DEVELOPMENT PROCESSING AND I DEVITRIFIED TITANIUM..(U) … · 2017. 3. 17. · l d-a149 699 alloy development processing and characterization of 1/2 i devitrified titanium..(u) northeastern

L D-A149 699 ALLOY DEVELOPMENT PROCESSING AND CHARACTERIZATION OF 1/2I DEVITRIFIED TITANIUM..(U) NORTHEASTERN UNIV BOSTON MRI BARNETT INST OF CHEMICAL ANALYSIS. S H WHANG DEC 84

UNCLASSIFIED N4--82-K-959F/G 11/6NL

EE~hhuEEILm

11111 l11U

JJJ1.15 1 A IIL.

MICROCOPY RESOLUTION TEST CHARTNAT IONAL BUREAU OF STANDARDS -963 A

% % %.......................

~~ N_ I- e...

RFP)n~CF~AT GOVERNMENT F)(PFNSE

a P-O VELOP'' 1E PRCESS ING AD) Q.RACTER ZAT~ ON

D)EVITRFEfD TA4II BASE O PLOY

ANNUAL REPORTI

0 ~ CTMCT -M42K 9

OFF ICE CF: NAVAL. :SEARCH

RNGTN VA 22217

MC~wER,1984

BA~ ~INSTTUT OF: Oe41CAL A4LYSIS AND W4ERIALS SC'EM~

BARNET UH1ERSITYD oSTON, OZ ~115

DTICJA 5185DThis do::~~thas been ppov

Ict p. r.~j~nand sale; its

85 01 08 0)59

SECURITY CLASSIFICATION OF TNIS PAGE (When Dots Entfred)

REPORT DOCUMENTATION PAGE REAO ISTRUCTIONSI. oREPORT MU. I, ORI. REORT NU,. 2 N, S3. RECIPtNT'S CATALOG NUM.ER

4. TITLE (and Subettle) S. TYPE OF REPORT & PERIOD COVEREDALLOY DEVELOPMENT, PROCESSING AND CHARACTERIZATIO1 Annual ReportOF DEVITRIFIED TITANIUM BASE MICROCRYSTALLINE September 1983-August 1984ALLOYS 6. PERFORMING ORG. REPORT NUMBER

7. AUTNOR(o) S. CONTRACTOR GRANTNUMsER(s)

S.H. Whang N00014-82-K-0597 -. -U

S. PERFORMING ORGANIZATION NAME AND ADDRESS 10. PROGRAM ELEMENT. PROJECT. TASK

341 Mugar, Northeastern University AREA & WORK UNIT NUMOERS

Barnett Institute of Chem. Anal. & Mat. ScienceBoston, MA 02115

It. CONTROLLING OFFICE NAME AND ADDRESS 12. REPORT DATE

Office of Naval Research December 1984800 N. Quincy St. ,,. lIUMUER OF PAGESArlington, VA 22217 109

14. MONITORING AGENCY NAME & AOORESSIf dillerent from Controlling Office) IS. SECURITY CLASS. (of (his report)

Unclassified

IS. OECLASSIFICATIONOOWNGRAOINGSCHEDULE

16. DISTRIBUTION STATEMENT (of this Report)

This document has been approved for public release and its distribution isunlimited. Reproduction in whole or in part is permitted by the U.S.Government.

~.2.EICTE,

1. SUPPLEMENTARY NOTES

19. KEY WORDS (Coninue on reverse edo II neceeear and identify by block number)

Rapidly solidified Ti alloy, arc melt spinning, rare earth dispersoid, Ostwaldripening of rare earth dispersoids, age hardening in rapidly solidified Tialloy, alpha Ti alloys, high temperature Ti alloys

20. ABSTRACT 'Continuo on reverse side It neceeemy and Identify by block number)

Principles that govern metastability in rapidly solidified Ti alloys have beeninvestigated with respect to solid solubility and phase transformation. Sub-sequent heat treatment in ternary Ti alloys results in precipitation reactionbetween group liA or IVA elements and rare earth metals in the absence of Ti.The Ostwald ripening of these binary dispersoids (rare earth metals Y, La, ErAl or Sn compo',nds) has been and is being studied. Characteristics such as

. transient coal .ing behavior and very slow coarsening in these dispersoids

DD O ,, 1473 UnclassifiedSECURITY CL.ASSIItCATION OF -4iS PAGE When Date EnretrG

% % . .'.. ." ."'-'' - ' : ': ." " -"-.". -._--'-"--," ...--. " i .., '.*.",' ', ," '-i' ""'L%. % - ."-"- ."" ."- '"

DEPARAMENT OF DEFENSE FORKS

DD Furm 1473: Report Dexcumettiation Pufer

ICWumTV CLAIMPICAT-O O@ ?f e C1P -tins

were identified. Mechanical properties (hardness) of al-lay ribbon and splat were studied through isochronal andisothermal heat treatment. The alloy ingot was con-solidated from alloy flakes by HIPing and is being studielwith respect to mechanical properites.

Docuo"TT~~~~~~.- CIA04~~no "g Aav-a -

_7-

TABLE OF CONTENTS

ABSTRACT

1. SUMMARY OF PRINCIPAL RESULTS 1

2. RAPIDLY SOLIDIFIED Ti ALLOYS CONTAINING NOVEL ADDITIVES 4

3. RAPIDLY SOLIDIFIED Ti ALLOYS CONTAINING METALLOIDS AND RAREIEARTH METALS -THEIR MICROSTRUCTURE AND MECHANICAL PROPERTIES 12

4. THERMAL STABILITY OF PRECIPITATES IN A RAPIDLY QUENCHEDTi-Al-Si ALLOY 20

S. MICROSTRUCTURAL CHARACTERISTICS OF RAPIDLY QUENCHED ALPHA-TiALLOYS CONTAINING La 28

6. SECOND PHASE COARSENING IN Ti-SSi-4.SL& SYSTEM 67

7. PILOT SCALE ARC MELT SPINNING PROCESS FOR Ti RICH ALLOYS 72

8. MICROSTRUCTURES AND AGE HARDENING OF RAPIDLY QUENCHEDTi-Zr-Si ALLOYS 77

*9. PRECIPITATION HARDENING OF RAPIDLY QUENCHED Ti-Zr-B ALLOYS 92

Ac.c .ession- For

NTIS GRAIDT: TAB

L Distribgatlon/

AvalIability t'o4Og

* ~AvaJ! 4nti/ar *

Dist Spf-!I A I

ABSTRACT

Principles that govern metastability in rapidly solidified Ti alloys

phave been investigated with respect to solid solubility and phase

transformation. Subsequent heat treatment in ternary Ti alloys results in

precipitation reaction between group IIA or IVA elements and rare earth

metals in the absence of Ti. The Ostwald ripening of these binary

* dispersoids (rare earth metals [Y. La, Enl - Al or Sn compounds) has been

- and is being studied. Characteristics such as transient coarsening behavior

-~ r and very slow coarsening in these dispersoids were identified. Mechanical

* properties (hardness) of alloy ribbon and splat were studied through

- isochronal and isothermal heat treatment. The alloy ingot was consolidated

* from alloy flakes by WIing and is being studied with respect to mechanical

S properties.iiio................................

