Phase transformations in a Cu-142Al-15.0Ni alloy

(1)

Phase transformations in a Cu–14.2Al–15.0Ni alloy

C.H. Chen

∗

, T.F. Liu

Department of Materials Science and Engineering, National Chiao Tung University, Hsinchu, Taiwan 300, ROC Received 21 May 2002; received in revised form 7 June 2002; accepted 14 June 2002

Abstract

In the as-quenched condition, the microstructure of the Cu–14.2Al–15.0Ni alloy was D03phase containing extremely fine L–J precipitates. Since both fine D03domains with a/21 0 0 anti-phase boundaries (APBs) and small B2 domains with a/41 1 1 APBs could be observed, the D03phase existing in the as-quenched alloy should be formed by an A2→ B2 → D03continuous ordering transition during quenching. It is worthwhile to note that the a/41 1 1 APBs have never been observed by other workers in the Cu–Al–Ni ternary alloys before.

When the as-quenched alloy was aged at temperatures ranging from 400 to 1000◦C for longer times, the phase transformation sequence as the aging temperature increased was found to be(␣+ B2 precipitate) → (B2 + B2 precipitate) → B2 → A2. This transformation has also never been observed by other workers in the Cu–Al–Ni ternary alloys before.

Keywords: Cu–Al–Ni alloy; Phase transformation; Continuous ordering transition; Anti-phase boundaries; TEM

1. Introduction

Phase transformations in the Cu–Al–Ni ternary alloys

have been studied by many workers[1–15]. On the basis of

their studies, it is found that when an alloy with a chemical composition in the range of Cu–(12.8–15.1)Al–(3.0–7.7)Ni

was solution heat-treated in the single ␤ phase (A2,

dis-ordered body-centered cubic) region and then quenched into room temperature water or iced-brine rapidly, the

mi-crostructure was single D03 phase [1–4], or D03 phase

containing extremely fine 2H-type precipitates[5–7]. After

being aged at temperatures ranging from 325 to 550◦C for

moderate times and then quenched,␥2 particles started to

occur within the D03 matrix at the aging temperature and

the remaining D03matrix would transform to␥₁martensite

during quenching[3,4,7]. When the as-quenched alloy was

aged at this temperature range for longer times, fine B2 precipitates were observed to appear within the well-grown

␥2 particles and the remaining D03 matrix would

com-pletely transform to␣ phase (A1, disordered face-centered

cubic) or a mixture of (␣ + ␤) phases at the aging

tem-perature [7,9]. When the as-quenched alloy was aged in

the range from 600 to 770◦C for longer times, the stable

microstructure was found to be(␤ + ␥2) phases[7,9,10].

∗_{Corresponding author.}

E-mail address: [email protected] (C.H. Chen).

In the previous studies [1–14], it is clearly seen that

al-though the phase transformations in the Cu–Al–Ni alloys have been extensively studied, most of the examinations were focused on the Cu–Al–Ni alloys with lower nickel content. Information concerning the microstructural devel-opment of the Cu–Al–Ni alloys with higher nickel con-tent is very deficient. Therefore, the present investigation is an attempt to clarify the phase transformations in the Cu–14.2Al–15.0Ni alloy.

2. Experimental procedure

The alloy, Cu–14.2Al–15.0Ni, was prepared in a vac-uum induction furnace by using 99.9% copper, 99.9% aluminum and 99.9% nickel. The melt was chill cast into

a 30 mm × 50 mm × 200 mm copper mold. After being

homogenized at 1000◦C for 72 h, the ingot was sectioned

into 2.0 mm-thick slices. These slices were subsequently

solution heat-treated at 1100◦C for 1 h and then quenched

into room temperature water. The aging processes were

performed at temperatures ranging from 400 to 1000◦C in

a vacuum heat-treated furnace for various times and then quenched into room temperature water rapidly.

