Hydrostatic pressure induced ductility transitions in pure bismuth and tin-bismuth alloys

HYDROSTATIC PRESSURE PURE BISMUTH PETER

V.

DEMBOWSKIt,

INDUCED DUCTILITY AND TIN-BISMUTH JOSEPH

PEPE:

and

TRANSITIONS ALLOYS*

THOMAS

E.

IN

D.iVIDSON§

The mechanical behavior of pure (99.999?6) bismuth and tin-bismuth alloys of various compositions has been observed over a range of superimposed hydrostatic pressures. Results indicate that maxima in ductility (as measured by percent reduction in each at the fracture surface) in specimens tested at atmospheric pressure occur at compositions bordering pure tin and the eutectic compositions. _%t sufficiently high pressures all compositions failed by rupture, i.e. necking to virtually 100 per cent R-1. For pure bismuth, pressure was observed to retard failure due to the formation of cracks at twin-grain boundary interactions; this result was consistent with the hypothesis that the effect of pressure is to shift the mode of crack propagation by decreasing the normal tensile component of stress acting on a crack. TRANSITIONS DE DUCTILITE DUES A UXE PRESSIOS HYDROSTATIQL7E DASS LE BIS&rUTH PUR ET DANS LES ALLIAGES ETAIN-BISMUTH Les auteurs ont Btudi6 les propriMs m6caniques du bismuth pup (99,999%) et de divers alliages etainbismuth en prCsence de pressions hydrostatiques. Les &ultats indiquent que pour les 6chantillons d6form6s B la pression atmosph&ique, la ductilit6 maximale (mesurb par la rbduction de section en :b B la surface de rupture) se produit pour des compositions voisines de l’6tain pur et des eutectiques. Pour des pressions stisantes, tous les allies cassent avec des strictions de pratiquement 100%. Dans le bismuth pur, la pression retarde la cassure due B la formation de &sures aus intersections de joints de grains et de joints de macle; ce r&s&at s’accorde avec l’hypoth&se que la pression change le mode de propagation d’une fissure en diminuant la contra&e normale agissant sur elle. DUKTILITliTS-t%ERG&GE LEGIERT_JXGEN

IX REIXEM WISJICT U-SD IN ZIXS-WISMUTUXTER EIXEM HYDROSTATISCHEN DRUCK

Das mechanische Verhalten von reinem Wismut (99,9990,/0) und von Zinn-Wismut-Legierungen wurde bei verschiedenen hydrostatischen Drucken beobachtet. Die Ergebnisse zeigen, da9 Duktilitatsmaxima (Mel3grbBe ist prozentuale Reduzierung an BruchflBche) in den bei Atmosph&rendruck untersuchten Proben in nahezu reinem Zinn und in Proben mit nahezu eutektischer Zusammensetzung auftreten. Bei geniigend hohen Drucken kamen Proben aller Zusammensetzungen zum Bruch. d.h. Halsbildung his virtuell 100% RA. In reinem Wismut verziigert hydrostatischer Druck den Bruch aufgrund einer RiDbildung an Zwilling-Korngrenzen-Schnittlinien. Dieses Ergebnis ist in thereinstimmung mit der Hypothese, da!3 der Druck den Ril3ausbreitungsmechanismus durch ein Verminderung der normalen auf den Ril3 wirkenden Zugspannungskomponente verschiebt.

INTRODUCTION

In general, it has been observed that ductility, measured by either percent reduction in area at the fracture surface or the natural strain to fracture (ln (_&/A,)), is increased by a superimposed hydrostatic pressure. The ductility pressure function can be either linear with a characteristic slope,(1.2) nonlinear,(3*‘) or abruptly increasing within a limited B comprehensive review of the region of pressure.+‘) effects of pressure on mechanical properties of materials and techniques for determining these effects till not be attempted here, and the reader is referred to a review by Brandes@) for this information. ‘;IIost investigations concerning the effects of pressure on the ductility of materials have been limited to single phase alloys or pure materials. In * Received January 10, 1974; revised February 28, 1974. t Advanced Engineering Division, Benet Weapons Laboratory, Watervliet ArsenaI, Watervliet, Sew York, U.S.A. $ Materials Engineering Division, Benet Weapons Laboratory, Watervliet Arsenal, Watervliet, New York, U.S.A. 5 Xaterials Engineering Division, Benet Weapons Laboratory, Watervliet Arsenal, Watervliet, Sew York, U.S.A. ACT-1 5