r.67

SUMMARY OF PRINCIPAL RESULTS

The scope of rapidly solidified (iS) Ti alloy research at NortheasternUniversity covers three principal areas: 1) a study of heat transport inarc melting and casting during arc melt spinning of Ti alloys; 2)microstructural studies as a function of heat treatment, including

* - precipitation, and identification, morphology and stability of dispersoids:*3) an investigation of structure-property relations in RS Ti alloy at*i various stages of heat treatment.

PROCESSING

Rapid solidification processing (USP) by arc melt spinning hasadvantages over other techniques in many ways. It is contam nat on-free,inexpensively operated, and results in high quench rates (10V 10V [/sec).Since the heat transport and temperature profile of Ti melt in the coldcopper crucible are not well known, melting experiments in a cold coppercrucible by arc heating were conducted using Ti-6.3Si alloy. Thetemperature measurement at both the surface and the bottom of the alloybutton were performed by optical pyrometer and thermocouple, respectively,as a function of nominal input power. The results show that: 1) in themolten state, the heat transfer coefficient of the alloy at the interfacebetween the melt and crucible is -0.6 W/ca -K: 2) the temperature gradientin the melt reaches as high as -100 de/mm 3) radiation loss at the surfaceof the molten alloy is significant at this temperature; 4) the estimatedquench rate of Ti lloy ribbol and splat from the heat transfer coefficientranges from 5 x 10 to 5 x 10 1/sec.

Currently, one charge of 20g alpha-Ti alloy is routinely processed intoribbon. Continuous melting and casting requires preheating of feed alloy.The spinning unit for this purpose is planned for modification in the fiscalyear 1985/1986.

MICROSTRUCTURE

a) SOLID SOLUBILITY BY RAPID QUENCHING

Increased solid solubility of alloys by rapid quenching is widelyknown. In binary Ti alloys, the extended solid solubility variessignificantly depending upon phase diagram features. More specifically, thesolubility of solute increases significantly in alloy systems having -

eutectold transformation such as Ti-Cr, Mn, Fe, Co, Ni, etc. By contrast,. - no distinct increase in solubility due to rapid quenching is observed in

alloy systems having peritectoid transformations except for B and C.Examples are Ti-Ge, Ti-R.E. (where R.E. is Ce, La.... ). A large solubilityof B and C in Ti is a result of interstitial solution, but some questionsabout the occupying lattice site have been raised since the increasedlattice parameter of Ti-B is smaller than would occur when all boron atomsoccupy interstitial sites. On the other hand, rapid quenching stabilizesphase against a phase in 0-isomorphous type phase diagrams. For instance,as in Ti-Mo, Ti-Ta, Ti-Nb, etc.

b) PRECIPITATION

Super saturated solid solution of additive elements results in

-. 1

.' ? %p%~ ~ .;.. ~ ;~ ~ *.* . *-. * *. ~ .... *

K o

precipitation upon heat treatment. The precipitates in ternary Ti alloys(under high vacuum) take non-Ti, non-oxide binary compounds. Dispersoids inTi-SSn-4.SLa were identified as La2Sn (B82)• Currently, precipitates in Ti-5Sn-3Y, Ti-SSn-S.4Er and Ti-2.SAI-S.4Er alloys are being investigated.Experimental techniques for identification includes carbon extractionreplica, diffraction ring pattern, micro-diffraction and x-ray diffraction.

c) PARTICLE COARSENING

In the past, qualitative observations have shown that particlecoarsening of rare earth dispersoids in Ti is slower than in metalloiddispersoids. Present investigations aim at exploring the kinetic aspects ofparticle coarsening in ternary Ti alloy systems. Some of the experimentalresults are listed in Table 1. The low coarsening rate of the precipitatesin Ti-SSn-3Y and Ti-SSn-4.SLa are distinctive, which indicates that the mlower diffusivity of rare earth metals is responsible for the slow

* coarsening, provided the diffusivity of Sn is similar to the self- Z---

diffusivity of Ti at 8000C (they are virtually the same above 9S0°C).Nevertheless, a high coarsening rate in a Ti-Al-Er system appears to

r- contradict the foregoing results. In order to clear up the confusion, it isimportant to investigate the possibility that a coupled diffusion betweenrare earth metals (Er) and alloying elements might operate. Currently, theTi-Sn-Er system is under study for this objective. During the study of

• particle coarsening in the Ti-Sn-Y and Ti-Al-Er alloys, it was found thatnone of the single diffusion mechanisms can explain the transient coarseningoccurring in the Ti-SSn-3Y and the Ti-2.SAI-S.4Er alloys at 8000C.

I Nevertheless, the polynominal equation (an, where n - 1, 2, 3,...) can fitthe coarsening curve. In addition, the microstructure shows that in the

.- early stage of annealing a large fraction of precipitates is associated witha subgrain boundary. Further detailed studies are underway to understandthe diffusion mechanisms involved in this transient coarsening behavior.

In summary, current results suggest that further investigation isI needed in: a) the controlling mechanism of rare earth particle coarsening,

i.e., rare earth diffusion control or a coupled diffusion mechanism; b)diffusion mechanisms involving the transient coarsening behavior beinginvestigated; ) the prime reason for the discrepancy in coarsening ratesamong rare earth metals.

MECHANICAL PROPERTIES

Emphasis is placed on mechanical behavior of alloy ribbon at the agedstate and the HIPed from flake materials.

Age hardened ribbon alloy exhibits a relatively moderate increase inhardness in contrast with splat alloy. This appears to be associated withless fine microstructres resulting from the low cooling rate, which in turn

. has to do with the 'lift-off' in the substrate-side ribbon surface.Presently, produced ribbons were chopped into pieces 2-3mm long and

HIPed at 9000C. The mechanical properties (hot hardness) and themicrostructures are being characterized.

. . . o

oo.'..'o.'o•'o~'. -o * .°

". " - °',.° . .. '.°°° '°°o = "%'., ,°o.- o°O- '%.. . . . . . .. . . . . . . . .- o. .o...... . . . .o.-.. . . . .- o. .o. .o. .-.-. .

