Superplasticity in splat-quenched PbSn eutectic

Materials Science and Engineering, 45 (1980) 83 - 93

83

© Elsevier Sequoia S.A., Lausanne -- Printed in the Netherlands

Superplasticity in Splat-quenched Pb-Sn Eutectic R. CHEESE and B. CANTOR

School of Engineering and Applied Sciences, University of Sussex, Falmer, Brighton, Sussex (Gt. Britain) (Received May 9, 1979; in revised form January 25, 1980)

SUMMARY

Pb-Sn eutectic alloy tapes were prepared by melt spinning over a wide range of solidification conditions. In all cases the rapid solidification produced a microstructure consisting of equiaxed lead-rich and tin-rich grains, approximately 4 t~m in size. This microstructure could result from a divorced eutectic solidification process or more probably from post-solidification coarsening. Under tensile loading the melt-spun Pb-Sn eutectic tapes exhibited superplastic deformation with large failure strains, a sigmoidal variation of the logarithm of flow stress with the logarithm of flow strain, a strain rate sensitivity o f 0.42 and an activation energy for superplastic flow of 54 kJ mo1-1.

1. INTRODUCTION

Since the work of Duwez and coworkers [1, 2] there has been considerable scientific interest in the structure and properties of splat-quenched alloys [3 - 7]. Liquid alloy droplets have been splat quenched at cooling rates of greater than about 105 K s-1 usually either by the gun technique [ 8] in which the liquid is projected onto a solid substrate or by a piston technique [9] in which the liquid is squeezed between two solid surfaces. Depending on the particular alloy system, splat quenching can produce amorphous structures [ 1, 10], metastable crystalline phases [ 9, 11], extended solubility [12, 13], anomalous vacancy concentrations [ 14, 15] or suppression of martensite formation [16 - 18]. In many cases these novel microstructures have been shown to have potentially useful

electrical, magnetic, mechanical or electrochemical properties [3 - 7, 19, 20]. In recent years a variety of amorphous alloys have been prepared by a new technique called chill-block casting or melt spinning [21, 22], in which a stream of liquid is solidified in contact with the outer edge of a rotating drum to produce a continuous splatquenched solid tape. Because of the dimensional regularity of the resulting melt-spun tapes, they are particularly suitable for investigating the mechanical properties of splat-quenched alloys by conventional tensile testing [23]. In addition the cooling rate is reasonably uniform throughout the tape so that the microstructure shows less variability than is often found in specimens prepared by other splat-quenching techniques [23]. The melt-spinning technique has been used almost exclusively to study amorphous alloys, and the objective of the present investigation was to extend the technique to crystalline splatquenched alloys using Pb-Sn eutectic as a model material. The effect of melt-spinning conditions on the solidification of Pb-Sn is described elsewhere [24]; in this paper we describe the tensile properties of the meltspun Pb-Sn eutectic tapes.

2. EXPERIMENTAL TECHNIQUE

Eutectic Pb-Sn was obtained by the induction melting of 99.99% pure lead and tin in a recrystallized alumina crucible under a dynamic argon atmosphere. Splat-quenched tapes were prepared by melt spinning as follows. Approximately 3 g of eutectic Pb-Sn was placed in a quartz tube 100 mm long and 9 mm in diameter which narrowed to a 1 mm nozzle at the bottom and which was flushed by argon gas from the top. The alloy was

84

induction melted and then ejected onto a highly polished copper drum 150 mm in diameter which was a b o u t 1 mm from the nozzle and was rotating at a rate of up to several thousand revolutions per minute. The ejection gas was argon at 0.01 Pa above atmospheric pressure. A series of tapes were melt spun; they were typically 2 - 3 m long and a b o u t 1 mm wide, with thicknesses depending on the rotation rate of the copper drum. By varying the rotation rate the surface speed of the outer edge of the drum was varied between 3 and 14 m s -z to produce tape thicknesses in the range 30 - 120/~m. Further details of the relationships between tape thickness and width, drum speed, ejection pressure and nozzle diameter will be published elsewhere [24]. Mechanical properties were measured in a conventional Instron tensile testing machine. The thickness of each tensile specimen was taken from the mean of six micrometer measurements, the width was taken from the mean of four measurements with a travelling microscope and the gauge length was measured with a vernier height gauge. Several specimens of each tape were tested. A typical gauge length was a b o u t 100 mm, and specimens were gripped at each end by being wrapped around a steel capstan 35 mm in diameter before being clamped into position. This gripping system was effective at preventing stress concentrations at the grips and very few specimens failed outside the gauge length. During preliminary testing at strain rates of 5 X 10 -4 s -z, specimens frequently exhibited strains in excess of 100% at a steadily decreasing flow stress. This suggested that the tapes were superplastic, so in subsequent tensile tests the flow stress was measured for each tape thickness as a function of strain rate in the range 3 X 10 - s - 2 X 1 0 - 2 S- l . This was achieved by the technique of varying the cross-head speed during tensile testing, so data were obtained at a relatively small total specimen strain. To investigate the activation energy for the flow process, the variation in the flow stress with strain rate was determined for two tape thicknesses over a range of temperatures from 273 to 373 K by heating in a h o t air oven. The fine scale of a splat-quenched microstructure can make it difficult to etch specimens for microscopic examination [ 23, 25].

