment at the higher temperature of 823 K, the microstructure again showed high volume fraction of coarse intermetallic particles (Fig. 8(c)).

Structural Characterisation and Mechanical Properties of Nanocomposite Al-based Alloys

2021

Fig. 7 Bright field TEM images of the alloys after heat treatment beyond the first transformation step: (a) Ti-2 after 45 min at 703 K, (b) V-1 after 45 min at 628 K, (c) Mn-1 after 45 min at 603 K.

Fig. 8 Bright field TEM images of the alloys after heat treatment beyond the second transformation step, i.e. for 45 min at 823 K: (a) Ti-2, (b) V-1, (c) Mn-1.

3.3 Mechanical properties 3.3.1 Tensile testing of melt spun alloys Figure 9 shows typical examples of stress-strain curves obtaining from tensile tests of the four as melt spun alloys at room temperature. The Ti-1 alloy was very brittle and fractured at low tensile stress, in the range of 200–250 MPa. The Ti-2 alloy, with a lower volume fraction of the icosahedral phase than the Ti-1 alloy, had higher ultimate tensile strength and showed a small plastic deformation zone, with fracture stresses and total strains (elastic and plastic) typically of 510– 575 MPa and 3–3.5% respectively. The V-1 alloy had the highest strength but without showing macroscopic ductile behaviour. Conversely, the Mn-1 alloy had comparatively good ductility, with a parabolic behaviour similar to conventional ductile Al alloys, albeit with a low yield strength, with fracture stresses and strains typically of 210–300 MPa and 3.5– 4.5% respectively. 3.3.2 Fractography of melt-spun alloys The fracture surfaces of the as melt spun Ti-1 samples showed no evidence of plastic flow. They were flat with some

undulations as a consequence of catastrophic brittle fracture in a stressed material (Fig. 10(a)). On the other hand, the fracture surfaces of the as melt spun Mn-1 samples exhibited necking and a dimpled fracture surface characteristic of plastic deformation in crystalline Al alloys, as shown in Fig. 10(b). The fracture surfaces of the as melt spun Ti-2 samples showed the characteristic ribs of ductile fracture, as shown in Fig. 10(c) (see details in the inset of Figure) with slip bands close to the fracture surface. The fracture surfaces of the as melt spun V-1 samples, that had not shown macroscopic plastic deformation, again showed ribs and vein patterns characteristic of a ductile fracture (Fig. 10(d)) similar to those found in ductile amorphous or partially amorphous alloys.29, 30) Furthermore typical shear bands, as are observed in glassy metals,30) were found near the fracture surface. 3.3.3 Microhardness as a function of heat treatment Figure 11 shows microhardness results for Ti-2, V-1 and Mn-1 alloys as melt spun and after heat treatment at the different temperatures corresponding to the end of each transformation step. The Ti-2 and V-1 alloys had similar microhard-

2022

F. Audebert et al. 700

Slip Band Process

/ MPa

V-1

/ MPa Stress,

Ti-2 500

400

Crack Propagation

Stress,

600

0.031

Ti-2

0.032

Strain

0.033

Mn-1 300

Ti-1 200

100

0 0.00

0.02

0.04

0.06

0.08

Strain Fig. 9 Room temperature stress-strain tensile curves from the as melt-spun alloys. The inset shows an enlarged view of the region of plasticity in the curve for the Ti-2 alloy.