Electron microscopy specimens were prepared by means of a double-jet electropolisher with an electrolyte of 60% methanol and 40% nitric acid. The polishing temperature

was kept in the range from−30 to −10◦C, and the current

(2)

Fig. 1. Electron micrographs of the as-quenched alloy: (a) BF; (b)–(e) four SADPs. The zone axes of the D03 phase are (b) [1 0 0], (c) [1 1 0], (d) [3 3 1] and (e) [1 1 1], respectively (h k l = D03 phase,h k l1or2= L–J phase, 1: variant 1; 2: variant 2); (f) (0 2 01) L–J DF; (g) and (h) (0 0 2) and(¯1 1 1) D03 DF, respectively.

(3)

Fig. 1. (Continued ).

(SADPs) of the as-quenched alloy. When compared with our

previous studies in the Cu–14.6Al–4.3Ni and Cu2.2Mn0.8Al

alloys[15,16], it is found in these SADPs that the brighter

and well-arranged reflection spots are of the ordered D03

phase, and the extra spots with streaks are derived from the

L–J precipitates with two variants.Fig. 1(f)is a (0 2 01) L–J

dark-field (DF) electron micrograph, clearly revealing the presence of the extremely fine L–J precipitates. Accordingly,

the microstructure of the as-quenched alloy was D03phase

containing extremely fine L–J precipitates. This is similar to that found by the present workers in the Cu–14.6Al–4.3Ni

alloy[15].Fig. 1(g), a (0 0 2) D03 DF electron micrograph

of the same area asFig. 1(a), shows the presence of the small

B2 domains with a/41 1 1 anti-phase boundaries (APBs).

Fig. 1(h)is a(¯1 1 1) D03DF electron micrograph,

indicat-ing the presence of the extremely fine D03 domains with

a/21 0 0 APBs. InFig. 1(g) and (h), it is also seen that the

sizes of both B2 and D03domains are very small. Therefore,

it is confirmed that the D03phase existing in the as-quenched

alloy was formed by an A2→ B2 → D03continuous

or-dering transition during quenching[17–21].

When the as-quenched alloy was aged at 400◦C for

mod-erate times and then quenched, both the D03and L–J phases

disappeared and other two kinds of phases started to occur.

A typical example is shown inFig. 2(a).Fig. 2(b)–(d), three

SADPs, indicate that the microstructure present inFig. 2(a)

is a mixture of B2 phase and ␥₁ martensite with internal

twins [3,13], and the orientation relationship between the

B2 phase and the ␥₁ martensite is [0 0 1]B2//[1 0 ¯1]_␥

1 and (1 ¯1 0)B2//(1 2 1)_␥

1.Fig. 2(e) and (f)show the [1 ¯2 1]␥1 and

(1 0 0) B2 DF electron micrographs of the same area as Fig. 2(a), clearly revealing the presence of the ␥₁ marten-site and the extremely fine B2 precipitates, respectively. Transmission electron microscopy (TEM) examinations

in-dicated that when the as-quenched alloy was aged at 400◦C

for longer times, the size of the B2 precipitates increased with increasing the aging time and the remaining matrix

had gradually decomposed into the disordered␣ phase, as

illustrated inFig. 3.Fig. 3(a)is a BF electron micrograph

of the alloy aged at 400◦C for 24 h and then quenched.

Fig. 3(b) and (c), two SADPs taken from an area covering

the B2 precipitates and their surrounding␣ matrix, indicate

that the orientation relationship between the B2 precipitate and the␣ matrix is [1 1 ¯1]B2//[1 0 ¯1]␣and (0 1 1)B2//(1 1 1)␣, which corresponds to the Kurdjumov–Sachs (K–S)

orien-tation relationship. Fig. 3(d) is a (0 0 1) B2 DF electron

micrograph, showing the presence of the well-grown B2 precipitates. It is thus concluded that the stable

microstruc-ture of the alloy at 400◦C should be a mixture of (␣ + B2

precipitate).

TEM observations of thin foils indicated that the mixture of (␣ + B2 precipitate) could be preserved up to 650◦C.

A typical example is shown in Fig. 4(a), which is a BF

electron micrograph of the alloy aged 650◦C for 24 h.