The best results in the present experiments were obtained w i t h o u t etching, by observing the atomic number contrast between the leadand tin-rich phases in a Cambridge Stereoscan 2A scanning electron microscope. Specimens were examined in this way both before and after tensile deformation.

3. RESULTS

Davies et al. [26] have described the characteristics of superplasticity as follows. In general the flow stress of a metal can be expressed as o = K1 e"~ m

(1)

where o is the flow stress, e the strain, g the strain rate, n the strain-hardening exponent, m the strain rate sensitivity and Kz a constant. At temperatures below a b o u t 0.4 times the melting point, m is small and deformation is controlled by strain hardening so that eqn. (1) reduces to (2)

O = K 2 en

with K2 approximately constant. At higher temperatures fine-grained two-phase materials exhibit a characteristic sigmoidal variation of flow stress with strain rate (Fig. 1). The details of the deformation mechanism have n o t y e t been fully resolved. However, it is well established that at intermediate strain rates {region II) deformation is superplastic and is controlled by grain boundary sliding. In this superplastic region the flow stress is

1

H

m ~o~

%

f

"0-5

~ 0.1 0.05

|

i

I

!

10-5

16 4

15 3

16 2

STRAIN

RATE

,

S-1

Fig. 1. Variation in flow stress with strain rate for n i n e t h i c k n e s s e s of melt-spun splat-quenched P b - S n e u t e c t i c t a p e s t e n s i l e t e s t e d at room temperature (297 K): t , 31 pro; A, 44 pro; z~, 54/~m; v , 58 pro; +, 63 pm; x, 75 pro; ", 80 pm; o, 92 pm; V, 116/zm.

85

strongly dependent on strain rate with a strain rate sensitivity m of more than 0.3; the flow stress is given by

s u p e r p l a s t i c m a t e r i a l , t h e f l o w stress v a r i e d sigmoidally with strain rate and the strain rate sensitivity decreased w h e n the strain rate was e i t h e r g r e a t e r t h a n 2.0 × 1 0 - 3 s - 1 o r less t h a n 1 0 - 5 s- 1 . T o d e t e r m i n e w h e t h e r t h e t a p e t h i c k n e s s h a d a n y e f f e c t o n t h e relat i o n s h i p b e t w e e n f l o w stress a n d s t r a i n r a t e , e q n . (3) w a s r e w r i t t e n as

(3)

0 = K3 ~m

with K 3 approximately constant. At the lowest strain rates (region I) there is sufficient time for diffusive flow; at the highest strain rates (region III) dislocation glide becomes important. In both cases the strain rate sensitivity falls below 0.3. Figure 1 shows the variation in flow stress with strain rate at room temperature for nine thicknesses of melt-spun P b - S n eutectic tapes. In the strain rate range from approximately 10 -5 to 1 0 - 3 S- 1 the splat-quenched P b - S n eutectic was superplastic with a strain rate sensitivity of a b o u t 0.4. As expected for a

lOgl0o = log10 K3 + m log10 e

L in e a r regresssion analysis was applied to the log10 o - l o g l o g d a t a i n t h e s t r a i n r a t e r a n g e 1 0 - 5 - 1 0 - 3 s- 1 t o d e t e r m i n e t h e c o r r e l a t i o n c o e f f i c i e n t r, t h e s t r a i n r a t e s e n s i t i v i t y m a n d t h e c o n s t a n t o f p r o p o r t i o n a l i t y K3. T h e r e s u l t s are s h o w n i n T a b l e 1. W h e n t h e regression analysis was applied separately to each of the nine tape thicknesses of 30 -

TABLE 1 Strain rate sensitivity of melt-spun P b - S n eutectic tapes as a function of tape thickness and testing temperature Thickness (pm)

(4)

Testing temperature T(K)