ness values in the as melt-spun state and again after heat treatment at high temperature (823 K). The microhardness values of the Ti-2 and Mn-1 alloys decreased continuously, whereas the microhardness of the V-1 alloy increased initially, after heat treatment at intermediate temperature (628 K), before decreasing, after heat treatment at higher temperature (823 K). The Mn-1 alloy had the lowest microhardness values and decreased faster with the increasing heat treatment temperature in comparison with the other alloys. 4. Discussion 4.1 Structure and stability All samples had the α-Al phase as a matrix, although the alloys of the different systems showed different microstructures. The Ti-1 and Ti-2 alloys had finer microstructures than observed previously for a melt spun sample with a composition close to the icosahedral phase,20) and for the same composition produced in bulk by warm extrusion of atomised powders.31) The XRD profiles of both Ti-1 and Ti-2 showed only the P-type i-phase peaks, and no other peaks related to the F-type ordered i-phase found previously by Inoue et al.20) in similar alloys containing Fe. The weakness of the (18/29) and (20/32) i-phase peaks may be related to the presence of a disordered P-type i-phase, and reflections from the distorted planes of approximant phases, such as monoclinic θ Al–(Cr, Fe).24, 25, 28) The two other small peaks in XRD traces of the Ti-2 alloy could be assigned to t-Al3 Ti and θ -Al– (Cr, Fe). The sample heat treated at 703 K showed clearer peaks from θ -Al–(Cr, Fe), and after heat treatment at 823 K all the equilibrium phases were formed following decomposition of the i-phase. The i-phase at higher alloy content than Ti1 or Ti-2, i.e. with a composition close to Al84.2 Fe7 Cr6.3 Ti2.5 (at%), has been shown previously to transform via two DSC peaks at ∼ 850 K and ∼ 925 K.19) The Ti-2 sample, of the

present study has a lower alloy content, and also shows two peaks, although at lower temperatures (Fig. 3). The difference in transformation temperatures results from the lower alloy content and the P-type disordered structure of the i-phase. XRD (Fig. 1) and TEM (Fig. 2(c)) show that as melt-spun V-1 has a microstructure composed of a nanogranular amorphous phase embedded in an α-Al matrix. The formation of this kind of microstructure is related to glass formation in icosahedral alloys.32–34) These alloys have, in the liquid, an icosahedral short range order around the transition metal atoms which provides a barrier to the formation of crystalline phases.33, 35) Therefore, an icosahedral phase can form composed of interpenetrating clusters of icosahedral symmetry.36) When the transition metal in the i-cluster has low atomic diffusivity in the Al liquid (such as Ti and V), quasicrystalline formation becomes difficult and a nanogranular amorphous region with icosahedral short range order can be retained in the α-Al matrix. It has been reported34, 37) that a decrease in cooling rate allows the formation of the quasicrystalline phase. The subsequent annealing of the V-1 alloy shows a first heat treatment step where no microstructural changes are observed (Figs. 5 and 7(b)), perhaps caused by structural relaxation. On the other hand, the second transformation step is characterized by two exothermic DSC peaks around 729 K and 743 K (Fig. 3). Similar behaviour has been observed for the ternary nanogranular amorphous alloys Al93 Fe3 Ti4 and Al94 Fe2 V4 32, 38) although at lower temperature, which confirms the higher stability of the quaternary alloy. After the second heat treatment step, only the crystalline stable phases are observed (Fig. 5). The as melt spun Mn-1 alloy has a conventional microstructure with fine particles in the grain boundaries and inside the α-Al grains. The presence of the i-phase could not be confirmed. Haas et al.17) have reported the presence of the i-phase for a sample with a similar composition produced by warm extrusion. That means that the formation of the i-phase in this alloy is very sensitive to the chemical composition and the kinetic conditions. The lower alloy content of the Mn-1 alloy with respect to Ti-1, Ti-2 and V-1, and the higher solubility of Cu in α-Al (3 to 8 times higher than Fe, Cr, Mn and Ti) may hinder the formation of the i-phase. In agreement with this hypothesis, there is a small displacement in the positions of the α-Al peaks observed in the XRD profiles (Fig. 1), indicating that the α-Al matrix has a higher solute content. The first heat treatment step (Fig. 2) is characterised by intermetallic coarsening (Fig. 7(c)) and decomposition of the unindexed metastable phase (Fig. 6). The second DSC step, with a broad exothermic peak at ∼ 740 K is associated with the formation of coarse intermetallic compounds (Fig. 8(c)). The decomposition of the quasicrystalline particles was reported by Haas et al.,17) to occur at ∼ 700 K at 10 K/min. Considering the peak displacement with the DSC scan rate,39) that means that the stability of the metastable alloy is reached at a similar temperature for the decomposition of either the metastable crystalline or quasicrystalline precursor phase. 4.2 Mechanical properties 4.2.1 Al–Fe–Cr–Ti system The tensile stress-strain curves in Fig. 9 and the SEM fractographs in Fig. 10, show that the volume fraction of

Structural Characterisation and Mechanical Properties of Nanocomposite Al-based Alloys

2023

Fig. 10 Secondary electron SEM image of the fracture surfaces of the tensile tested as melt-spun alloys. (a) Ti-1, (b) Mn-1, (c) slip steps in Ti-2, with the inset showing ribs on the fracture surface, and (d) V-1, with the inset showing an enlarged view of a region with a vein pattern.

