Directionally solidified copperCuZrSi pseudo-binary eutectic alloys

Directionally solidified copperCuZrSi pseudo-binary eutectic alloys

Materials Scienceand Engineering, 11 (1973)203-209 © American Societyfor Metals, Metals Park, Ohio, and ElsevierSequoia S,A., Lausanne - Printed in th...

928KB Sizes 0 Downloads 15 Views

Materials Scienceand Engineering, 11 (1973)203-209 © American Societyfor Metals, Metals Park, Ohio, and ElsevierSequoia S,A., Lausanne - Printed in the Netherlands

203

Directionally Solidified Copper--CuZrSi Pseudo-Binary Eutectic Alloys A. J. PERRY

Brown BoveriResearch Centre, CH-5401 Baden/Switzerland) (Received July 25, 1972)

Summar3 ~, The tensile properties of as-grown copper-CuZrSi E phase pseudo-binary eutectie alloys have been measured at ambient temperatures. The tensile curves display the characteristics of a Liiders band propagation on yielding. The volume fraction of the CuZrSi phase is only just sufficient for fibre strengthening. The total elongation is of the order of 8 10°/o. The ultimate tensile strength increases with growth rate until colonies are formed giving the maximum UTS as about 42 kg/mm 2. At high growth rates multiple fracturin 9 of the CuZrSi phase lamellae was observed.

NiTiSi found by Shoemaker and ShoemakerL Nickl and Sprenger ~ found that the CuZrSi E phase comes into direct pseudo-binary eutectictype equilibrium with copper. The composition and temperature of the eutectic were given variously as : Cu-2.3 mol. % CuZrSi at 1055°C +_5 deg C ref. 1. Cu-2.63 mol. % CuZrSi at 1045°C + 5 deg C ref. 2.

In recent work Nickl, Sprenger and Schmidt 1"2 found that directionally solidified copper-CuZrSi pseudo-binary eutectic alloys possess a combination of reasonable strength and conductivity at ambient temperatures and also at 300°C. They cite as best values 36 kg/mm 2 and 44 m/flmm 2 at ambient temperatures for the ultimate tensile strength 1 (UTS) and electrical conductivity 2, respectively, in samples grown at 12 15 cm/h. The total elongation was of the order of 2-3 % and fracture was associated with growth bands. In the present work we have made a further study of the tensile properties of these alloys, utilizing samples which were free from growth bands.

F r o m studies of as-cast material we have found the intermediate composition of about 2.5 tool. To to be correct" 95.52 wt. To Cu, 3.43 wt.% Zr and 1.05 wt.% Si, and have measured the eutectic temperature as 1048+2°C by differential thermal analysis. The three compositions have been studied here, namely eutectic, hypo- and hyper-eutectic alloys. These last were chosen because their tensile properties are reported ~ to be somewhat better than the eutectic itself. The volume fractions of E phase are 8.49 vol. %, 9.18 vol. % and 9.62 vol. OJ~o,respectively, calculated taking the theoretical density of the E phase as 6.54 g/cm 3 from the published lattice parameters 4. Directional solidification of the eutectic produces a blade or broken lamellar type of structure which is typified by the longitudinal and transverse microsections given in Fig. 1. On the copper-rich side of the eutectic the excess copper appears in dendritic form. In hyper-eutectic alloys the primary E phase is angular but becomes blade-like following directional solidification.

THE COPPER CuZrSi EUTECTIC

EXPERIMENTAL PROCEDURE

The CuZrSi phase is equiatomic and a member of the E-phase series of which the prototype is

All alloys used in this work were made from 99.93 % pure O F H C copper, zirconium of 99.99 % purity and semiconductor silicon. Pre-cast alloy rods 3 mm in diameter were made by melting together the pre-weighed elements under flowing

INTRODUCTION

* R6sum6 en franCais5. la fin de l'article. Deutsche Zusammenfassung am SchluB des Artikels.