Table 1

Alloy System Coarsening Rat: at Compound EstimatedAt. % 00 naphsStutr dif io

coefficiejt at8000C(c /$ )

Ti-5Sn-3Y -5 z 02 under understudy study

Ti-SSn-4.SLa -9 x 02 La Sn -4 x 10-14 (La)

_8 z 0-26(B12)

Ti-2.SAl-S.4Er aunder understudy study

*Ti-SAI-2Si -3 x 1-6T Si-1 x 10-11 (Si)

I3

Coarsening rate W r ro 3)/t where r and t are average particle ridiun andannealing time, respectively. Self-diffusivity of Ti is -2 a 101 cm /sec(8000c).

ii3

I RAPID SOLIDIFIED Ti ALLOYS CONTAINING NOVEL ADDITIVES

S.H. Whang

Journal of Metals, Vol. 36, No. 4, 1984, pp. 34-40

Journal of Metals; Vol.36, No.4, April 1984, pg. 34-40

Rapidly Solidified Ti Alloys ContainingNovel Additives

6S. H. Whang

SUMMARY

The rapid solidification of Ti alloys by an arc plasma melt spinning •technique is reviewed. Attention is given to the melt spinning technique, andguidelines for the selection of alloying additives are presented. As a newapproach for Ti alloy processing, the devitrification of Ti-rich glasses isintroduced and discussed. The microstructures of several rapidly solidifiedTi alloys which remain crystalline are described for the as-quenched as wellas heat-treated conditions. Preliminary results on the age-hardening re-sponse of rapidly solidified Ti alloys are also discussed and their thermalstability is reviewed.

INTRODUCTION

In recent years, there is an increased interest in the development ofhigh-specific-strength and high-temperature alloys through rapid solidifica-tion processing (RSP). RSP offers a distinct advantage: small amounts ofelements that normally have nearly no solid solubility in titanium at room Stemperature can be homogeneously dissolved in titanium through RSP.Subsequent heat treatment of such alloys can form precipitates which giverise to additional strengthening at moderate or, depending on the precipi-tate stability, even at highly elevated temperature. As a result, RSP offersalloy designers more flexibility than ever before to control microstructureas well as to select alloy compositions.

Rapid solidification of Ti alloys has been previously studied using meltextraction techniques based on the conventional alloy compositions withoutadditives.' However, new approaches for the syntheses of rapidly solidified(RS) Ti alloys include the devitrification (or crystallization) of glassy phase 2

and the rapid solidification of the molten alloys containing metalloids 3"5 orrare-earth elements.6 .7 Upon heat treatment, these alloys yield various --.

second phases which comprise of compounds based on (a) Ti-metalloids(B, Si), (b) rare-earth metalloids, (c) rare-earth group III or IVA metals, and(d) rare-earth interstitial (C, N, 0). Microstructures and mechanical proper-ties of rapidly solidified Ti alloys depend significantly on the type ofsecond phases. Aspects of RSP of Ti alloys as well as some current resultsof such processing are reviewed in the following.

A Arc Melting Comptmef"tWalw' ~~ FOK ill o Spinningl Compartment ".NW"-Ct e C RIon Pr4

vessO Chembe c MELT SPINNING TECHNIQUE

Regulawet-. . While some of the techniques for rapidly solidified Ti alloy processingAuto-Feed have been previously summarized,4 the technique of melt spinning of Ti

'rA. © * lvle alloys is new. The advantage of the melt spinning processes lies in the factArc P'.u -r - that extremely refined microstructures of good uniformity can be attained

Solenoid Vetv• . through the high cooling rate derived from the geometry of the ribbon.w~t Inert :aPreviously, many different melt spinning processes for Ti alloy have beenWoler Cool ed, Ri available; these include the pendant drop melt extraction, the crucible melt

- .extraction, 8 and the arc plasma melt spinning.9 The extraction processesCCo Ble .ush. are different from the arc plasma process in that the extraction spinning

depends on a natural contact mode (i.e., "free fall and contact") in thedrop melt extraction and a "contact and drag" mode in the crucible melt

. ,mp extraction. This contrasts with the orifice-shaped, high-velocity molten jet" - in the arc plasma melt spinning. Thus, instead of a half-circled wire, the

product in the plasma melt spinning is a thin ribbon. A pilot-scale, arcFigure 1. Schematic drawing of pilot arc plasma melt spinning unit is shown in Figure 1 and consists primarily ofplasma melting spinning unit. three functional parts: melting compartment A. melt spinning compart-

ment B, and ribbon processing chamber C. The pre-melted alloy ingot is . - "charged directly into a cold copper crucible while additional granule alloy -. - .is continually supplied by an auto-feeder to maintain the molten alloylevel. When the melt is adequately superheated. additional inert gas is -.-introduced into compartment A creating a pressure differential Detween

34 JOURNAL OF METALS* April 1984 . S

.• . . . °- -

- _, _ _ __ ", -", "_ _ -- . . _,..-. .". .. "-".. .,- . . .,"' ' ' -.-.-.. ,-.-, .' ' ' ..-.-. '. ' ' . ' " ' -.-. ,

the melting and the spinning compartments. As a result, the molten alloyforms a jet through the orifice centered at the bottom of the crucible. Thejet subsequently impinges onto the highly conductive spinning disk under-neath the crucible and rapidly solidifies into ribbons or flakes depending *on the disk employed. The ribbon or the flake can be further processedinto particulates by means of mechanicdl pulverization in the chamberbefore discharging or canning under the protective environment. Detailed 'W

description of the pilot-scale arc melt spinning process is given elsewhere.' 0

ADDITIVE ELEMENTS

It is well known that a small amount of Si additions to Ti alloysimprove their creep resistance. 11 In RS Ti alloys, other metalloids andrare-earth elements can also be used as additives to provide the second 0.5,/1 "phases for high-temperature properties. Since extended solid solubility ofnormally insoluble element can be achieved via rapid quenching fromliquid, the choice of the additive element is quite broad, in particular, itincludes metalloids to rare-earth metals.