Fig. 4(b), a (0 0 1) B2 DF electron micrograph, reveals that the B2 precipitates were much larger than those observed in

the alloy aged at 400◦C.Fig. 4(c)is an SADP taken from

an area covering a B2 precipitate and its surrounding ␣

matrix, indicating that the orientation relationship between

the B2 precipitate and the ␣ matrix is [1 1 ¯1]B2//[1 0 ¯1]␣

and (0 1 1)B2//(1 1 1)␣, which is also of the K–S orientation

relationship.

Fig. 5(a)shows a BF electron micrograph of the alloy aged

at 700◦C for 2 h and then quenched, revealing that the B2

precipitates could be observed within the matrix. Shown in Fig. 5(b)is an SADP taken from the matrix, indicating that

the microstructure was the mixture of(D03+ L–J) phases.

Fig. 5(c) and (d)are (0 2 01) L–J and(¯1 1 1) D03 DF

elec-tron micrographs of the same area as Fig. 5(a), exhibiting

the presence of the L–J precipitates and D03domains,

re-spectively. It is clear in Fig. 5(c) and (d)that the sizes of

both the L–J precipitates and the D03domains are very

ex-tremely fine. It is therefore reasonable to believe that these two phases were formed during quenching from the aging temperature; otherwise, their sizes should be increased at

(4)

Fig. 2. Electron micrographs of the alloy aged at 400◦C for 10 min: (a) BF; (b)–(d) three SADPs. The zone axes of the B2 precipitate are (b) [0 0 1], (c) [0 1 1] and (d) [1 1 1], respectively (h k l = B2 precipitate, h k l = ␥₁martensite,h k lT= internal twin); (e) and (f) (1 ¯2 1)␥1and (1 0 0) B2 DF, respectively.

electron micrograph, shows that the B2 domains have grown

to reach the whole grain and no evidence of the a/41 1 1

APBs could be detected. This indicates that the microstruc-ture of the matrix was the B2 phase. Apparently, the stable

microstructure of the alloy present at 700◦C was the B2

phase containing B2 precipitates. It is worthwhile to note that the coexistence of two kinds of ordered B2 phase has not

previously been observed by other workers in the Cu–Al–Ni alloys before.

Shown inFig. 6(a)is a (0 0 2) D03DF electron micrograph

of the alloy aged at 750◦C for 1 h and then quenched. It

reveals that along with the growth of the B2 domains, the

a/41 1 1 APBs had gradually disappeared. Fig. 6(b) and (c)(¯1 1 1) D03and (0 2 01) L–J DF electron micrographs of

(5)

Fig. 3. Electron micrographs of the alloy aged at 400◦C for 24 h: (a) BF; (b) and (c) two SADPs, taken from an area covering the B2 precipitates and their surrounding a matrix. The zone axes of the B2 precipitate are (b) [1 1 0] and (c) [1 1 ¯1], respectively (h k l = B2 precipitate, h k l = ␣ phase); (d) (0 0 1) B2 DF.

the same area asFig. 6(a), show that only quenched-in D03

domains and L–J precipitates (the sizes being comparable to those observed in the as-quenched alloy) could be observed. This indicates that the microstructure of the alloy present at

750◦C was the single B2 phase.

Progressively higher temperature aging and quenching ex-periments indicated that the single B2 phase was preserved

up to 975◦C. An example is shown inFig. 7(a). However,

in an alloy aged at 1000◦C for 1 h and then quenched,

only quenched-in B2 domains could be detected, as

illus-trated in Fig. 7(b). This indicates that the microstructure

existing at 1000◦C or above was a single disordered ␤

phase.

4. Discussion

The presence of the a/41 1 1 APBs is a remarkable

feature in the present study. In the previous studies of the

Cu–(12.8–15.1)Al–(3.0–7.7)Ni alloys[1–15], it is clear that

in the as-quenched condition, no a/41 1 1 APBs could be

observed. For the absence of the a/41 1 1 APBs, several

workers proposed that although the D03phase was formed

by an A2 → B2 → D03transition during quenching from

the single␤ (A2) phase region, the a/41 1 1 APB energy in

the Cu–Al–Ni and Cu–Al–Mn alloys was very high, which would lead the B2 domains to grow up to the whole grains

during quenching[4,23,24]. Therefore, no a/41 1 1 APBs

could be observed. In contrast to the above proposition, other

workers claimed that the D03phase was occurred–an A2→

D03transition, rather than the A2→ B2 → D03transition

[1,3,7]. The A2→ B2 transition produced a/41 1 1 APBs

and the B2 → D03 transition produced a/21 0 0 APBs

[17,18]. Therefore, no a/41 1 1 APBs were formed. In the

present study, it is obvious that the a/41 1 1 APBs indeed

could be detected. This result strongly confirms that the D03

phase existing in the as-quenched alloy should be formed

through an A2→ B2 → D03 transition during quenching.