204

A . J . PERRY Argon in Motor coupling

Flexible coup|lng Support Tube and Crucible move downwards

sohddmd

rod

Base prate

Water ou Argon out

Fig. 2. The directional solidification equipment used in the present study. Fig. 1. Longitudinal and transverse microsections of a directionally solidified copper-CuZrSi eutectic alloy (courtesy of M. Biller and P. R. Sahm).

argon in graphite crucibles using RF induction heating. After allowing sufficient time for the melt to become homogeneous, the liquid alloy was drawn up into a quartz tube (which had previously been graphite-coated with cracked alcohol vapour) by opening the tube to a vacuum line. This method of producing the pre-casts was modified from that developed by Sahm and Lorenz 5. Directional solidification of the alloys was again carried out under argon in graphite crucibles using the equipment shown schematically in Fig. 2. As shown, the entire crucible-containing tube was water-cooled which bestowed thermal stability on the system. One specimen was run with an axially-mounted thermocouple of 0.1 mm diameter wire which was used to determine the tempera-

ture gradient as 640 deg C/cm ahead of the solid/ liquid interface and the length of the molten zone to be about 2 cm. The speed of growth was controlled by a gear-box assembly which gave 1.91 cm/h and 5.75 cm/h and their decade multiples and submultiples. The as-grown rods produced in this way were 4 mm in diameter, about 100 mm long and were machined into two threaded tensile specimens of 10 mm and 2.5 mm gauge length and diameter, respectively. These were tested in simple tension at ambient temperature with an Instron machine at a strain rate of 0.025 rain- 1, normally with a strain gauge extensometer (10 mm gauge length) coupled in up to the yield point and finally sectioned for optical microscopy. As discussed below, a number of samples were vacuum heat-treated (2 x 10-s Torr) for 100 h at 500°C. At the end of such treatment the samples were retained in vacuum during either furnace or

205

Cu-CuZrSi PSEUDO-BINARY EUTECTIC ALLOYS o A

eutechc h y p o - e u t e c t ic



.

o

heat-treated

is

h y p e r - eutect=c heat- treated

5O E ~O

~0

/ /

t-

O

T= c(u L.

U3

f

J

/

30

t

A

(u

(.)

In ¢-

I,)

(,)

20 O

E _~

(a)

10

I

I

I

I

I I

I

I

I

|

I

I

I

I I

I

I

I

I

1

I

I

I

IO Growth

I I

100

Rate

cm/hr.

15

(b)

C o

10 o

tO iii

(o) 0

I

I

I

|

J |

I

I

I

I

I

I

1

I

t

Jl

I

I

10 Growth

Rate

I

I

I

I

III

I

100 cm/hr.

Fig. 3. The (a) ultimate tensile strength, (b) elongation to fracture, of the copper-CuZrSi alloys before and after thermal treatment. A number of data points are included in parentheses from samples which contained some form of flaw as mentioned in the text.

air cooling to ambient temperature; there was no difference between the results. Chemical analysis showed that a small degree of carbon pick-up occurred (typically 0.004~) during growth. A small loss of zirconium was also remarked and is presumed to have been due to oxidation by gas occluded in the mould. The mechanical properties of alloys which were corrected for this loss were found to be unchanged; all the results obtained are therefore presented impartially in Fig. 3. During the development of the experimental

sequence discussed above, a number of specimens were made which contained flaws in the form of oxide films or undissolved silicon from an imperfect protective atmosphere or insufficient solution time during the making of the pre-casts. Some of the results from these various samples are included in parentheses in Fig. 3 as mentioned below. Parallel tests on samples of the annealed copper basis metal were also performed for reference: Conductivity 59.21 m/~mm 2 UTS

17.99 kg/mm z.

206

A.J. PERRY

30 7

copper -Cu Zr S~

%

6

E

E

/

Oe~

0 uI

/

20 In vl o i,..

4-

5

/

"" x .1=

"O w

/

101-

/

/

¢O

2

00

I

2

3

4

5

6

7

8

9

0

Strmn % Fig. 4. A typical stres~strain curve ofa eutecticalloy together with the strain recorded (right hand ordinate) by the strain gauge extensometer. Its gauge length was 10 mm as compared with 20 mm of the specimen.The sample shown had been directionally solidified at 5.75 cm/h.