The benefits of these alloying additives in Ti alloys are twofold: (a) whileboth types of additives have very low solubilities at room temperature,they may be rapidly solidified into solution and (b) they both have thepotential of forming thermally stable cow.pounds. In fact, both metalloidsand rare-earth metals have similarities as well as dissimilarities in atomicnature; they have unfilled P shells and a large atomic size difference fromthat of titanium. The dissimilarities are that metalloids have a largeelectronegativity difference as well as high negative heat of solution intitanium, whereas rare-earth elements have the opposite characteristics. Asa result, the metalloid prefers to form compound phases directly withtitanium, but the rare-earth element tends to form an immiscible phasedue to the positive heat of solution. Alternatively, the rare-earth elementscan produce compound phases in the Ti matrix with other elements suchas metalloids (B, Si), interstitials (C, N, 0) or group ILIA, IVA metals (Al, bSn. Ge, etc.) Figure 2. Mlcrostructures of TleaZreAI 4Ni0Bi 2

There is also an important contrast between the solubility of the metal- glass after crystallization at 550°C/2 h; (a)loid and the rare-earth element upon rapid quenching. The metalloids (B, bright field and (b) selected area diffrac-C. Si) have an extended solubility range of 6-10 at.%, this is larger than tion pattern.the maximum equilibrium solubility at the eutectoid temperature.6 On theother hand, the rare-earth elements iLa, Ce) have a range of less than 0.6at.% which in turn is smaller than the maximum equilibrium solubility ata-transus.6 Microscopically, the small metalloid atom can occupy either theinterstitial site or substitutional site or both, while the rare-earth atomcan only replace Ti atom at the substitutional site. The combined effects of Althe unfavorable atomic size as well as atomic interaction make the rare-earth elements the least soluble elements in titanium. The morphology andthermal stability of second phases based either metalloids or rare-earthelements are an important basis for determining which type of additives ismost promising; this will be discussed later.

DEVITRIFIED (DV) Ti ALLOYS

The principal advantage of a devitrified (DV) Ti alloy (i.e.. one whichhas been quenched to an amorphous state and is subsequently crystallizedby heat treatment) is that extremely fine and homogeneous microstruc-tures can be generated from a glassy phase. The microstructures in DV Tialloys are genetically different from those in rapidly solidified microcrystal-line alloys. When the glassy phase undergoes crystallization, the initialcrystallized phase is that which has the lowest critical energy forcrystallization, not necessarily the one which has the lowest free energy.-2

For this reason, frequently unknown metastable alloy phases occur in theearly stage of the crystallization. As a result, the crystallization process inglassy metals often produces different phase mixtures from those in therapidly solidified crystalline alloys.

The synthesis of DV alloy must satisfy one stringent compositionalrequirements; the alloy composition must be glass-forming at the givencooling rate. Although glass forming tendencies in binary and ternaryalloy systems can be estimated using a phenomenological model." : '5 thefollowing empirical expression is useful in determining the glass formingability of a given alloy at the cooling rate of -106 Ks

Ti, 'Zr. Hf. V. Mo etc. ) iSi. B. Fe. Co. Ni. Cu etc.-where x - 50 at.1, y 30 at.%r, and z = 20-30 at. . As the cooling rateincreases, the z value can be reduced below 20 at."i. The elements in thefirst parenthesis normally have extended solid solubility in titanium while

JOURNAL OF METALS *April 1984 :35. . ... ... .. ' ' - : . .. ". .."- - -, - - '- '-' .' .' - ,- - -.- ,- -.- ' -. ' - .. -. _ . . -2 .i~ ii ;l . .- . :.

0~ O25kn'

d aOm~eL

Figure 3. Microstructures of splat-quenched the elements in the second parenthesis tend to form eutectics " ith titdiUIt.Ti alloys, (a) Ti-5A1-25Sn-1B, (b) Ti-5A-3La. Using the above relationship, alloys based on the late-transition metal-rich.(c) Ti-5A1-2.5Sn-3Ce. (d) Ti-18.MZ-1.31E. glass or metalloid-rich glass can be synthesized In either case. the cr\ ti.-(a) To-171Dr-3.3Si. and (f) Ti4Mo-2.3A-3.5Si. lization results in ver- high -vol ume fraction of one or nmore intermettalic

phasesAs. an example of a DV Ti allov. Ti,.Zr Al4 Ni: -B _ at fhas lev-n

quenched into glassy' phase and crystallized b\ the succeeding heat trtat-ment -'Figures 2a and b show% the microstructlure of the crvstaflhzed ah. -and the corresponding diffraction pattern in which split ring , derixed fronmthe typical amorphous diffused ring are distinctive The crystallized ri

sieat this stage is extremely% small, being less than 100i A dianhier,

RAPIDLY SOLIDIFIED (MICROCRYSTALLINE) Ti ALLOYS

RSP of crystalline Ti alloys offers improv-ements in mechanical proper--U ties based not only on the refinement of exis-ting ow- microstructurv-during rapid quenching but also on newk phase forniation through heattreatment Both the ability" to form finelyi disper.sed precipitate particle,and their subsequent highi-temperature stahilitY are e!,ertial t, rapidl%soilidified Ti allov d( ig-n licrostructural characteristics of ra;'tdl solidl-f-cl Ti all% l-ii h . ,nvenientl \ described in three difle-rent stage-

it- iut n4 xed, Rnn.-aled ai intermediate temperatures. arid atnnejeltd at H;gl.

Microstructures

As-Qidcn hed In general. microstructural refint-enti ol Ti dal' \ can battained through rapid quenching Furtherniore. a smai~l! am--ant of anlladditive element can alter micro.st-uctural characteristics fI th, a quLeni te~istate- Such additive elements exi-t in the ii-qui-nched stati I:t'. i tnt(

Oform oI a supersaturation solid S-d il in r In- t ht firM 01 - ,Itt C i-tt

Figures 3a. b. c sho\A micrograph- for ,-I I Hlt.IV- coritainirg B -A. anidttre~pcctivelk In all a~ lov-. a weli I tc r\ finec Hcica~ia micr- -T.

turt is- present A careful observat i-c. -, - that wAhen- the r.n i--n-rare-earth metail, exceed- the mavinun. exte-nde-d sbhhipr t-.i..occur- along at subh-r-ain homndir\ net a irk Thi- net,,k-rk .!-thoiindir it-- on!\ ohst-rxed it at. et ani-anit of tht- rar, *,-- t.t ti

4 ~~~i- pre-t nt The sub--grain sie( o- it'- 111r,,at-l ,i\ r-t.~ -'-! -

to- tht (-win-t 4 the rar, tarih t-,ni-t jr the Zr rill a. t i it

1\ refinied martensiti- Sttrtr 1 pre-, r!- atil I- -f.tli- 1it(

struCtUrt in the hackgcriund a- shan" Ii I-icui,- ill dIi) t H -at, Ii t

%1-- rich all,% 1igur( :if (-\hili - a it- cti -%rti !!i X-------------seilif of toattn-itt -itructor,