METALLURGICA,

VOL.

f”,

SEPTEMBER

1954

this current study, a two phase alloy, wherin the t,wo phases exhibit marked differences in intrinsic ductility and pressure response is examined. Chosen for study was the tin-bismuth alloy @em. Here the nature of the system is such t,hat there is little room temperature solid solubility and t,he existence of a eutectic reaction permits variat,ion of mixture morphology. Prior workfg) has shon-n that bismuth undergoes an abrupt increase in ductility over a narrow pressure range. It has also been reported(lO) that a 50 at. % Sn-50 at. % Bi alloy shows a similar abrupt transition, but that further increases in pressure above the transition pressure results in a decrease in ductility. Such a phenomena is contrary to the observed behavior of many other materiaIs and inconsistent with proposed explanations of the pressure effect on ductility. Reported herein is the response of ductility to pressure for several Sn-Bi compositions and microstructures. The pressure effects upon ductility are interpreted in terms of the fracture modes observed. 1121

_1CT_1

112’7

JlET_1LLURGIC_1, TABLE

loo:/, 976 307b 42% 75% 90yo

Sote: bismuth.

Bi Bi Bi Bi Bi

15O’C 150°C Ambient Ambient Ambient Ambient

Hot Hydrostatic Hydrostatic Hydrostatic Hydrostatic

Virtually

XATERI.%LS

Temperature

Hot

Bi Sn-bal Sn-bal Sn-be1 Sn-bal Sn-bal

1. Fabrication

Method estrusion

Composition

no change in intercept

AND

VOL.

PROCEDURE

Post extrusion

Condition prior to annealing Hot extruded and air cooled Hot extruded and water quenched

100~~ Bi

9o& Sn-bal Bi 3006 Sn-bal Bi

Hot extruded and air cooled Hydrostatically extruded

Sn-bal Bi

do.

75?!-, Sn-bal Bi 90”; Sn-bal Bi

do. do.

42”;

Sote:

“ST”

Lubricant

treatment

Sone

a. -Air cooled b. Water quenched _1ir Cooled Sone Sone Sone None

Sone Pb-MOS, Sane Pb-MOS, Sane

of cooling rate was observed

alloys

Annealing schedule Sone Xone

(a) !ZZne (b) ST 13% for 25 hr, air cooled ST ST

Comment3 If the extrusion temp. is 66C, the grain dia immediately after extrusion is very small and incresses to approximately 12.5 mm in 24 hr Pb-JloS,

coating not removed -

is defined as heating at 19OC for 233 hr, quenching into water at O”C, upquenching to 66C for 4 hr, followed by air cooling to room temperature.

a second upquench

in pure

through use of the fixture shown in Fig. 1. Tensile force is introduced into the specimen by the advance of the main piston into the pressure cavity, with this motion in turn transmitted into the specimen by two movable legs. The other two legs are stationary and bear against the bottom closure of the high pressure cavity, thereby fixing the position of one end of the specimen. Since straining of the specimen is accomplished by the advance of the piston into the high pressure cavity, the tests are not isobaric and an increase in pressure occurs as the specimen is extended. In reporting data t.he pressure at fracture is plotted. Crosshead movement was estimated from a record of piston displacement and was maintained at a rate of 0.05 in/min. All tests were performed at ambient temperatures. Due to the relative softness of the materials used, mounting specimens for metallographic examination in an epoxy compound (Hysol) was adequate to insure reasonably good edge retention. Pure bismuth was esamined in the as polished condition using polarized light and the Sn-Bi alloys were examined in either the as polished condition, or were etched using a solution of 10 ml hydrochloric acid and 0.6 g chromic acid in 250 ml of water prior to viewing. Electron micrographs were prepared using a two stage replication technique emplo_ying chromium as a shadoting material.