Strain rate sensitivity m

Correlation coefficient r

lOgl O K3

31 ± 2

297

0.48

0.995

1.86

44 ± 4

297 312 328 334 348 368

0.42 0.42 0.45 0.45 0.43 0.40

0.998 0.999 0.998 0.997 0.997 0.995

1.68 1.48 1.43 1.38 1.12 0.87

54 ± 4

297

0.40

0.996

1.57

58 ± 2

297

0.40

0.993

1.58

63 ± 3

297

0.41

0.997

1.63

75 ± 6

297 310 319 324 329 339 342 349 359

0.40 0.43 0.43 0.42 0.43 0.46 0.45 0.42 0.47

0.999 0.999 0.999 0.998 0.998 0.997 0.997 0.996 0.996

1.56 1.49 1.36 1.31 1.32 1.31 1.17 1.05 1.06

80 ± 4

297

0.42

0.994

1.64

92 ± 5

297

0.41

0.995

1.59

116 ± 11

297

0.44

0.994

1.64

T = 297 K (all specimens)

0.42

0.979

1.63

nS, T -- 297 K

0.42 ± 0.02

n~, T > 297 K

0.43 ± 0.02

r~ (all specimens)

0.43 ± 0.02

86

5

T ~297 K ~o- ~ "e~''311

~=E

o ~ ~ ,=., ~ ~ ~

2.0

K

strain rate

111328 K

,~ s& D,¢~ , I 0 / .k K / • AV

/ 0." 359 K

1.0

= .~• ~

0.8

a" 0.6

0"1

~ -~

~,.q"~

•"'

=E o~

~,e ,,o

~s

..v

"

v~

•

~

,~

/

~ ~

•

0"05

•" 0'4

I

I

I

165

164

153

STRAIN

RATE

I

~.~~ ."

7"

&,'"

,V • V

=

.•;

=¢

o.;

/

s /

162

S-1

Fig. 2. Variation in flow stress with strain rate for two thicknesses of melt-spun splat-quenched P b - S n eutectic tapes tensile tested at the various temperatures indicated in the figure: 0, ~, ~, v, 44 p m ; o , ~, m , v , t , 75 pro.

120/~m, the room temperature strain rate sensitivity was always in the range 0.40 - 0.48 with a correlation coefficient greater than 0.99. When the regression analysis was applied to the complete set of room temperature data irrespective of tape thickness, the strain rate sensitivity was 0.42; significantly the correlation coefficient remained very high at 0.98, indicating that the variation in flow stress with strain rate was independent of tape thickness. At a given strain rate the flow stress decreased as the testing temperature was raised above room temperature. Typical curves of flow stress against strain rate are shown in Fig. 2 for two tape thicknesses and several temperatures. Regression analysis applied separately to the different temperatures and tape thicknesses showed that the strain rate sensitivity was still always in the range 0.40 0.48 and the correlation coefficients were greater than 0.99 (see Table 1). The constant of proportionality K3 varied with the testing temperature, so it was n o t possible to apply a regression analysis to the complete set of data at all tape thicknesses and temperatures. However, the different tape thicknesses tested at room temperature had a mean strain rate sensitivity of 0.42 with a standard deviation of +0.02; for the specimens tested above room temperature, the mean strain rate sensitivity was 0.43 with again a standard deviation of +0.02 (see Table 1), confirming that the strain rate sensitivity was constant within a b o u t 5%, independent of testing temperature as well as tape thickness.

0.2

/'•

~"

.o"

T-1 ,

10-3K -1

Fig. 3. Arrhenius plots of the logarithm of flow stress against inverse temperature for two thicknesses of melt-spun P b - S n eutectic tapes tensile tested at the various strain rates indicated in the figure: v, A, ©, 4 4 / j m ; v, A, o, 75 pro.

et al.

Davies [26] give the temperature dependence of superplastic flow as

~= K4daobexp(--R~ )

(5)

where K4 is another constant, d is the grain size, the e x p o n e n t a is in the range --2 > a > --3, the e x p o n e n t b = m -1, Q is the activation energy for superplastic flow, R is the gas constant and T is the absolute temperature. The flow stress-strain rate data at different temperatures were replotted as log10 o against T -1 according to the Arrhenius equation (eqn. (5)) and typical results are shown in Fig. 3. From eqn. (5) the gradient of the Arrhenius plot is and regression analysis was applied to the log10 o - T -1 data to obtain values for the activation energy for superplastic flow. The results given in Table 2 show that for each tape thickness (44 or 75 pm) and for each strain rate in the range 10 -5 - 2.0 × 10 -3 s-1 the activation energy was between 49 and 59 kJ mo1-1 with a correlation coefficient always greater than 0.98. The activation energy had a mean value of 54 kJ mo1-1 with a standard deviation of 3 kJ mol-1 and was therefore constant within a b o u t 5% for all strain rates and tape thicknesses. At a given strain rate the Arrhenius plots for the two tape thicknesses were very