RESULTS AND DISCUSSION The form of the stress-strain curve, a typical example of which is shown in Fig. 4, is quite interesting. Below the yield point the change in slope is attributed to the yielding of the copper matrix. Beyond the yield drop, which was not observed in specimens solidified at the highest speeds, the form is that normally associated with the propagation of a Liiders Band. To support this interpretation two types of subsidiary experiments were made. First, a square-sectioned specimen was tested and the surface was photographed to show the existence and propagation of a deformation band. As found below, the bands normally started at one end of the specimen (presumably in the grips); however, in this instance it developed from a flaw in the gauge length. One of the photographs

is shown in Fig. 5, where the band can be seen and further that its two edges are parallel. Also evident is a second band which was initiated at the same point; this specimen finally failed there before propagation was complete. After failure the angle of the front was determined as 47 ° to the tensile axis. Second, the extensometer was retained coupled to several specimens up to the point of mechanical instability. The local extension recorded by the 10 m m gauge length of the extensometer is shown in Fig. 4 (right hand ordinate) against the total elongation of the whole specimen. The passage of the Ltiders band through the gauge length is clearly evident ; no deformation is recorded by the extensometer when the band is beyond its limits, evidencing the local nature of the deformation. Subsequently the specimen work-hardened uniformly along its length.

Cu CuZrSi PSEUDO-BINARY EUTECTIC ALLOYS

Fig. 5. A broad Liiders band, together with a narrower one which started later, photographed shortly after the yield drop during the tensile testing of a square-sectioned specimen containing a flaw in the gauge length. The sample had been directionally solidified at 5.75 cm/h.

One puzzling aspect of the Lfiders deformation is that the stress was not constant during its propagation but increased in all samples, albeit in a discontinuous manner. This effect is not understood. It could be either work hardening of the copper matrix independently of the location of the band or the cumulative effect of several yield bands closely spaced. However, the lack of extension recorded by the extensometer when the band is beyond its limits does not appear to support either proposal, unless the spacing of successive bands is very small. The UTS and total elongation to fracture of all specimens are given in Figs. 3a and 3b, respectively, as a function of the growth rate. The results are not affected by extended heat-treatment; the structure is thus stable against thermal exposure. Except in specimens containing some form of flaw, the fracture strain is markedly greater than

207 previously reported 1. As shown in Fig. 3, any flaw of the type discussed in the previous section caused fracture mostly after an extension of the order of the yield strain, i.e. some 2 3 ~ ; flaws in this context would extend to include the growth bands of Nickl and Sprenger. As anticipated the hyper-eutectic alloys are somewhat stronger; the eutectic and hypo-eutectic series exhibit almost identical properties. Below a growth rate of 5.75 cm/h the UTS increases steadily with growth rate. At this value the strength becomes growth rate independent and finally falls slightly, at least in the hypo-eutectic alloys. This type of behavior was already observed at elevated temperatures in NiA1-Cr alloys by Walter and Cline 6 and attributed to colony formation of an increasingly fine nature. Microsections of the present samples showed colonies (Fig. 6a) at and beyond 5.75 cm/h. It was also found that the lamellae were multiply fractured in the necked region of samples grown at all speeds (Fig. 6b), independently of composition. At speeds of 57.5 cm/h and higher the repeated fracturing extended throughout the whole length of the sample; this has also been observed in carbide strengthened nickel alloys by Lemkey and Thompson 7 and attributed to a strength distribution in the fibres. It was observed further th,at the lamellae became increasingly irregular with growth rate following the onset of colony formation. The foregoing observations are interpreted in the following manner. The regular increase of the UTS with growth rate is in accord with the general observation that strength increases with decreasing lamellar spacing until colonies are formed, which offset the increase in strength. The multiple fractures in the neck indicate that the lamellae can deform plastically up to the point of mechanical instability. Further, that the volume fraction of E phase is nearly subcritical for fibre reinforcement. The increasing irregularity of the lamellae has a concomitant increase in flaw occurrence thereby causing them to weaken. Finally the strength of the lamellae becomes subcritical and results in repeated fracturing and "fibre-weakening" of the structure. Support is available for the argument. Some samples were grown at a speed of about 20 cm/h but with an increased temperature gradient to prevent colony formation (the details of this equipment will be published at a later date). The samples were colony-free and possessed a UTS consistent with the extrapolation from lower speeds as shown in Fig. 3a.