TABLE 2. Annealing schedule for Bi ancl Sn-Bi Compo3ition

19i-l

conditions

gram size as a function

The bismuth and tin used in the fabrication of the alloy compositions were reagent grade stock; the nominal purity level of the .former being 99.8 per cent Bi and the latter 99.9 per cent Sn. The material used for investigating the pressure-ductility response of pure bismuth was of high purity (99.999 per cent) 4 in. polycrystalline bar stock. Alloys were prepared by melting in a Pyrex crucible open to the atmosphere and using SnCl, as a tlux. Wherever possible, bismuth was added to the tin to minimize oxidation. Fabrication conditions for bar stock are given in Table 1 and the specimen annealing schedule in Table 2. All temperatures listed were held to within 12°C. All specimens were packed in fire clay powder prior to annealing. Specimens were cylindrical in shape, having a nominal gauge length of 0.562 in. (the shoulder to shoulder distance) and diameters of either 0.160 (pure bismuth) or 0.090 in. (tin-bismuth alloys). The apparatus used in the course of the investigation was a 30 lib Bridgman-Birch hydrostatic system Khich has been previously described;(‘) pressure measurement was accomplished through the use of a manganin wire transducer used in conjunction nith a Foxboro recorder. The estimated error in pressure measurement was 53 per cent. Estension of the specimen was accomplished

‘2,

to 24C for 14 hr and

_-. _-.--

,

.

1

PRESSURE

Fm. 1. High pressure tensiie test fisture.

I

IO

)

!

3

’

J 20

(N/fix

lo”,

FIG. Z(b). Pressure-ductility curve Bi alloy.

for a 90s;

Sn-bal

FIG.

for a ;5:;

Sn-bal

RESULTS

EJfcct ofpre.swre cm dmtilitg and fracture appearance The effecta of pressure on t,he various materials esatied is summarized in the pressure versus reduction in area plots shoxn in Figs. Cfa-e). For simplicity, each of the materials will be separately considered.

The effect of superimposed pressure upon the tensile ductility of pure bismuth is shown in Fig. i’(af5 along with the data of Pugh.c4) It should be noted that a sharp ductility transition is observed for this material at a particular pressure which will be defined as the Brittle to Ductile Transition Pressure (BDTP) and wih correspond to the lowest pressure

2(e). Preswre-ductilitp curw Bi alloy.

NOTE: SPECIMENS IN THIS RANGE ELONGATE TO THE LiMtT OF TRAVEL OF THE TENSILE H5LDER WtTXOUT FRACTI;RE

PRESSURE

(N/m2

FIGURE

X10’)

ZA

Prew_nv-ductility curve for lOO?&Si, corn. paring reeulta for columnar (as cast) and equiased (extruded) microstructures to data derived from results reported by Pugh.!”

F’ra.

PRESSURE

2(u).

FIG.

Bi

tN/i’rfxl&

2(d). Pressure-ductilit~p curve for a 42th SS-bsl Results derived from da& reported by alloy. L&hits et CX~.““~are shown for comparison.

r

I

F I ?

1 4

L

t

PRESSURE k-x‘:.

I

f

I

IO

2 ,e) \ ,. Pressure-ductility

I 20

(N/m2XIO*)

curve Bi alloy.

for

a

9:;

Sn-bat

resulting in virtually 100 per cent reduction in area. It should be noted that the BDTP is highly structure sensitive. The BDTP was observed to be approximately ri -. 1tmp (6 kb-) w-hen the structure consisted of cquiased grains with an average grain boundary intercept distance of 0.33 mm. However, x-hen the material was in the as-cast condition the BDTP n-as obserr-ed to be approsimatcly 11.5 :: 10’S m2 !11..5 kb). In addition to being cast! this material had a si,gnificnntly larger grain size. the averagr :rain size being as much as one-half the specimen ctiniue;er or approximately 1.15 mm in some areas of tke specimen. Referring to Figs. 3 and 4(a)+ the fracture mode below the transition pressure is transgran!.&r cleavage. Numerous stable microcracks are present behind the fracture surface of a tensile specimen fractured at atmospheric pressure and the number of mechanical twins increased as the fracture surface is apprwched (Fig. 4b). As can be seen bcomlwiny Fig. i(a) with S(b), the macroscopic fracture lnode changes with pressure from a flat crystalline t;r-pc to ductile rupture. Carrful examination of the region in the ricinity of the tensile fracture surface rerealed that the predominant mechanism of crack nucleation \vas the intersection of twins with grain boundaries as seen in Figs. 3(a) and (b). Although the predominant crack in Fig. S(b) is not apparently associated with a tnin-grain boundary interaction, this is simply a manifestation of the polished surface not intersecting the point of crack nucleation as is schematically shown in Fi:g. 6. Increasing the level of superimposed pressure towards the BDTP results in a small increase in duct&t)-. The resultant mechanical twin density is increased, but There is a decrease in the number of microcracks

FIG. 4(a). The microstructure teded to failure at utmxphoric Polarizwl

of purr bismuth specimens pressure. .N x, Cnetched, Light.