--Qm/R,

87 TABLE 2 A c t i v a t i o n energy for superplastic flow in melt-spun P b - S n eutectic tapes as a f u n c t i o n of tape thickness and strain rate

Strain rate ~ (s -1)

1.2 2.3 5.9 1.2 2.3 5.6 1.0 2.1

× 10 - 5 x 10 - 5 X 10 -5 x 10 - 4 × 10 - 4 × 10 - 4 × 10 - 3 x 10 - 3

(all specimens)

Activation energy Q (kJ mo1-1) for super-

Correlation coefficient r

plastic flow Tape 44 pm thick

Tape 44 pm thick

49 52 55 54 50

Tape 75 tim thick

56 57 59 58 56 52 50 49

Tapes both 44 and 75 pm thick

53 56 57 55 50

0.998 0.998 0.999 0.999 0.997

Tape 75 pm thick

0.989 0.983 0.983 0.986 0.990 0.990 0.989 0.975

Tapes both 44 and 75 pm thick

0.971 0.981 0.982 0.984 0.985

54 _+ 3

(a)

(b)

Fig. 4. Scanning electron micrographs of (a) the t o p surface and (b) the b o t t o m surface o f a melt-spun P b - S n e u t e c t i c tape. The dark regions are tin rich and the light regions are lead rich. On b o t h surfaces the lead- and tin-rich regions are equiaxed, a b o u t 4 p m in size. The b o t t o m surface shows imprints f r o m grooves in the polished c o p p e r drum.

similar (Fig. 3) and when regression analysis was applied irrespective of tape thickness the correlation coefficients remained high at greater than 0.97 (Table 2, last column). From eqn. (5) this suggested that the grain size was approximately the same for both tape thicknesses. Figure 4 shows the microstructure of a melt-spun Pb-Sn eutectic tape observed directly by atomic number contrast in the

scanning electron microscope. The lead atoms which have a high atomic number produce more secondary and backscattered electrons than do the tin atoms which have a relatively low atomic number. Thus the microstructure in Fig. 4 shows equiaxed lead-rich {light) and tin-rich (dark) regions about 4 pm in size. This identification of lead-rich and tin-rich regions was confirmed by scanning X-ray microanalysis (Fig. 5). The microstructure

88

(a)

(b)

(c) Fig. 5. Scanning electron micrographs of melt-spun P b - S n eutectic tape to confirm the interpretation of the dark regions as tin rich and the light regions as lead rich: (a) the secondary electron image; (b) the Pb Ms X-ray image; (c) the Sn L~ X-ray image.

was similar on both surfaces of each tape and appeared to be independent of tape thickness, as had been suggested by the high temperature tensile tests (as discussed above). This was investigated further by using micrographs such as Fig. 4 to determine the apparent grain size. On each surface of each specimen the apparent grain size was determined from the mean of 30 measurements of the number of lead-rich and tin-rich regions intersected by a 30 gm line. The results given in Table 3 show that the apparent grain size was approxi-

mately 4 ~m, that it was the same within about 1 um on both tape surfaces and that it was independent of tape thickness. By re-examining the tape surfaces after tensile testing, the deformation process was shown to involve grain boundary sliding, as would be expected for superplastic flow. Extensive grain boundary sliding produced the characteristic surface steps shown in Fig. 6. BY examining an otherwise undeformed tape surface in the region of a Knoop microhardness indent (Fig. 7), we confirmed

89 TABLE 3 Grain size in melt-spun P b - S n eutectic tapes as a function of tape thickness

Thickness

(~m)

31~2 44~4 54~4 63~ 3 75±6 80~4 92~5 116~ 11

Grain size d (pm) As solidified, As solidified, As deformed top surface bottom surface 4.1±1.0 3.9±1.0 3.9±0.8 4.0±0.9 4.4±1.2 4.0±1.0 3.9±1.3 4.0±1.0

d(all 4.0 specimens)

4.2±1.2 3.8±0.7 3.8±0.8 3.3±0.8 3.8±1.1 4.3±1.0 3.6~ 1.2 3.7±1.2

2.9±0.6 2.6±0.5 3.5±0.9 2.8±0.6 2.9±0.7 3.2±0.4

3.8

3.0

(a)