Annealing at Intermediate Temperature. When rapidly solidified Ti alloysare subjected to heat treatment at a range of 500-7000C, the resultingmicrostructural changes include G.P. sone-like precipitation and disappear.ance of both martensite and sub-grain boundaries. No apparent graingrowth can be detected at this stage. In the rapidly solidified alloyscontaining either metalloids or rare-earth elements, precipitation is a uni-versal phenomenon. The Ti-3La alloy annealed at 7000C/2 h exhibits cuboi-dal precipitates (Figure 4a) while the Ti-5AI-2.5Sn-3La alloy annealedunder the same conditions yields fine precipitates of spherical shape (Figure4b). These precipitates in the Ti-5A1-3La and Ti-5AI-2.5Sn-3La have beenidentified as AI4La (orthorhombic) and AI3La(CI5) from the selected areadiffraction rings and as a non-oxide form from EELS.6 Previously, thedispersoids in a Ti-rare-earth system was found to be rare-earth oxide, i.e.,Y20 3 in Ti-8A1-4Y (at.%),17 etc. In addition to the nucleation and growthof compound phases, both the fine-scale martensite structure and the sub-grain boundaries disappear as the grain structure becomes more distinctive(Figures 4c and d). At this stage, the particle shapes and morphologies canbe identified. For example, the boride particle has a needle or rod shape incontrast with the polyhedral or spherical shape of silicide and carbideparticles. However, all rare-earth precipitates in ternary and quaternaryalloys are spherical. These discrepancies are summarized in Table I.

Annealing at High Temperature. After exposure of the rapidly solidified Tialloys to high temperatures significant microstructural coarsening mayoccur (900-950*C). Extensive Ostwald ripening occurs in all of the metalloid-containing alloys (Figures 5a-d). In contrast, the coarsening of precipitateis minimal in the alloys containing rare-earth additives (Figures Se and f).In all cases, the grain size increases to 1-2 ;m diameter after the exposureat 900-950°C/2 h.

Depending upon additive element, the morphological characteristics ofthe precipitates become distinctively different after high-temperatureannealing. The silicide particle tends to coarsen along the grain boundarywhile the boride with a rod shape takes random orientation. The rare-earth precipitates remain relatively uniform in distribution but someprecipitate-free zone formation begins at this stage.

Mechanical Properties

Rapidly solidified Ti alloys can show a strong enhancement in hardnessin the aged condition even after high-temperature annealing. This is prima-rily a consequence of precipitation hardening, the stages of which aredescribed below for the rapidly solidified Ti alloys.

As-Quenched State. A significant hardness increase occurs after rapid quenchprimarily due to the supersaturation of additive elements and microstructuralrefinements. The supersaturation of 1 at.% metalloid in the Ti alloy in-creases the hardness by 4-13% from that of the base alloy. In contrast, the

TABLE I: Dependence of Particle Space and Morphology on Additive ElementsAge Hardening Particle

Additive Compound (Temperature) Particle Shape Morphology ReSi TisSi3 Yes (550-700°C) Polyhedral or Grain boundary 6

spherical precipitate andcoarsening

B TiB Yes (550-700°Ci Needle or rod Random 3C TiC No (700°C) Spherical Uniform 6Y Y20 3 Yes (500-700°C) Spherical or Uniform 17,3

cuboidalLa AI4La & Yes (500-700°C) Spherical Uniform 6

AI3La

TABLE I1: Mlcrohardnesa of As-Ouenched and Heat-Treated Alloys*

Alloy Composition, As-Quenched, Aged, High TemperatureW•% GPs GPa Annealing, GPa

Ti-5AI-2.Sn-3C, 4.3 5.8 (600°C/2 h) 4.0 (900°C/2 hItTi-SAI-2.5Sn-IB 5.7 6.8 (600°C/2 h) 4.9 (9000C/2 h)Ti-5A1-4Zr-2.5Sn-3La 6.4 7.5 (7000C/2 h) 5.4 (900°C/2 hITi-18Zr-4.4Si 5.5 5.6 (550°C/4 h) 3.9 (850°C/2 h)Ti-7.7Zr-3.4A1-3.6Si 4.5 5.6 (700°C/2 h) 4.7 (900°C/2 h)Ti-8.2Mo-2.3A-1.4B 4.8 5.4 (700°C/2 h) 4.5 (900°C/2 h) %Ti-6A-4V-2.2Si 4.1 6 4 (5000C12 h) 4.6 (900°0C2 h)Ta-SAI-2.5Sn 3.6 ......

*""sdohn sIst fois of 20 ,m thek

JOURNAL OF METALS April 1984 37

- - - -- |

b • .

K"

Figure 4. Microstructures of aged alloys;(a) TI-3La, 700rC/2 h, (b) TI-5Al-2.5Sn-3La, &

70C2 h, (c) TI-IS.2Z-1.3B, 550*/24h, * '.O2mr and (d) TI-i 7.1Zr-3.31SI, 550*C/20 h. *

c d

same amount of rare-earth metals causes an increase of 20-47%1 from thesame standard.s These large increases are proportional to the atomic misfitparameter of the rare-earth atom and Ti atom.6 However, when the concen-tration of rare-earth additives is increased above 1 at.%, the hardnessinstead begins to decrease.Aged State. The most important and universal phenomenon in rapidlysolidified Ti alloys is the presence of age hardening. A typical age harden-ing response is shown in Figure 6. The hardness increase per atomicpercent is much higher in the rare-earth-containing alloy than in themetal loid -containing alloy. As an exception, the alloys containing carbonapparently do not age harden (at least at 700'C (, even though a very largesolubility of carbon is observed as well as a large strength increase in theas-quenched state is noted. As is the case in the as-quenched state, agehardening is a function of an atomic misfit parameter in the alloyscontaining rare-earth additives (Figure 7).3 Isochronal annealing treat-ments of as-quenched specimens indicate a difference in behavior betweenalloys based on single-phase a of ar-0 Ti alloys. Specifically, over-aging--occurs at a much lower temperature in -0 alloys than in* a-type alloy(Figure 8). This is probably due to the fact that the high diffusi'vity (andlpossibly a higher solubility) of the additive elements in 13 phase accelerates -

particle coarsening. In other alloys, a large drop in hardness occurs at orabove 800'C. Also, it is observed that alloying elements such as zirconiumand tin are effective in delaying or impeding the over-aging process inrapidly solidified alloys.