A. J. PERRY

208 i

Fig. 6a. A transverse microsection showing the formation of colonies in a sample grown at 57.5 cm/h.

Fig. 6b. A longitudinal microsection of the fracture surface of a tensile specimen. The CuZrSi E phase lamellae are seen to have cracked perpendicularly to the tensile axis. No cracking is evident beyond the necked region.

ACKNOWLEDGEMENTS

REFERENCES

The author is indebted to Dr. P. R. Sahm for suggesting the work discussed here. He is also grateful to Dr. D. J. Rowcliffe for a number of discussions on the subject of Liiders band propagation. Finally he wishes to thank Mrs. B. H 6 h n and Mrs. G. Keser for much of the metallography and Mr. E. Schoenfeld for photographing the Ltiders bands during tensile tests.

1 J. J. Nickl and H. Sprenger, Z. Metallk., 61 (1970) 229. 2 J. J. Nickl, H. Sprenger and P. W. Schmidt, Verbundwerkstoffe, Dr. Riederer Verlag, Stuttgart, 1972, p. 105. 3 C. B. Shoemaker and D. P. Shoemaker, Acta Cryst., 18 (1965) 900. 4 J. J. Nickl and H. Sprenger, Naturwiss., 54 (1967) 18, 248. 5 P.R. Sahm and M. Lorenz, J. Mater. Sci., 7 (1972) 793. 6 J. L. Walter and H. E. Cline, Met. Trans., 1 (1970) 1221. 7 F.D. Lemkey and E. R. Thompson, Met. Trans., 2 (1971) 1537.

Cu-CuZrSi PSEUDO-BINARY EUTECTIC ALLOYS

209

Alliages eutectiques pseudo-binaires cuivr~CuZrSi obtenus par solidification dirigOe

Gerichtet erstarrte pseudobindre eutektische Kul?ferCuZr Si-Legierungen

Les caractdristiques de traction d'alliages eutectiques pseudobinaires de cuivre et de phase E CuZrSi ont 6t6 mesurdes fi la temp6rature ambiante, dans l'6tat brut de fabrication. Les courbes de traction r6vblent un ddbut de d6formation plastique caract6ristique de la propagation de bandes de Liiders. La fraction volumique de la phase CuZrSi suffit tout juste pour produire un effet de renforcement par fibres. L'allongement total est de l'ordre de 8 ",i 10~/o. La charge de rupture augmente avec la vitesse de croissance jusqu'au moment off des colonies se forment. On atteint ainsi un maximum de 42 kg/mm 2 pour la charge de rupture. Dans les alliages pr6par6s avec une vitesse de croissance 61ev6e on observe des fissurations multiples dans les lamelles de la phase CuZrSi.

Die Zugeigenschaften von pseudobin~iren eutektischen Kupfe~CuZrSi-Legierungen (E-Phase) wurden bei Raumtemperatur untersucht. Die Verfestigungskurven zeigen die charakteristischen Eigenschaften der Ltidersbandausbreitung beim Fliegen. Der Volumenanteil der CuZrSi-Phase ist gerade ausreichend ffir eine Fiber-Verst~irkung. Die Gesamtdehnung ist von der Gr613enordnung 8 100/<~.Die Zugfestigkeit (ultimate tensile strength, UTS) w~ichst mit der Wachstumsgeschwindigkeit bis sich Zellen ausbilden, die eine maximale UTS von etwa 42 kg/mm 2 liefern. Bei hohen Wachstumsgeschwindigkeiten wurde mehrfacher Bruch der CuZrSi-Phase beobachtet.