The top surface is the surface not in contact with the rotating copper drum during solidification. Each grain size was determined as the mean of 30 measurements of the number of grain intersections along a 30 pm line using micrographs such as Figs. 4 and 5. Scatter bands are given as plus or minus one standard deviation.

that these steps were caused by deformation. With increasing distance from the indent the number of surface steps decreased because of the decreasing amount of deformation. On micrographs such as Figs. 6 and 7, steps could be seen at all three types of boundary, i.e. Pb-Pb, Pb-Sn and Sn-Sn, indicating that sliding along each type of boundary contributed to the deformation process. The apparent grain size after deformation was measured from micrographs such as Fig. 6, using the same procedure as that described above. The results, included in Table 3, show that the apparent grain size decreased from about 4 ~m before deformation to about 3 pm after deformation. The main reason for this decrease was that the atomic number contrast revealed only Pb-Sn boundaries in the as-solidified microstructure (Fig. 4); after deformation a similar number of Pb-Sn boundaries were present, but Pb-Pb and Sn-Sn boundaries could also be seen because of the slip steps produced during grain boundary sliding {Fig. 6).

(b) Fig. 6. Scanning electron micrographs of (a) the top surface and (b) the b o t t o m surface of a melt-spun Pb-Sn eutectic tape after superplastic deformation. Both surfaces show surface steps caused by grain boundary sliding at Pb-Pb, Sn-Sn and Pb-Sn boundaries.

4. DISCUSSION

Previous workers have shown that Pb-Sn eutectic is superplastic when solidified conventionally and then heavily deformed to produce a fine-scale two-phase microstructure [26 - 29]. Avery and Backofen [27] have quoted the strain rate sensitivity for Pb-Sn eutectic prepared in this way as 0.48 at

90

Fig. 7. Scanning electron micrograph of a melt-spun Pb-Sn eutectic tape in the region of a Knoop indent. The number of surface steps decreases with increasing distance from the indent, confirming that the steps are caused by deformation.

room temperature when the grain size is a b o u t 4 tam and the strain rate is 10 -4 s -1. From the temperature variation of flow stress, Cline and Alden [29] have measured the superplastic activation energy for P b - S n eutectic to be 48 kJ mo1-1. Our results show clearly that melt-spun P b - S n eutectic tapes exhibit classical superplastic behaviour during tensile deformation. The melt-spun P b - S n tapes exhibit a sigmoidal variation of flow stress with strain rate as expected for a superplastic material tested by the cross-head variation technique; the strain rate sensitivity is 0.42 in the strain rate range 10 -5 10 -3 s-1, in good agreement with Avery and Backofen's [27] data for conventionally prepared superplastic P b - S n tested under comparable conditions; the activation energy for flow is 54 kJ mol-1, also in good agreement with Cline and Alden's [29] data for conventional superplastic Pb-Sn. Two further pieces of evidence support the conclusion that melt-spun P b - S n tapes are superplastic. Firstly, microstructural examination of the tape surfaces (Figs. 4 - 7) shows that grain boundary sliding is an important deformation mechanism, and this is characteristic of superplastic flow. Secondly, almost all specimens tested to failure in the strain rate range 10 -5 - 10 -3 s -1 exhibit large failure

strains of 80 - 200% and fail by necking to a point, as expected for a superplastic material. Tapes cannot extend by more than a b o u t 200% because of their very small initial thickness (30 - 120 tam), and in a few cases premature failure is initiated at a defect in the unpolished as-solidified tape surface. There are t w o possible explanations of the miCrostructure of melt-spun P b - S n eutectic (Fig. 4) which causes superplastic behaviour during tensile deformation. Either the microstructure is a direct result of the solidification process or it is caused by a post-solidification coarsening. Let us consider the solidification process first. During melt spinning, heat is removed from the solidifying tape by radial heat conduction into the rotating copper drum. At a given position along the length of the tape the heat flow is essentially unidirectional, so that the solidification front is approximately parallel to the surface of the drum and moves from the b o t t o m surface of the tape (in contact with the drum) through to the free top surface to complete the solidification process. It is well established [30, 31] that during unidirectional solidification eutectic microstructures are determined by two main parameters, namely the rate R of m o v e m e n t of the solidification front and the temperature gradient G at the solid-liquid interface. Let us assume that the tape becomes completely solid at a distance x from the point of impingement of the liquid stream. Then the solidification rate R is given by (6)