Furthermore, in addition to contain higher nickel content, the chemical composition of the present alloy is similar to that investigated by other workers in the Cu–Al–Ni alloys [1–11]. It seems to imply that the increase of the nickel

(6)

Fig. 4. Electron micrographs of the alloy aged at 650◦C for 24 h: (a) BF; (b) (0 0 1) B2 DF; (c) an SADP taken from an area covering the B2 precipitate and its surrounding␣ matrix. The zone axes of the B2 precipitate and ␣ phase are [1 1 ¯1] and [1 0 ¯1], respectively (h k l = B2 precipitate, h k l = ␣ phase).

content could decrease the a/41 1 1 APB energy; therefore,

the a/41 1 1 APBs became visible.

In the as-quenched condition, the microstructure of the

present alloy was D03phase containing extremely fine L–J

precipitates and no ␥₁ martensite could be observed. This

indicates that in the alloy with the chemical composition of

Cu–14.2Al–15.0Ni, the D03 → ␥₁martensitic

transforma-tion temperature should be below room temperature.

How-ever, when the present alloy was aged at 400◦C for moderate

times and then quenched, the microstructure was the mixture of (␥₁+B2 precipitate). This feature has never been observed

Table 1

Chemical compositions of the phases revealed by EDS

Heat treatment Phase Chemical composition (wt.%)

Cu Al Ni

As-quenched D03+ L–J 70.77± 0.72 14.19± 0.35 15.04± 0.48

400◦C for 10 min B2 precipitate 56.06± 1.56 18.67± 1.05 25.27± 1.11

␥

1 Martensite (remaining D03phase) 83.71± 1.02 9.82± 0.98 6.47± 1.05

400◦C for 24 h B2 precipitate 12.29± 0.67 25.91± 0.51 61.80± 0.56

␣ Phase 91.71± 1.02 6.40± 0.36 1.89± 0.32

by other workers in the Cu–Al–Ni alloys before. In order to clarify this feature, an STEM–EDS study was performed. Fig. 8(a)–(c)represent three typical EDS spectra taken from

the as-quenched alloy and a B2 precipitate as well as the␥₁

martensite in the alloy aged at 400◦C for 10 min,

respec-tively. The average weight percentages of alloying elements examined by analyzing at least 10 different EDS spectra of

each phase are listed inTable 1. For comparison, the

chemi-cal compositions of the B2 and␣ phases existing in the alloy

aged at 400◦C for 24 h are also listed inTable 1. It is noted

(7)

Fig. 5. Electron micrographs of the alloy aged at 700◦C for 2 h: (a) BF; (b) an SADP taken from the matrix. The zone axis of the D03phase is [1 1 0] (h k l = D03 phase,h k l1or2= L–J phase, 1: variant 1; 2: variant 2); (c) (0 2 01) L–J DF; (d) and (e) (¯1 1 1) and (0 0 2) D03 DF, respectively.

made in the STEM mode on thin films (not on the extracted phase) and the size of the B2 precipitates (about 60 nm) ex-amined is slightly larger than that of the electron beam spot (40 nm) produced on the JEOL JEM-2000FX STEM, some errors in the exact percentage of elemental concentrations in the B2 precipitates are inevitable. However, it is seen in Fig. 8andTable 1that both the nickel and aluminum con-centrations in the B2 precipitates are much greater than those

in the as-quenched alloy,␥₁martensite or␣ phase.