High Temperature Exposure. As would be expected. annealing the age-hardened rapidly solidified Ti alloys at 900'C causes a decline in hardnessassociated primarily with the second-phase coarsening. and secondarilywith other microstructural coarsening. It is observed that the precipitatecoarsening varies significantly depending upon the type of alloying additive.The grain size remains sub-micron in most alloys even after 900'C 2 hThe hardness, which at this stage is still as much as 20O7 higher than that ofthe as-quenched condition, after aging, and after high -temperature anneal. i-

listed in Table 11

SECOND-PHASE COARSENING

The potential for developing high-temperature Ti alloys based on RSPrelies heavily on the thermal stability of the second phase derived fromthe alloying additives In Figure 5a-dextensive particle coarsening occurs

38 JOURNAL OF METALS April 1984

. .." .. ig r s .s s n in F e .. . .. . .

in the alloys containing metalloids, (C, B, Si) upon being exposed toanneals in the range 800-900"C. The coarsening is characterized by morpho-logical differences between silicide, carbide, and boride particles.

A study of the coarsening kinetics for Ti-5AI-2Si alloy has been conductedby transmission electron microscopy in the range of 600-800C. The aver-age particle diameter as a function of the annealing time for 700-800 C ofparticle diameter and the soaking time agree with Wagner's lattice diffu-sion model, but the data for 600°C is better described by a high diffusivitymodel.' s Similarly rapid coarsening kinetics at low temperature has beenpreviously observed in rapidly quenched Al alloy systems."9 It should benoted that lattice diffusion coarsening may occur in the bulk form ofthese alloys due to an absence of a high dislocation density. .. '

In contrast to the metalloid-containing alloys, Figures 5e and f show aresistance to coarsening in the rare-earth alloys. Therefore, at 900°C theparticle coarsening in the Ti alloy containing the rare-earth elements (La,Ce) appears to be lower by an order of magnitude than those of the alloyscontaining the metalloids (C, B, Si). As a result, particles of 1000 Adiameter remain stable (see Figures 5e and f).

EFFECTS OF COOLING RATE

It is well established that microstructural features such as dendrite armspacing, grain size, cell size, second-phase inclusion size, etc., are exponen-tial functions of the average cooling rate. Any enhancement of the mechani-cal properties is a consequence of the combined effects of one or more ofthe microstructural refinements. Therefore, the relationship between thetwo quantities, i.e., the cooling rate and the mechanical property is onlymeaningful under well-defined circumstances. The effects of the coolingrates on hardness in rapidly solidified Ti alloys are evident both in theas-quenched state and after intermediate temperature annealing. There isan approximately 10% decrease in microhardness from 20 ;m to the 40pm.6 The estimated change in cooling rates is a factor of two, assuming aNewtonian cooling in which the average cooling rate is inversely propor-tional to the sample thickness and directly to the heat transfer coefficientbetween the sample and the substrate.20 In the present case, the ribbon iscooled by a substrate from one surface and by a gas medium from theother surface. The gas cooling is three or four orders of magnitude smaller Figufe S. Mlcrostructurss of annealedthan that of substrate quenching. Hence, the cooling rate difference be- alloys; (a) TI-5A-2sn-C, 50C/2 h, (b)tween the ribbon materials and splat materials is roughly a factor of two TI-18.2Zr-1.30, 9O1C/2 h, (c) TI-17.SZr-.3SJ,for the same thickness. It is important to note that the thickness of the 9S0C/2 h, (d) TI-7.7Zr-3.4A1-3.6S, 950"C.as-quenched alloy ribbon thus is a critical parameter if a rapidly solidified 2h, (e) TI-SAI-2.55n-3La, 900C" h, (fM TI-5A-Ti alloy is to benefit from rapid solidification processing. 2.SSn- e, 900C/2 h.

"" t05/Am IL, ..- , O'm 5~ M .:.

.N F T Ar 8

a a %"' ,o m''

SJOURNAL OF METALS •April 1984 39.

CONCLUSION- ~ While preliminary results on some mechanical and microstructural prop-

erties of rapidly solidified Ti alloys appear promising there is no doubt- that the critical areas remain to be explored in the y'ears ahead. Critical to

""" the development of Ti alloys based on RSP are the following areas:

5 '50 1. Identification. morphology, and thermal stability of second-phase parti-6 cles used to strengthen the alloys.

~~~~2 Efs.ss.c fects of the solidification rate on the both as-quenched and annealed40 -5S.W.. icrostructures and the resulting mechanical properties in conjunc-

tion with the evaluation of various RSP techniques.30 3.TeaiiytOosliaerpdysldfe - losit ukfr

AO 1 2 4 0 20and to retain the enhanced me(' 'nical properties.7. TMEhour 4.The effectiveness of age hardening in structure-sensitive properties

Figure S. Age treatment at 600C of RS such as fatigue, fracture toughness, creep. etc.. of bulk materials.Tl-SAI-2.5Sn-1 8 and RIS TI-5A-2.5Sn-3Ce.

ACKNOWLEDGMENT

The major portion of this article is based on research sponsored by theTh -5.4-25S-~(Office of Naval Research. Contract No. ONR-N00014-82-K-0597. The author

RS 75AI25S-REthanks Prof. D.A. Koss. Michigan Technological University for many sug-.~' gestions in the preparation of this manuscript. The author is indebted to

a. Dr. C.S. Chi and Mr. Y.Z Lu for their enthusiasm and dedication to rapidly-3 La solidified Ti alloy research. The author thanks Ms. N. Barnett for the pre-

gt-2 paration of illustrations. Contribution * 194 of the Barnett IsiueoChemical Analysis and Materials Science.

- References0 1E W Collings. C E %tabley*. RE. Maringer. ani H L Gegel. "selected Propertis. if %felt.Fotracted Titanium'

Bai .. vrvtln Alloys. in Thir'd Int Conf Ropidl. Quenthed Mtal, Ill. edited bs B C'antor The Metakl

i-i Third Conf Rapid Solidification Prrcessuu., edited by R. .Mehrabian. NBS. 1982. pp '286.291)l3, C S Chi and S H Whang. "Effects of Novel Additives 1B. Cei in Microstructural and Mechanical Prope.rtiesin Rapidly solidified Alpha Ti Alloys." T.MS-AIME. Paper No Fl1.14. The Metallurgical Societv of AIME.

/Warrendale. Pennsylvania."RpdSldfctoPresig4Tanu40, 4. S.M.L Santy . TC Peng. PJ. Meochter. and JE. O'Neal. "ai oiiiainPoesn fTtnu

Alloys." J. Metals. 35,9,,19831. pp. 21-28.- -- -- -- -- -- -- -- -5 AGI Jackson. T F Broderick. F H. Frses, and J Moteff. *Effect of Aging in the Microsttructure of Rapidi -

>AQ T, - 5.1-2 55n Cooled Ti.5Al.2.5Sn with Si Additions." in Proc. Third Cowt Rapid Solidification Prossinx. edited hi R.Mehraian. NBS. 1982. pp. 585-89.

301 i C.S. Chi and 5.H Whang. "Rapidly Solidified Ti Alloys Containing Metalloids% and Roae Earth Metals-30 1 Their Microstructures and Mechanical Properties.' paper presented 19&.3 Annual Materials Research society*1 '0 II 1 1 3Meeting. November 14-17, 1983. Boaton. To be published in the proceedings

R a E/RT, 7 S.M.L. Sastrv. T C. Peng. and J E. O'Neal. "Consolidation. Thermomechanical Processing. and Mechanical -

Properties of Rapidlv Solidified. Dispersion.Strengthened Titanium Alloys,'in Prost Third ('on) Rapid Soihl ii

Figure 7. A correlation between microhard- cation Prnceooing. edited by R. Mehrabian. NBS. 1982. pp 579.5848 R. Maringer and C Mobley. "Casting of Metallic Filament and Fiber." J. Vac. Sci. Technology. 11 1974,.*ness and atomic misfit parameter In TI-5A1. pp 106741071

2.5Sn-2Y (-3Ce, -31-a). 9 S H, Whang and B C Giessen. 'An Arc Furnace Melt-Spinner for the RSR Processing of Refractory andReactive Alloy' s." in Pros' Third ('omf Rapid Solidificatiiin Prissing. edited by Robert Mehrabias. NBS. 19S2.pp 439-44210. SH. Whang. "Pilot Scale Arc Plasma Melt Spinning for Ti and Refractory Alloy." to be 'submitted toI ~ Materials Letters, in preparation

_____11 N.E Paton and M W Mahoney. "Creep of Titanium'Sil[icon Alloys." Met. Trans.. 7A 1976.fi. pp 1685-1694- ,. .~...*12 U Koster sod Ui Herold, "Crystilization of Metallic Glasser." in laos Metals 1, edited by Hf J1 iGutherdt

- - ~and H Beck. Springer'Verlag Berlin. Heidelberg. New York. 19181. pp 225-25913. S.H Whang. "New Predictin of Glass Formting Abilty in Binary Alloys Using a Temperature.Compoivitois~ ,, ~'Map." Mat. Sci. & Eng., 57,198:3,. pp A7.9.514 S H Whang. "Glass Forning Ability for Binary, Alloy Systems bv Modified T-C Map is Relation to Phase

* - .9 i' ..- '' s.'sI Diagram." presented at Fifth tnt Conf. Liquid & Amorphous Metals. August 1983 J. Non-Crvsi. s'ot..i . .*61 & 62. Part it. January 19S4. pp A41-846I , ' 15 S H Whang. 'Glass Forming Property 'if a Ti'Zr'Si System Determined by Tsemperature'('ompoceiiion Mlap.

'7 ""~~Z'""Scripts Met.. 18 4i '1984-a- . o- 16 H Jones. "'some Principles of Solidification at High Cooling Rates.'* in Pr- First 1st 'ott 'is Rapid

Solidification Processing. edited by R Mabrabian. B Kesr. and M1 Cohen. Claitors Publishing Di, Baton -

-. -~- 'a:~'sRouge. 1977. pp 28.4517 D G Konitzer. B C Muddle, and H L Fraser. 'A Comparison of the Microistrui'tures if As'i'ast and Laser -

________________________________ Surface Melted Ti.8Al'4Y." Met. Trans. A. 14A 19WI3. pp 1979-l9A85.~ 500~ "0 no18 Y Z Lu. S H Whing. and B C Gieson. 'Thermal Stability ofi Precipitate In Rapidly Quencnved TI -. Al.-S

'tiiitO05i '2Alloys." paper presented at Materials Research Society. November 19M3. Boston Tio be publi.hed is the

Figure 8. Mlcrohardness as a function of Ptceit5isocronl anea~ng(2 hurs inrapdly 19 1. Pontikakiw .ind H Jones. "Coarsening of Intermetallic Particles in Rapily Solidified Aluminum-Transition-Isocrona anealig (2hous) i rapdly Metal Alloys." Metall Science. 16 '1982'. pp 27.30