R = dV/x

where d is the tape thickness and V is the surface velocity of the drum. In the present experiments, d ~- 50 tam, V ~ 10 m s-1 and D = 150 mm (where D is the drum diameter). Fully solid tapes can be seen to leave the drum after approximately a tenth of one revolution so that x < 50 mm; however, experiments on higher melting point metals show that the melt pool extends only a few millimetres from the point of impingement [32, 33]. Thus the solidification rate R is in the range 10 -2 10 -1 m s-1. Previous estimates of solidification rates in other splatquenching techniques are also in the range 10 -2 - 10 -1 m s -1 [22, 34, 35]. It is more difficult to estimate the temperature gradient during melt spinning. If there is no barrier t o -

91 TABLE 4 Solidification parameters during melt spinning compared with single-crystal growth by directional solidification [36] and chill casting [37] Solidification process

Temperature gradient G ( K m -1 )

Solidification rate R ( m s -1 )

G/R (K s m -2)

Melt spinning

~>4 × 106

10 -2 - 10 -1

>~4 × 107

Single-crystal growth by directional solidification [36]

103

Chill casting [37]

2 x 103

3 × 10 - 5

2 × 10 - 4

heat flow at the d r u m - t a p e interface and if the drum is a perfect heat sink, then the d r u m - t a p e interface remains at room temperature TR. At the point of impingement, as solidification begins the temperature gradient is very high close to the d r u m - t a p e interface. As solidification proceeds, the temperature gradient falls until the outer liquid surface reaches the melting point TM and the temperature gradient is approximately (TM -- TR)/d, i.e. 4 × 106 K m -1 in the present experiments. The slowest solidification rate is also at the outer surface of the tape, so the solidification ratio G/R is greater than a b o u t 4 X l 0 s K s m -2. Inefficient heat transfer at the d r u m - t a p e interface may lead to a somewhat lower value, say a b o u t 4 × 107 K s m -2. For comparison, Table 4 shows values of G, R and G/R for melt spinning, directional solidification in a Bridgman crystal grower [36] and small-scale chill casting [ 3 7 ] . It is interesting to note that G and R are both 102 - 10 a greater in melt spinning than in conventional directional crystal growth experiments b u t that the parameter G/R is approximately the same in both processes. Directionally solidified P b - S n and other simple eutectics such as A1-A12Cu have been studied rather extensively [38, 39] because of the current interest in using directionally solidified nickel- and cobalt-based eutectics as jet-turbine blade materials [40 - 4 2 ] . In a crystal grower with G and R values as in Table 4, directionally solidified P b - S n , in c o m m o n with many other eutectics, has a highly regular morphology with alternate lamellae of the two eutectic phases aligned parallel to the heat flow direction [38, 4 3 ] . When G/R falls below a critical value which depends on the level of ternary impurities,

5 × 107

107

a planar solid-liquid interface becomes unstable and the lamellae become less well aligned, forming the characteristic wheatsheaf arrangement of eutectic colonies [30]. When R is small the lamellae can become degenerate [ 3 0 ] , and the melt~spun P b - S n eutectic microstructure in Fig. 4 is not unlike the degenerate microstructure seen in both directionally solidified [44] and splatquenched [45] A1-A12Cu eutectic. However, it is improbable that Fig. 4 shows a degenerate eutectic microstructure for three reasons. Firstly, degenerate P b - S n eutectic usually consists of wavy lamellae rather than a relatively equiaxed microstructure as in A1A12Cu [43]. Secondly, a degenerate eutectic structure is rather improbable with the high values of G, R and G/R in melt spinning (Table 4). Thirdly, Cline and Livingston [39] have seen rather different microstmctures in P b - S n eutectic directionally solidified with G > 104 K m -1 and R as high as 6 × 10 _3 m s -1, i.e. with conditions approaching those in melt spinning. At high R the region of coupled eutectic growth moved to tin-rich compositions; thus, with R > 3 × 10 -3 m s -1, alloys of eutectic composition solidified with primary lead dendrites rather than with a fully eutectic microstructure. It is more probable that the microstructure in Fig. 4 is formed when the solidification structure coarsens. Lamellar eutectic microstructures are usually very stable, even after exposure for long periods at high temperature [46, 4 7 ] . However, the melting point of P b - S n eutectic is n o t very much above room temperature, and quite rapid coarsening has been seen in the degenerate wavy lamellar microstructure [43]. Moreover, the coarsening rate can be enhanced very considerably,