500

HV100

V-1 400

Ti-2 300

Mn-1 200

100

Microhardness at Room Temperature

300

400

500

600

700

800

Heat Treatment Temperature, T / K Fig. 11 Room temperature microhardness of the Ti-2, V-1 and Mn-1 alloys as melt-spun and after heat treatments.

the icosahedral phase has a great influence on mechanical properties in the Al–Fe–Cr–Ti system (Ti-1 and Ti-2). The less highly alloyed Ti-2 material shows a tensile strength (∼ 550 MPa) similar to that reported by Inoue et al.19) on a bulk warm extruded sample, although the failure strain (∼ 3.5%) is slightly lower. The volume fraction of particles can be estimated, assuming that particles are arranged like points of a fcc lattice, from the relationship: f = rp2 /(λ/2)2 where rp is the mean planar particle radius and λ is the mean distance between the centres of the planar particles.40) From TEM observation the average values are 2rp  ∼ 65 nm and λ ∼ 100 nm for the Ti-2 alloy, giving f ∼ 0.42, similar to the value of 0.45 reported by Inoue et al.19) Similar calculation for the more highly alloyed Ti-1 material give f ∼ 0.46. The importance of the volume fraction of the second phase in nanocomposites has been demonstrated by Scanlon et al.41) They observed in nanoscale mixtures of a pure plastic metal and hard oxide nanoparticles, that the microhardness has a different relationship with respect to the volume fraction after the percolation threshold of ∼ 0.5. A similar concept has been proposed to explain the ductile-brittle transition as a function of chemical composition in heterogeneous amorphous metals42) and nanocomposites containing hard nanocrystals.11) The high volume fraction Ti-1 alloy has the highest microhardaness value (µHV100 ∼ 450 kg/mm2 ) and shows completely brittle behaviour (Fig. 9) without traces of microplas-

ticity or evidence of second phase particles on the fracture surface (Fig. 10). This mode of fracture results presumably from the presence of pre-existing internal or surface flaws or cracks, so that crack nucleation is not necessary. The stress is concentrated progressively ahead of the pre-existing flaws or cracks, crack propagation takes place throughout the matrix between the high volume fraction of hard quasicrystalline particles as soon as it is accompanied by reduction in total system energy,43) leading to brittle fracture. Neither sheared nor cleaved quasicrystalline particles have been observed on fracture surfaces in other nanoquasicrystalline Albased alloys.17, 18) The low volume fraction Ti-2 alloy, shows significant ductility (Fig. 9), and shear flow associated with heterogeneous microscopic plasticity can be seen clearly on the fracture surface (inset of Fig. 10(c)). The lower volume fraction of second phase particles allows plastic deformation and prevents catastrophic brittle failure at pre-existing flaws or cracks. Fracture is initiated at higher stress level than the more brittle Ti-1 alloy, when the applied stress is sufficient to activate multiple slip or twinning leading to formation of microcracks.43) The enlarged view of the high stress region of plasticity in the stress-strain curve of the Ti-2 alloy (Fig. 9), shows serrated yielding at constant stress, as was observed previously in rapid solidification processed Al–Li alloys.44) This yield stress is that required to nucleate and grow slip bands which propagate until specimen failure by crack propagation throughout the Al matrix, as can be deduced by the rib marks with small grains in the fracture surface (inset of Fig. 10(c)). Two different regions can be observed in the fracture surface (inset of Fig. 10(c)) that can be explained as follows: (i) The surface of the ribbon is unable to support a stress, deforms under plane stress conditions, and produces shear lips. (ii) The central portion of the ribbon is subjected to plane strain conditions and the cracks advance faster as a result of a smaller plastic zone size and a triaxial stress state. This results in a reduced fracture toughness but with a ductile fracture.43) Heat treatment for 45 min at 703 K, i.e. after the first transformation step, on the Ti-2 alloy produces ∼ 40% reduction in the microhardness from ∼ 396 to ∼ 240 kg/mm2 . This softening is related to the relaxation of residual stresses in the α-Al matrix, and the reduction of solute content in the matrix, as intermetallic particles are formed. The decomposition of the icosahedral phase and the coarsening of the intermetallic particles reduces the microhardness value still further, by ∼ 28% to ∼ 173 kg/mm2 (Fig. 11).