FIG. 4fb). The microstructure of pure bismuth specimens texed to failure at atmospheric preaaure, 15 ..I. Cnetched, Polnrizrd Light.

5. The microstrucwre FIG _. .^ .

of a pure bismuth specimen Urucka furrned by the mtersectmn of turns with gram Twin-crack pairs felt to nucleate below the plane of polish Polarized

obscr-red immecliateIp behind the tensile fracture surface as seen in Fig. T(a). Finally at pressures abore the BDTP, cleavage q-as suppressed and the specimens failed by rupture, [Fi,.v Tbf. The eridence of mechanical twins is less and eren in tie\7 of the large plastic strains, the grains are equiased. This is indicatke that the high strains were sticient to CBWX post test recry3tallization prior to metallographic preparation.

tested to failure HT.atmospheric pressure shoring: ia) .-i ~. . . - 1 . _ . bounclar~ea (i5 x , L netehwi, Yolartzed Llghri. ancl !I>! bl; the mechanism outlitd in Fig. 6. (I00 :. . lenetched, Light).

for pure tin and for near-eutectic compositions TCtlt. small Bi additions severe& ion-ering the ductility of the high Sn allovs. C’ompsrison of the tx-o CLWXY shops a direct relationship to exist betx-een the effects of pressure and ~ol~~positioilon ductility. Those areas having high cluotiiit~ (pure Sn and near-eutectic compoGtions) correspond to the low xransition pressures, and compositions haring a lox inherent ductility required a higher Talue of superimpxed hydrostatic prejaure in order to fail by rupture.

The trentl in duct&t>- exhibited by various Sn-Bi For specimens of 90 per cent Sn-baf Bi tested to ali0J.s compositions at atmospheric pressure and the failure at atmospheric pressure? failure occurs along rariation in transit&n pressure of these compositions a section nominafty perper~~~~iIar to the Iot~~itudinal is slwvn in Fig. 8. Ductility is seen to be a xna~irn~~m axis ’ as the pressure increases, the appearance of the &actzwe correspouds to failure by- ductile rupture. I At atmospheric and low values of superimposed hydrostatic pressure. cracks are seen to originate along the interface between Bi particles and the Sn matris (Fig. 9), \vhereas particle-matrix separation is not, seen in specimens tested closer to the BDTP. Sumerous voids3 hating the general shape of microcracks, were seen in areas adjacent to the fracture PLANE OF POLISH -i surface at atmospheric and near-atmospheric pressures (Fig. lt!a), but not at pressures approaching the BDTP (Fig. lob). At atmospheric and low ~~~~~~ pressures, the mode of fracture appears to be mixed, with eridence for the occurrence of both charage at interphase boundaries and ductile rearing present I-KANEOF \c, , POLISH as seenin Fig. 11(a). As the superimposed hydrostatic : 1 pressure approaches the BDTP, the fracture surface PIG. 6. d schematic representation of the appearance gradually converts to a dimpled structure, Fig. II(h). of cracks without metnllographically observed obviolw means of nucleation. The plane of polish intersect3 the with these dimples decreaaiug in size at the BDTP as crack above the point of nucleation, and results in the shown in Fig. 11(c). structures seen in Fig. 3(h).

a I

(A) FIG.

of:

The microstrucbure of pure bismuth specimens which fractured at superimposed hydrostatic Dre~Sure-j fn) 6.3 x 10” S/m’. and (h) 14.2 X 1iW S/m’. 30 X, t’iWdwtl, Polnrizecl Light. T*:nslle aXis &cheated by arrows. i.

As compared to the former allo>-, this alto>- contains a higher density of bismuth particles along the Sn grain bound&es and cry&allographic planes. This results in a higher microcrack density and a greater propensity towards cleavage fracture. The changes with pressure are otherwise effectively the same. 42 prr cc& Sn-bal

Bi alloy

This near eutectic alloy composition WAS seen to contain voids in the interior of the specimen behind the fracture surface when tested to failure at atmospheric pressure (Fig. Ea). Of interest is the shape ofthe voids; the “V” shape indicates failure by shear,

,/’

‘@ :

‘. .+:

::

100

80

60

40

WT.%

Sn

FIG. 9(a). The microstructure of a 90°0 &&al Bi specimen where pressure st fracture equalist 0.3 y 10s S/m’. Sate the grain size and cracks ber;reen matriv and particies straddling grain boundsris 500 :c, Vnetched.