92

either by the presence of primary dendrites [ 4 7 ] w h i c h are e x p e c t e d f r o m Cline a n d Livingston's [39] experiments or by deformat i o n [48] w h i c h m i g h t be p r o d u c e d b y t h e shearing e f f e c t o f t h e liquid f l o w d u r i n g m e l t spinning. I n o u r e x p e r i m e n t s an a t t e m p t was m a d e to c o n f i r m this i n t e r p r e t a t i o n b y observing a melt-spun Pb-Sn eutectic tape a l m o s t i m m e d i a t e l y a f t e r solidification. T h e m i c r o s t r u c t u r e was similar to t h a t in Fig. 4 b u t on a m u c h f i n e r scale, w h i c h c o u l d well i n d i c a t e a less w e l l - d e v e l o p e d c o a r s e n i n g o f t h e as-solidified m i c r o s t r u c t u r e . Previous experiments on splat-quenched Pb-Sn e u t e c t i c h a v e s h o w n t h a t t h e r e is little or n o e x t e n s i o n o f s o l u b i l i t y in t h e t w o e u t e c t i c p h a s e s [49, 50] a n d t h a t a m e t a s t a b l e h e x a g o n a l p h a s e f o r m s w h e n t h e e u t e c t i c is q u e n c h e d t o l o w t e m p e r a t u r e s [51, 5 2 ] ; however, there have been no reports about t h e e u t e c t i c m i c r o s t r u c t u r e . T h i s is c u r r e n t l y b e i n g i n v e s t i g a t e d f u r t h e r a n d t h e results will be d e s c r i b e d in a f u t u r e p u b l i c a t i o n .

ACKNOWLEDGMENTS We w o u l d like t o t h a n k t h e Science R e s e a r c h C o u n c i l a n d the U.S. O f f i c e o f Naval Research (Contract N-00014-78-G-0048) for financial s u p p o r t o f this r e s e a r c h p r o g r a m m e a n d P r o f e s s o r R. W. C a h n f o r p r o v i s i o n o f l a b o r a t o r y facilities.

REFERENCES 1 W. Klement, Jr., R. H. Willens and P. Duwez, Nature (London), 187 (1960) 809. 2 P. Duwez, R. H. Willens and W. Klement, Jr., J. Appl. Phys., 31 (1960) 1136. 3 Proc. Int. Conf. on Metastable Metallic Alloys, Brela, Yugoslavia, 1970, in Fizika (Zagreb), 2, Suppl. 2 (1970). 4 N. J. Grant and B. C. Giessen (eds.), Proc. 2nd Int. Conf. on Rapidly Quenched Metals, Cambridge, Massachusetts, 1975, Vol. 1, Massachusetts Institute of Technology Press, Cambridge, Massachusetts, 1976; Vol. 2, in Mater. Sci. Eng., 23 (1976) 81 - 308. 5 H. Jones, Rep. Prog. Phys., 36 (1973) 1425. 6 T. R . Anantharaman and C. Suryanarayana, J. Mater. Sci., 6 (1971) 1111. 7 B. Cantor (ed.), Proc. 3rd Int. Conf. on Rapidly Quenched Me tals, Brigh ton, 1978, Metals Society, London, 1978. 8 P. Duwez and R. H. Willens, Trans. Metall. Soc. AIME, 227 (1963) 362.