2024

F. Audebert et al.

4.2.2 Al–Fe–Cr–V system The V-1 alloy has around 0.5 volume fraction of the nanogranular amorphous phase, which acts as a strengthening agent. The homogeneous distribution of a high volume fraction of very fine hard particles is a well known alloy strengthening mechanism.43) In agreement, V-1 alloy shows a very high ultimate tensile strength, σUTS ∼ 650 MPa (Fig. 9). This value is higher than that reported by Kimura et al.20) for an alloy with the same composition produced by warm extrusion of atomised powders (σUTS ∼ 500 MPa), with a microstructure of 120 nm sized icosahedral particles distributed in an αAl matrix. The larger icosahedral particle size (120 nm compared with the 5 to 10 nm sized amorphous regions) may explain the difference in σUTS values between the extruded and the present melt-spun alloys. Several σUTS values have been reported for Al-based alloys with a microstructure similar to the V-1 alloy: 1000 MPa for Al94 Fe2 V2 Ti2 ,45) 1350 MPa for Al94 Fe2 V4 ,21) 1320 MPa for Al93 Fe3 Ti4 32) and 760 MPa for Al92 Fe2 V4 Si2 .34) The latter alloy has the same Al concentration as the V-1 alloy, and has a σUTS value closer to that of the V-1 alloy than the other alloys. The similarity in behaviour between Al92 Fe2 V4 Si2 and V-1 may be related to a similar volume fraction of hard particles with similar macroscopically brittle failure mode (Fig. 9). Fractographic investigation shows that fracture in V-1 is a ductile process, resulting in a fracture surface with rib marks and vein patterns (Fig. 10(d)) characteristic of adiabatic fractures in fully or partially amorphous metals.8, 29) A vein pattern on the fracture surface of nanogranular amorphous Albased alloys has been observed previously by Inoue et al.38) The V-1 alloy also has an acceptable ductility in bending tests, but nanocomposites obtained by crystallisation with a similar volume fraction of α-Al nanograins in the amorphous matrix are often completely brittle.10, 11) This indicates the importance of the mechanical properties of the matrix on the mechanical behaviour of such nanocomposites. Structural relaxation in the amorphous phase and coherent precipitation in the α-Al phase, may be responsible for the increase in the microhardness from ∼ 395 to ∼ 514 kg/mm2 after heat treatment for 45 min at 628 K (Fig. 11). After heat treatment for 45 min at 823 K, transformation of the fine nanoscale microstructure to a coarser distribution of submicron intermetallic compounds precipitated in the α-Al matrix, drastically reduces the microhardness to a value of ∼ 165 kg/mm2 . 4.2.3 Al–Mn–Cr–Cu system The Mn-1 alloy was the most ductile of the four alloys studied, showing significant macroscopic plastic deformation (Fig. 9), in agreement with a microstructure containing the lowest volume fraction of strengthening particles (Fig. 2). Plastic deformation takes place by conventional dislocation movement in the α-Al matrix grains. Dislocations pile up around the strengthening particles increasing the local normal stresses leads to particle fracture and microvoid nucleation at the particle interface.43) The voids begin to grow or elongate in the direction of the applied tensile stress and the final fracture stage is eventually reached when the voids link up by rupture of the α-Al matrix. The ultimate tensile strength σUTS and fracture strain ε are ∼ 300 MPa and ∼ 4.5% respectively, lower than those