I _____

__.

xi

__

0

FIG. 8. The variation in ductility and ductility tmnsicion presswe as a function of composition.

Frc. 9(b). The microstructure ofa %I~, Sn-baIBispecimea where pressure at frueture equalled 1.35 .- IW Sim”. Soce the smaller grain size than that &oxn 5~ Fig. 9 (a) and lack of purticle-matrix separation. .Wlt _=:, CI’netched.

(A) FIG. IO. The microstructure of YO”’,, Sn-bal Bi specimens immediatelv behind the fracture surface in qecimens tested to failure at: (a) atmospheric pressure, nnd (b) I.-5 s 10’ S/m”. 50 x , Etched.

I

FIG. 11(b),

FIG. flfa). An electron fractograph of a SY’& Sn-bill Bi specimen which fractured at 1.05 IO” S/m?. Evidence of both ductile fracture and clearage is seen. mioh cleavage apparently following the inter&e between the Sn matrix snd Bi partieles which precipirated nlonf: crystallographic planes, 3250 Y.

(cl. ElectrotlfractographsofBOqo Sn-bal Bi specimens which had fractured at (b) 1.53 :-: 1G5 S/m’ and ‘(c) 3.2 :-: IO’ S/m” showing a dimpled structure typical of ductile fracture. 3000 Y _

(A)

-TENSILE

AXIS-

(8)

FIG. I_‘!&). (b). The microstructure of 42% Sn-bsl Bi specimens tested to failure at: (a) Atmospheric pre~ur~, showing void formation along the approximate centerline of the specimen. 50 x, Etched. (b) 4.3 x lo3 S/m’. showing ultimate fracture to occur by rupture. 50 : , Etched.

Bi FIG. 17(c). The microstructure of a 4-3”; Sn-bal specimen tested to f&we at 1.5 .x: 10: S/m’. Surnerous instances of voids occurring at particle-matrix interfaces and evidence of cracking of Bi-rich particle3 are seen. .X0 .s., cnetched.

yet the ultimate fracture is along a plane perpendicular to the longitudinal asis. Figure l?(c) shows cracks initiating within Bi particlesl and their link-up through the matrix at the edge of the specimen This behux-ior allows the fracture surface to exhibit macroscopic features typical of brittle fracture eren thou& the alloy is relatirely ductile. Figure E(b) contrasts the appearance of specimens tested to fracture abore the BDTP to those fractured at atmospheric pressure (~Fig.l?(a)). Here, failure is hy rupture with no T-ok13abserred along the gauge length. and

Specimen5 containing a high percentage of primary Bi :i.e. hpp~~reutectic allo>-s) failed by clenmge within thy Bi and ductile failure in the Sn at ~-alueS of sup~ritnpo+xl h$.rowntic pressure 1x10~~ the BDTP (Fig. 13) and b>- rupture at pressures abo~ the

BDTP (Fig. 14(a)). Also observed KG a change in microstructure in regions of the specimens which had undergone more than minimal deformation prior to fracture. The microstructure shon?l in Fig. 13 and the post deformation zone area of Fig. 14(b) is representative of the microstructure of this compositiont with the Bi grain structure resulting from preferred rain orientation. This “oriented” struct’ure of Bi grains within the primary Bi stringers u-as destroyed in areas adjacent to the fracture surface (Fig. 14a). and clearly indicated the apparent boundar>- betxeen deformed and underformed material shown in Fig. 14(b). The &ml grain structure of these deformed areas ws not resolved, an effect felt to be due to recr+allization of these same areas.

(RI

-~EAS~LE

Ax15-

131

The microstructure of 9’); Sn-bal Bi specimens tested to fracture at 15.0 x 10%S/m”: ia) showinlr failuxe bv rapture, 59 x , Unetehed, Polarized Light. tbl same soecimen as in (al 1 , abox-e. but area shown is ok! II u ~: materiafvin the &init~- of the shoulders. Sate to tram&ion from an indistinct to a prono~uxcett grain s:rwture at the apparent boundnry between plastically deformed an3 undeformed material, 50 x, Cnetched, Polarized Light. FIG.

14.