9 R.W. Cahn, K. D. Krishananand, M. Laridjani, M. Greenholz and R. Hill, in N. J. Grant and B. C. Giessen (eds.), Proc. 2nd Int. Conf. on Rapidly Quenched Metals, Cambridge, Massachusetts, 1975, Vol. 2, inMater. Sci. Eng., 23 (1976) 83. 10 Metallic Glasses, American Society for Metals, Metals Park, Ohio, 1978. 11 P. Ramachandrao, M. G. Scott and G. A. Chadwick, Philos. Mag., 25 (1972) 961. 12 H. Jones and C. Suryanarayana, J. Mater. Sci., 8 (1973) 705. 13 B. C. Giessen, in B. C. Giessen (ed.), Developments in the Structural Chemistry of Alloy Phases, Plenum, New York, 1969. 14 G. Thomas and R. H. Willens, Acta Metali., 12 (1964) 191. 15 G. Thomas and R. H. Willens, Acta Metall., 13 (1965) 139; 14 (1966) 138. 16 Y. Inokuti and B. Cantor, Scr. Metall., I0 (1976) 655. 17 Y. Inokuti and B. Cantor, J. Mater. Sci., 12 (1977) 946. 18 F. Duflos and B. Cantor, in B. Cantor (ed.),Proc. 3rd Int. Conf. on Rapidly Quenched Metals, Brigh ton, 1978, Vol. 1, Metals Society, London, 1978, p. 110. 19 J. J. Rayment and B. Cantor, Met. Sci., 12 (1978) 156. 20 J. J. Rayment and B. Cantor, in B. Cantor (ed.), Proc. 3rd Int. Conf. on Rapidly Quenched Metals, Brighton, 1978, Vol. 1, Metals Society, London, 1978, p. 85. 21 S. Kavesh, in Metallic Glasses, American Society for Metals, Metals Park, Ohio, 1978, p. 36. 22 H. Jones, in N. J. Grant and B. C. Giessen (eds.), Proc. 2nd Int. Conf. on Rapidly Quenched Metals, Cambridge, Massachusetts, 1975, Vol. 1, Massachusetts Institute of Technology Press, Cambridge, Massachusetts, 1976, p. 1. 23 B. Cantor, to be published. 24 S. J. B. Charter, R. Cheese and B. Cantor, J. Mater. Sci., in the press. 25 M. G. Scott, personal communication, 1978. 26 G. J: Davies, J. W. Edington, C. P. Cutler and K. A. Padmanabhan, J. Mater. Sci., 5 (1970) 1091. 27 D. H. Avery and W. A. Backofen, Trans. Am. Soc. Met., 58 (1965) 551. 28 P. J. Martin and W. A. Backofen, Trans. Am. Soc. Met., 60 (1967) 352. 29 H. E. Cline and T. H. Alden, Trans. Metall. Soc. AIME, 239 (1967) 710. 30 G. A. Chadwick, Prog. Mater. Sci., 12 (1963) 97. 31 B. Cantor, in G. A. Chadwick and D. A. Smith (eds.), Interfaces, Academic Press, New York, in the press. 32 J. L. Walter, in B. Cantor (ed.), Proc. 3rd Int. Conf. on Rapidly Quenched Metals, Brighton, 1978, Vol. 1, Metals Society, London, 1978, p. 30. 33 H. Hillmann and H. R. Hilzinger, in B. Cantor (ed.), Proc. 3rd Int. Conf. on Rapidly Quenched Metals, Brighton, 1978, Vol. 1, Metals Society, London, 1978, p. 22. 34 D. R. Harbur, J. W. Anderson and W. J. Maraman, Trans. Metall. Soc. AIME, 245 (1969) 1055.

93 35 M. H. Burden and H. Jones, J. Inst. Met., 98 (1970) 249. 36 B. Cantor and G. A. Chadwick, J. Cryst. Growth, 23 (1974) 12. 37 F. Jabczynski and B. Cantor, in J. L. Walter, M. F. Gigliotti, B. F. Oliver and H..Bibriag (eds.), Proc. 3rd Conf. on In Situ Composites, 1978, Ginn Custom, Lexington, Mass., 1979, p. 303. 38 R. H. Hopkins and R. W. Kraft, Trans. Metall. Soc. AIME, 242 (1968) 1627. 39 H. E. Cline and J. D. Livingston, Trans. Metall. Soc. AIME, 245 {1969) 1987. 40 Proc. Conf. on In Situ Composites, in Publ. NMAB-1308, 1973, National Academy of Sciences, Washington, D.C. 41 M. R. Jackson, J. L. Walter, F. D. Lemkey and R. W. Hertzberg (eds.), Proc. 2nd Conf. on In Situ Composites, Xerox Corporation, Lexington, Massachusetts, 1976. 42 J. L. Walter, M. F. Gigliotti, B. F. Oliver and H. Bibriag (eds.), Proc. 3rd Conf. on In Situ Composites, 1978, Ginn Custom, Lexington,

Massachusetts, 1979. 43 J. D. Verhoeven, D. P. Mower and E. D. Gibson, Metall. Trans. A, 8 (1977) 1239. 44 T. Z. Kattamis and M. C. Flemings, Metall. Trans., 1 (1970) 1449. 45 M. G. Scott, M. Kijek and H. Matyja, J. Mater. Sci., 13 (1978) 1354. 46 L. D. Graham and R. W. Kraft, Trans. Metall. Soc. AIME, 236 (1966) 94. 47 B. Cantor and G. A. Chadwick, J. Cryst. Growth, 30 (1975) 140. 48 L.Y. Lin, T. H. Courtney and K. M. Ralis, Acta Metall., 25 (1977) 99. 49 I. V. Salli and I. S. Miroshnichenko, Dokl. Akad. Nauk SSSR, 132 (1960) 557. 50 N. I. Varich and A. A. Yakunin, Russ. J. Phys. Chem., 41 (1967) 437. 51 R. H. Kane, B. C. Giessen and N. J. Grant, Acta Metall., 14 (1966) 605. 52 I. V. Salli, V. A. Dzenzersky, V. S. Shvets and T. Y. Evina, Akad. Nauk Ukr. SSR, Metallofizika, 33 (1971) 59.