(σUTS ∼ 570–460 MPa, ε ∼ 10.3%) reported by Haas et al.17) in a warm extruded powder alloy of similar composition. The difference can be explained by the different microstructures obtained by the different processing routes. Haas et al.17) reported a microstructure composed of icosahedral particles embedded in an α-Al matrix, whereas the Mn-1 alloy, produced by liquid quenching, exhibits a fine scale distribution of particles at the α-Al grain boundaries which can decrease the toughness and hence the strength and the ductility.46) After heat treatment for 45 min at 603 K, the microhardness of the Mn-1 alloy decreases by about 60%, from ∼ 340 to ∼ 124 kg/mm2 , which could be related to the formation of additional incoherent precipitates in the α-Al matrix and coarsening of the particles (Fig. 7(c)). After the second stage of heat treatment for 45 min at 823 K, the Mn-1 alloy shows a microstructure of rod-shaped submicron intermetallic compounds in an α-Al matrix, and the microhardness decreases further, by ∼ 33% to ∼ 78 kg/mm2 . 5. Conclusion Four Al-based nanocomposite alloys with different kinds of microstructure composed of crystalline, amorphous or quasicrystalline phases dispersed in an α-Al matrix have been studied. The Ti-2 alloy contained icosahedral particles in an αAl matrix, and exhibited a high strength and a very stable microstructure during heat treatment, with the icosahedral particles resistant to coarsening prior to decomposition into crystalline phases. The V-1 alloy contained a nanogranular amorphous phase in an α-Al matrix, and exhibited a highest strength with a moderately stable microstructure up to 628 K, but relaxation in the amorphous structure made the alloy harder and more brittle. The Mn-1 alloy contained strengthening crystalline intermetallic particles in the α-Al matrix, and exhibited a high ductility, with a significant decrease in strength during heat treatment caused by incoherent precipitation and particle coarsening. A combination of classical strengthening mechanisms (solid solution, grain refinement and dispersion particles) was observed to be responsible for the high strength of the different kinds of nanocomposites studied. Acknowledgments The work described forms part of a European Union funded Research Training Network Project on Manufacture and Characterisation of Nanostructured Al Alloys, contract no. HPRN-CT-2000-00038. F. Audebert acknowledges the financial support of the Universidad de Buenos Aires, Argentina. REFERENCES 1) A. Inoue, K. Othera, A. Tsai and T. Masumoto: Jpn. J. Appl. Phys. 27 (1988) L280–L282. 2) G. Shiflet, Y. He and S. Poon: Scr. Metall. 22 (1988) 1661–1664. 3) A. Inoue and H. Kimura: Current Opinion in Solid State & Mater. Sci. 2 (1997) 305–310. 4) A. Inoue and H. Kimura: J. Light Metals 1 (2001) 31–41. 5) J. Q. Guo and K. Othera: Acta Mater. 46 (1998) 3829–3838.

Structural Characterisation and Mechanical Properties of Nanocomposite Al-based Alloys 6) K. Othera, A. Inoue, T. Terbayashi, H. Nagahama and T. Masumoto: Mater. Trans., JIM 33 (1992) 775–781. 7) T. Kobayashi: Aerospace Materials, (Series in Materials Science and Engineering, ed. by B. Cantor, H. Assender and P. Grant, Institute of Physics, UK, 2001) pp. 170–178. 8) A. Inoue, Y. Horio, Y. H. Kim and T. Masumoto: Mater. Trans., JIM 33 (1992) 669–674. 9) A. A. Csontos and G. J. Shiflet: Nano Structured Mater. 9 (1997) 281– 289. 10) Z. C. Zhong, X. Y. Jiang and A. L. Greer: Mater. Sci. Eng. A 226–228 (1997) 531–535. 11) H. S. Kim, P. J. Warren, B. Cantor and H. R. Lee: Nano Structured Mater. 11 (1999) 241–247. 12) A. Inoue, M. Watanabe, H. M. Kimura, F. Takahashi, A. Nagata and T. Masumoto: Mater. Trans., JIM 33 (1992) 723–729. 13) F. Schurack, J. Eckert and L. Schultz: Nano Structured Mater. 12 (1999) 107–110. 14) A. Inoue: Nano Structured Mater. 6 (1995) 53–64. 15) E. H. Buchler, E. Watanabe and N. S. Kazama: Int. J. Non-Equil. Processing 10 (1997) 35–47. 16) A. Inoue, H. M. Kimura, K. Sasamori and T. Masumoto: Mater. Trans., JIM 36 (1995) 6–15. 17) V. Haas, M. Cho, H. Ishil and A. Inoue: Nano Structured Mater. 12 (1999) 829–834. 18) F. Schurack, J. Eckert and L. Schultz: Acta Mater. 49 (2001) 1351– 1361. 19) A. Inoue and H. Kimura: Nano Structured 11 (1999) 221–231. 20) H. M. Kimura, K. Sasamori and A. Inoue: J. Mater. Res. 15 (2000) 2737–2744. 21) A. Inoue, H. M. Kimura, K. Sasamori and T. Masumoto: Mater. Sci. Eng. A 217–218 (1996) 401–406. 22) E. S. Humphreys, P. J. Warren, A. Cerezo and G. D. W. Smith: Mater. Sci. Eng. A 270 (1999) 48–54. 23) J. W. Cahn, D. Shechtman and D. Gratias: J. Mater. Res. 1 (1986) 13–26. 24) R. Manaila, V. Florescu, A. Jianu and O. Radulescu: Philos. Mag. B 60 (1989) 589–599. 25) F. Audebert, R. Colaco, R. Vilar and H. Sirkin: Scr. Mater. 40 (1999) 551–557. 26) K. Kobayashi, N. Tachibana and P. H. Shingu: J. Mater. Sci. 24 (1989)