DISCUSSIOS

Ductility at atmospheric pressure

Inherent ductility at, ambient, temperatures and atmospheric pressure is 100 per cent for pure tin and rirtually zero for pure bismuth. As an initial assumption, it would appear reasonable that ductility would be some cont~uo~~s function, decreasing as the volume fraction of Bi is increased. In contrast, the alloy sytem behares in a manner quite different from that expected. In prerious work x-ith tin-binary alloy having up to 20 wt.:; Bi, Xldentn) noted that specimens were somewhat brittle when annealed at ib-lOO”C, which he attributed to the embrittling effect of Bi particles along gain boundaries. It XT-asobserred in this current study that considerable Bi precipitation also occurred along specific crystallographic planes of the Sn phase. The IOK ductility of Sn rich alloys is a manifestation of the presence of these Bi particles and results in the foliovv-ing fracture sequence. First, particle-matris separation occurs along favorably oriented grain boundaries. As the tensile stress is increased, these separations join to form an intergranular crack. Further propagation then occurs by either further grain boundary fracture or (more probably) by travel along a favorably oriented collection of precipitate particles situated along some cystallographic plane in the tin. Eventual link-up of erncks to form the macroscopic fracture surface would be aecornpl~~~let~ ty the rupture of “bridges” of &-rich solid ~otution. The decrease in ductilit>xx-211incresing Bi concentration in the Pn.rich alloys

is attributed to the closer spacing of the Bi particles, which results in a more continuous plane of weakness, As the composition is changed so as to approach the eutectic composition, a substantial rise in ductility is observed. This rise, which is on the order of &3 per cent, appears to be due to a decrease in FGn size. Grain size ~as determined through linear intercept methods, aud the arerage grain diameter of the 90 per cent Sn alloy was approximately 1.2 :< 10-r cm. In the eutectic structure, the interphase distance was approximately 7.6 x 10-j cm. The difference in grain size (which is of sereral orders of maetitude) is considered the primary contributing factor for the disparity. Cracli initiation occurs by both the cleavage fracture of the Bi and failure at the Sn-Bi interface. However, even though there are more aI-ailable initiation sites in eutectic ‘z-s the Sn-rich allo’;3, the mean free path before the cracks encounter the ductile Sn phase is much less. Thus. more strain is required to extend cracks through the Su. The rapid drop in duct&t>- at Bi levels abo-i-e the eutectic composition results from the controlling efTect of large areas of the brittle Bi phase. The most probable mechanism for crack nucleation is the intersection of wins xith ,~ain bnundaries (i.e. the same mechanism as obserred in pure bismuth), \vith a possible assist from grain boundary sl.idin,g, as T = O.%3T,,. Although eridencr of cavit_v formation along gain boundaries is indicative of grain boundary slidiugt none Trere seen, nnd rareI_c- was nietallo~~~~~ic eridence of separation along grain boundaries obserx-ed. This Ii-t& of evidence kada to

the opinion mode

of

that

rain

crack

mechanism

to

intersections

boundan-

nucleation? be the

of txins

s1idin.g is a minor

and

the

nucleation

predominant

of cracks

at the

dimplin_rr, xx-ith the size of dimples decreasing

and the

hequency

prewre

of occurrence

increasest

indicating

be to reduce

and grain boundaries.

Duct ilit>- in terms of reduction of prewtre pure

Bi

at ambient,

and

Sn-Bi

ctara reported

alent

levels of ductility

in area as a function

temperatures

alloys

Khile

in

pure Bi. reported

Figs.

to occur

The anomalous ),J. Lirshits

character

cast

used

would

could

allow

their

Superimposed

hydrostatic stage

small if not negligible of a superimposed acting

shifts

pressure

nation

ail1 not affect

fracture

except

in a.

were to result in counter-

the normal

pressure,

impaired.!‘)

crack

stress b- the ma@-

propagation

This \vould effectively

lar fracture

and cleavage.

retard

of Fig. 7(a) that concomitant

decrease

in the

proposed

that

the application will result allouing extent

number

the retarding

of a superimposed

deformation not possible

enhancement,

of deformation

-grain boundary hydrostatic roids

and a It

by

pressure

obserred

to proceed

by

to an

pressure.

some eridence of bismuth

is

Also, for the

by slip and

to superimposed

pressure.

an effect forming

of pressure

on

would

a microscopic

be to

scale.

collapse

Since

the

promote for

for

defects

(such

as an indication

to impurity

of the BDTP

the

the

renlelting

and

apparent

that

oxidized

conditions

haT-e a large effect

nit,11 respect

of

areas

introduced

estrusion.