There-fore, these analyses are still enough to permit an inference that the B2 precipitates are enriched in both nickel and alu-minum. On the basis of the above analyses, it is clear that

when the as-quenched alloy was aged at 400◦C for

mod-erate times, a high density of the B2 precipitates started to appear. Since the nickel and aluminum concentrations in the B2 precipitates are very high, the remaining matrix would be

(8)

Fig. 6. Electron micrographs of the alloy aged at 750◦C for 1 h: (a) and (b) (0 0 2) and (¯1 1 1) D03DF, respectively; (c) (0 2 01) L–J DF.

Fig. 7. (0 0 2) D03 DF electron micrographs of the alloy aged at (a) 975◦C for 1 h and (b) 1000◦C for 1 h, respectively.

lowered in both nickel and aluminum. The lower concentra-tions of both nickel and aluminum would cause the marten-sitic transformation temperature of the remaining matrix to

be higher[2,7,11], which induced the D03→ ␥₁martensitic

transformation that would occur during quenching from the

aging temperature. This result is consistent with the

obser-vation inFig. 2. Furthermore, with increasing the aging time

at 400◦C, the B2 precipitates would grow. Along with the

growth of the B2 precipitates, the remaining matrix would be depleted in both nickel and aluminum. The insufficient

(9)

Fig. 8. Three typical EDS spectra obtained from (a) as-quenched alloy, (b) a B2 precipitate as well as (c)␥₁ martensite in the alloy aged at 400◦C for 10 min.

concentration of aluminum would cause the D03 matrix to

become unstable[7,25]. Consequently, the remaining D03

matrix would transform to the Cu-rich␣ phase, which is

consistent with the observation inFig. 3.

Finally, two more features are worthwhile to note as

follows: (1) the␥2 precipitates were always reported to be

observed in the Cu–(12.8–15.1)Al–(3.0–7.7)Ni alloys after

being aged at temperatures ranging from 200 to 770◦C for

various times[1–15]. However, in the present study, no

ev-idence of the␥2 precipitates could be detected. This result

implies that the higher nickel addition in the Cu–Al–Ni

alloy would unfavor the precipitation of the␥2 phase. (2)

In the previous studies of the Cu–Al–Ni ternary alloys [1–15], no information concerning the A2→ B2 transition temperature has been provided in the literatures. However, according to the phase diagram of the Cu–Al binary alloys [25,26], it is seen that the A2→ B2 transition temperature

of a Cu–14.2–Al alloy was about 710◦C. In the present

alloy, this transition temperature was found to be raised to

somewhere between 975 and 1000◦C. This indicates that

the nickel addition in the Cu–Al binary alloys would expand the B2 phase field.

5. Conclusions

The phase transformations in the Cu–14.2Al–15.0Ni alloy have been investigated by using TEM.

(1) In the as-quenched condition, the microstructure was

D03 phase containing extremely fine L–J precipitates.

The D03 phase was formed by an A2 → B2 → D03

continuous ordering transition during quenching.

(2) When the as-quenched alloy was aged at 400◦C for

moderate times and then quenched, the microstructure

was the mixture of (␥₁+B2 precipitate). This feature has

never been observed by other workers in the Cu–Al–Ni alloys before.

(3) When the as-quenched alloy was aged at

tempera-tures ranging from 400 to 1000◦C for longer times,

the phase transformation sequence as the temperature

increased was found to be (␣ + B2 precipitate) →

(B2 + B2 precipitate) → B2 → A2.

(4) The higher nickel addition in the Cu–Al–Ni alloy would

unfavor the precipitation of the␥2phase.

(5) The nickel addition in the Cu–Al binary alloys would expand the B2 phase field.

Acknowledgements

The author is pleased to acknowledge the financial support of this research by the National Science Council, Republic of China under Grant NSC90-2216-E-009-044. He is also grateful to M.H. Lin for typing.

References

[1] M.A. Dvorack, N. Kuwano, S. Polat, H. Chen, C.M. Wayman, Scripta Metall. 17 (1983) 1333–1336.

[2] N.F. Kennon, D.P. Dunne, L. Middleton, Metall. Trans. A 13 (1982) 551–555.

[3] N. Kuwano, C.M. Wayman, Metall. Trans. A 15 (1984) 621–626. [4] N. Zárubová, A. Gemperle, V. Novák, Mater. Sci. Eng. A 222 (1997)

(10)