2025

2437–2443. 27) D. Schectman and I. A. Blech: Metall. Trans. A 16A (1985) 1005–1012. 28) J. Copola, F. Audebert and S. Duhalde: J. Metastable and Nanocryst. Mater. 14 (2002) 27–32. 29) C. Pampillo: J. Mater. Sci. 10 (1975) 1194–1227. 30) G. J. Shiflet, Y. He and S. J. Poon: Scr. Metall. 22 (1988) 1661–1664. 31) H. M. Kimura, K. Sasamori and A. Inoue: Mater. Sci. Eng. 294–296 (2000) 168–172. 32) H. M. Kimura, K. Sasamori, T. Negishi and A. Inoue: Nano Structured Mater. 8 (1997) 833–844. 33) F. Audebert, B. Arcondo, D. Rodriguez and H. Sirkin: J. Metastable and Nanocryst. Mater. 10 (2001) 155–160. 34) E. S. Humphreys, P. J. Warren, J. M. Titchmarsh and A. Cerezo: Mater. Sci. Eng. A 304–306 (2001) 844–848. 35) V. Simonet, F. Hippert, M. Audier and R. Bellissent: Mater. Sci. Eng. 294–296 (2000) 116–119. 36) M. Boudard, M. De Boissieu, C. Janot, G. Heger, C. Beeli, H. Nissen, H. Vincent, R. Ibberson, M. Audier and J. M. Dubois: J. Phys.: Condens. Matter. 4 (1992) 10149–10168. 37) A. Inoue, H. M. Kimura, K. Sasamori and T. Masumoto: Mater. Trans., JIM 36 (1995) 1219–1228. 38) A. Inoue, H. M. Kimura, K. Sasamori and T. Masumoto: Nano Structured Mater. 7 (1996) 863–382. 39) L. C. Chen and F. Spaepen: J. Appl. Phys. 69 (1991) 679–687. 40) M. Ashby: Z. Metallk. 55 (1964) 5–17. 41) M. R. Scanlon, R. C. Cammarata: J. Appl. Phys. 76 (1994) 3387–3393. 42) F. Audebert, A. Garcia Escorial, J. Reyes Gasga and H. Sirkin: Proc. of First Journe´es Franco-Espagnoles des Noveaux Mat´eriaux, ed. by R. Yavari and J. L. Uriarte (Grenoble, France, 1997) 138–142. 43) T. H. Courtney: Mechanical Behavior of Materials, (McGraw-Hill, Series in Materials Science and Engineering, New York, 1990). 44) L. Zhen, D. Yang, S. Sun and G. Yu: Mater. Sci. Eng. A 248 (1998) 221–229. 45) A. Inoue, H. Kimura and K. Sasamori: Nano Structured Mater. 9 (1997) 493–496. 46) R. W. K. Honeycombe: The Plastic Deformation of Metals, (ed. by Edward Arnod pub., London, 1984).