parameters

xould

cast)

and possible

effects

su~JsecjUeIlt

initial

(as

the effects

v s equiased)

as internally

levels.

to 11.3 ‘<

as-received

to indicate

to any chemical

readily

IJ)-

by Pugh for a bismuth

BDTP

(columnar

in addition

It

describing

is the

to purit!_ and structure

on autjsequen:

mechanical

behavior. awhile thinning

is recognized

of deformation

the suggestion

decreasing

propagation

sliding n-hen esposed

casting

with an increase obserred.

at atmospheric

in the literature

microstructure etc.),

of the

Lvhich is taken

component

hydrostatic

reported

alloytg) and is taken

(11.5 lib)

h_vdrostatic

in ductility

by twinning

108S/m2

intergranu-

of crack

in the increase

exists

Also.

of cracks

to

6.5 x 10RS~m2 (6.5 kb) is approsimatel!

sensitirity

from esami-

It is apparent

tend

on shear strain.

9 x 10JS/tn2 (2 kb) higher than a value for the BDTP

mode

xvould be

in pressure is an increase in the number of thins

there

in

way. (~1 Howe\-er, if the effect

(i.e. decrease) of the

pressure

of clearape

dependent

the

stress contri-

It should be noted that the value for the BDTP pure bismuth

materials, in

the tendency

equal to two-thirds

which

Also of note is the shifting

be present

observed

fracture

curve

et ,Z.(lO’ in their

nith

would

to

of the

to occur.

the initiation

tude

n-hich

the

coupled

to the

(Fig. S), small

still

by reducing

butions

failure by mechanisms

derivecl from results

due

of pressure

probably

void.; at pressures

2 per cent silrer

However.

the effect

the normal

(c).

with composition

as the

yield stress and to decrease

alloys as

temperature

by Livshits

diffusion

than for

et al, (1*’ which was ascribed

differences,

alloys

and

of Sn-Bi

response of ductility

investigation, ductility

behavior

to occur for this alloy system

composition

for

at lower pressures,

was not duplicated.

to thy sensitire observed

2(a)

alloy rather

of the pressure-melting

fur this alloy’l*)

is shown

by Pugh(J) appears to show equir-

the data given u-as for a Bi-Ag

the

to collapse

flj .s~~pe.rimpo.s~d hydrmfatic pressures

that

void growth,

rate of vacancy Efjxts

increasing

of the

pressure

pressure,(l~~‘~J

rj6~61.‘~)that superimposed

will

affect

basic

the

Griffith

normal

stress

relationship(lgJ

b?;

the normal tensile stress by the magnitude

pressure,

crack propagation is considered

thereby

impairing

by clearage

obeys a critical

Therefore,

since

deformation

rise to an increase

for pure bismuth

normal

stress to fracture of micro-

by application

of a super-

pressure,

b- twinning

modes.

relevant

the propagation

cracks would be suppressed imposecl hydrostatic

or suppressing

or intergranular

to be especially

as bismuth laxv.(‘O’

at atmospheric

by Davidson

of’ the

as the predominant

the propensity

towards

should be enhanced,

in ductility.

At pressures

giting greater

than .5 x 10sX/m2 (.5 kb), grain boundaT

effects and

slip have been observed

It is quite

pOSSible

in pure Bi.(‘“,‘l’

the effect of pressure on the deformation

that

characteristics

of bismuth,

when combined

with the

preswre to collapse a void in a given material is given as approximately tx-o-thirds the yield strength”” of

effect on the normal tensile stress component and subsequent effects on crack propagation. gives rise

the material

to the abrupt ductile-brittle transition observed. The ductility response of dn-Bi alloys subjected

to

superimposed

to

and the tensile strength

as 1.4 :,: 104S/tn2 that

pressure

(2100 psi).

v-ould hare

growth

and

of tin is given(13J

it n*ould be espected

a strong

coalescence

effect of

on the for-

x-oids

in

the

mation, 611 matrix. Fractogaphs. Figs. 14(b) and (c). taken at the apparent ultimate fracture surface shox

hydrostatic

be a manifestation transition near

is considered

of the shift of the brittle-to-ductile

temperature

atmospheric

pressures (BDTT).

pressure.

this

Ho-xerer. effect

may

at

or

not

be

DEIMBOWSKI

et at:

PRESS‘C’RE

ISDCCED

seen as it may be masked by recrFtaU.ization or high temperature (>O.fiT,) rupture phenomena. CONCLUSIONS

1. The brittle-to-ductile transition pressure for pure (99.999 per cent) bismuth is approgimately 6.5 x 10W/m2 (6.5 lib) and is a sharp rather than gradual transition. The transition is sensitive to grain size, with the transition pressure increasing with the grain size. 2. Fracture in pure bismuth occurs by transgranular cleavage, and is nucleated at the intersection of twins and grain boundaries. 3. The effect of a superimposed hydrostatic pressure is to suppress fracture propagation, and enhance the operation of mechanisms of deformation (i.e. slip and twinning) and ductile fracture. 4. The low ductility of Sn-Bi alloys having a small volume fraction of Bi is due to fracture nucleating by the separation of the Sn matrix from the Bi particles situated along the grain boundaries and propagating crystallo~aphic planes. The ductility of hypereutectic Sn-Bi alloys is controlled by the Gacture behavior of primary bismuth. 6. The brittle-to-ductile transition of Sn-Bi alloys exposed to a range of superimposed hydrost,atic pressures is a logical consequence of a temperature induced brittle-to-ductile transition. ACKNOWLEDGEMENTS

The authors would like to acknowledge financial support of this project by the Army Materials and Mechanics Research Center, and the able assistance

DCCTILITY

TRlSSITIOSS

IX

BI

1131

and skill of Xs. Theresa Brassard in the metaliog.. raphp of the pure materials and alloys used, and Xessrs. Leo MacNamara and Harry Kazarian for their aid in electron fractogmphy and the conduct of experiments. REFERENCES 1. P. W. BRIDGXU-, Research (Lond.) 2,550 (1949). 2. P. W. BRIDGXAX, J. appt. Phys. 24, 560 (1953). 3. B. I. BERESSEV, L. F. VERESEAGIX, Yc. S. RE’IBISIS and L. D. LNSHITS, Some Probiems of Large P&tic Beiefomation of Xetals at Hiah Presswes. Transiation by-V. 31. XE&OS. JIacrniUan:Sea York (1963). 1st Int. Corif. on Jfetds, Phils. 4. H. LL. D. %GE, pa. (1964). Annual Winter Xtg., S.Y. o. -1. BOBRO~SXY, ASXE XOV. 29-Dec. 4 (1964). 6. J. R. GALLI and P. GIBBS, dcta -let. 12, 77.5 (1964). 7. T. E. DAVXDSOX.J. C. UP and -4. P. LEE. &ta Xet. 14,937 (1966). 8. XI. BR~XDES, The i%lechanical Behavior of Xaterials Unde+ Pressure, edited by H. Lt. D. Pvca, Chap. 6, Elsevier, Sew York (1970). on Irreversible Effects 9. H. Lt. D. PCGH, Sympo&m of High Pressure and Temperature on Materials. _U.T.>l. February 1964, Special Technical Publication So. 371. 10. L. D. LIVSBITS, B. I. BERES~;EVand Yo. S. Rmarsrs, Xaulz, SSSR, 161, (5), 1077 (1965). 11. T. H. SLDEX, dcta AMet.15, 469 (196if. to #he Theoy 12. 0. HOFFxhs and G. Sacrrs, Intr~~t~on of P.&&city fov Engineers, pp. 69-74. fIcGraw-Hill, Sew York (1953). 13. Xetals Handbook, 8th ed. (1961). 14. E. G. POXATOWSXY, Fiz Xetal, i XetaEZaved16,622(1963). 15. D. FRA~COIS and T. T. WILSHAW, J. appl. Phys. 39 (S), 4170 (1968). 227 16. T. E. DAVIDSON and C. G. Ho~L~s, Trans. IfME (l), 167-176 (1963). (193”). 17. H. J. GOCGH and H. L. Cou,J. Inst. JfetaZs48,Zi Sature 136, 9 1.5 (1935). 18. W. F. BERG and F. D. ST.U~TOFF, 19. A. A. GRIFFXTH, PTOC. 1st Int. Cong. Appl. Xech. pp. 55, Delft (1924). 20. >I. GEORCIEFF and E. SCHYID, 2. Phys. 64, 845 (1930). 21. T. E. D~T;IDso’?~,J. C. UY and A. P. LEE, Tram AIME 233, 820 (1966).