soldified TI alloys. 20 R C RahI. "Cooling Rate in Splat Cooling." Mat. Sct.. 1.1967,, pp 3'13l.1201

ABOUT THE AUTHOR

Sung H. Whang. Senior engineering at Seoul National University. thermal stability in A15 alloys, and metallicScientist. Barnett institute Korea, and D-Eng Sci in physical metallurgy glasses. effects of hydrogen on mechanicalof0 Chemical Analysis and from Columbia University. He has been with properties of RS alloys and glassy alloys. andMatenials Science. North- the Barnett Institute since 1978. Currently, he non-equilibrium phase diagram He is a men-eastern University, Sos- directs the research of a number of sponsored ber of The Metallurgical Society of AIME.

'i~ton. Massachusetts programs including rapidly solidified Ti alloy02115. research. His research interests include rapid

Dr. Whang received solidification processing. microstructural stabili-his BS in metallurgical ty and mechanical properties in PS alloys.

40 JOURNAL OF METALS aApril 1984

%.. . . . . . . . .

U RAPIDLY SOLIDIFIED Ti ALLOYS CONTAINING METALLOIDS

ANDI RARE EARTH METALS - THOIR MUMIUlRE AND MECHANEA PROPERR'I1

lei,

C.S. Chi and S.H. Whang

MRS Sym, Vol. 28, 1984, pp. 353-360

12~

1e- u . u n~ I- k -

a A

5 v iU.U uk-

10 an 0,a3 A

92A D A a1 sW . @ U -a a * ad k*0aa a' aN OW A

0' AA~ *..1

vi aa Aen.I .A. aa A-'0 ' a )

0 a *: . a, 7o on M,~- ~. Z -m - *. 0 V Au a0~ -1-a ~ofa An ~ ~ a .

ILa~ :2* *au oa~ A aa

on A

00 . 'u5. a. 0a ko. a~

aa *a. . 3 .U0 m 0 a.6

-~~9) 0Z~ A 2a~ Z ~ aa ~ S~a S u

ca ak _a .. A~I- Ma

0 ~ .p .3@a. UU~a a - a..0 a

a j- m0A1 -a. aP.. a C0 d.a 01 A A Wnn.

. -0 a 6 a~- - a - 00a.0 . . ,"A% zoo .. _

,4 C 0 a X Z*.. .. . 4 1 o ,uo

. •. Z. A. s3.-O h wk + * 1 IU * ..0-- a *• ,., ,- . O I e1.0:0:1 .4 ..009 05~., ..- e

C . 0-2a~ a6 6 0 0

va .. -00a t• a o 0 A 0 a 0

em. 00.0 0 A 0. A' a-~mee~ 0 W. A ' A4 d.o* • * 0 0 **- . .OU A.

0 .

-- 3 I ,,i00 N . * as iO a

-a i - . o 0 0 .

ow,* " . ' - j:"* -0 '*o" Ok

.~~ 30, a oA., e. ,-, .i . := . Z *. >l ,",,,.m. , m *,mo .,k, . ..

-uu:~ .- ,..A=. =;o=. * .. Z j ! * .;,c * ..-, , U - . 3. - . *A ..S , 0,* 64 k .-

2o a _ Aua . m 5. - , a - ,- ,..J . em ro m -Z.,i 0. 0 .0:...a"•I " -a o 'a AC

*o u , .*US U o..1.

A* U: 2. : :i0.ii: .

---Iu. --0.. uI I --- • ma l 8 g o * I ' Jm . -- S. I' - a , o t • o o O0I on S . *0 a .Io A 0- a 0 •

"0 0 : f : 0 O's-.- -e -6 A - u

m 0 .0. A* VI -iuU . aaaa 0

. a -0 ain0 0@. -. d 3 a

0 *0 a 0 00

.:.. u 2 3+~ a5~ ~ ,... , .... , uo @

a m u, A- 0 a w " . a d 06e u a 0

oaE.-l;..a .og o. .....5 ma a -lm Ueu--.oo .. m 0

a . -- m1.a s 71 ., u I -IOimuA , o 0 .a oaI.. a

:,, o -o... . - o- C- .,,,.C - I...-. -u . ' at,,

Q.. 0, .00 i.0 Igo ZI :k

SOw

a .a !: A j e o'

Z a ow 0

S0 a 0 0.I a 0 Aa 0 ao

* ': ".]=,, " auo~ o 9 - . .. . em m+ .

"4 ~~~~, a e

. - -J"* . ,I O. .A I ,." a

- a o . A.- * I .

*u eu.me,, .n,.. . ,,.. .. * .

.d 0 a.•- 0 d L'Ut

*

eAu.1S C . ... .... 0=: .- . ,omu.u . ...A~ 4

o . .. . . , . . . -0-. -0 Ch w 045

CmG A*1 al a* 0o 9C1

C A a a a a . -- -,, ., -. ,,. .. ..:.

a6 0 a. b .. A. .~. 00.. 1.. ACa a. -I o . 5- %.1.. I-0 A O:.- 00- "A03~ -- 0 0i~m01. a 3.' .. a C a a :U ON~

As *- a ~ - - a.- 2ol oA~ A. 2 ~ .m a.-.s UUsA. .- a . 0. 0 -0 In mA. A~~** 'goU0:mA mu *.. 4 :U :- -ml o. l0 o 5 Z.0 ua a a 1., 0 C ;a. -A.. .0 1 10 a eNA"-.- - qaaa 0...0 so a. 0 -0 a a0 a0 , 0 mZ9 t CC on a. sa0,CA5m 'A'A -0 W.=~ ASC.

P...2 - ................

*o E

bf'

_j0

A' rig

EE

3 (45 - * a 3 04

04 304 0 ~d% -

0- -. 0 I.- ~ U04 SI* .4* ~ ~ 0 0* 0-.. a.

4~.. 4/'

4 0 - 004 4Q04.4 ~> a 0

04 044 N

4. ~ /

004-i

4~

4.. 3 04 0 -, 0- rJ C0 41.. - *4 *4

* =4.4

3 0 04

0.4 .4.- 04S. 044 I

04 04a- 04... 4 N -.4 4.04 3 -433 . *...N .4 -4 04.43. -. 0-N *4 - ~0 ~ ..~ K 1:-04 4.S rN 044 4. 4~4 44.4 o~........-..

4.- 43 4 a 4 Jo8

,04 .t.. -- 3 .4 - ~.o.. . I* 404 4 0 ~o.a

I..4 ~ - i - 84~ . I>4 (-44 04 3 ~0I...

.. , K- - N f~4 *4 4 44

3 4 4 -. -~3-- 4A 4~ 4~ 4~ ~ > 0 -~0~3...04 - - Sn 04*4 5t4 5(4 5(4 4 4- u ~*4N N N N N 3

0- 3*44- ~ .~ .~0 0~44 (4 ~ 5(4 ~ 0 4 0 3)

43 0 44 ~ ~ '54 45Iii * 0

d4 SS~NQ~VN0~0I~

r.

0*

-:

-- 04 t80404404....404 04 ~04 04.4443. >4'5404 4fl453 *04 0404A>3.40...4 404 >..4 04>..*3*t3.444 4- 4 CS. 4 .04.. 304 43 4. - ~-~4 '4 404 4044 4 433 04.4) *N 0 a .4044>5 44 ~' 8844,a, 84 N 04 4 .4044 4.4 3..4-4... .444~~(%04 ~04UN04 --o *04 04 -4 0 ~ 404 0 4~ = *4- 0404W

3aN 4 IC. 04 - -

04.4 5.. 0~. -- 4 0443 ~ 3~ 4>430. S.f... 04~ ~o..4 ~ 4.4> 3..3.43 - -4.4 04

4o 4)043

04-.

4.813

>54o .4c~ 43 0 ~ 0.4

.4. .404.4043~ 04 3 ~-*0 - 4 - --~04 * 044 404.4 .4 GL.8 -4) 4 0 404.04.3 444...

a. >~ - - -4.:~ 4 304344.44 8~ ~ *-- ~.S .40

4 ~. ~o ~'. - .44 404 55 04=8.445404 44 ~.

4 44 0 - 43 .4* - 1 .4;~, 04*04-- 0. 3 8~4.4.4 - ~ ~ ~ 044 304

0 - >..3.4

4

04044

.404*04 -~ .4. 4~.. 0 -- *s*, 40 >4 .504 .~.4 4 ~.4 ~ .334 . 38 ~ -04333 - -- - 4 *8

0 4.. A

3. >~

- 48 04 - - - 404 - 4, 04 *3~ -043 4 >~> .4.4~ 04 404 .3404>5304.4. ~ 404

44444 >.4 ~44..40.4.40404 - ~ ~ '5~ 3>,~4 ~"3 3*4 35040443.40. .4 . ~ . 04*040. * A '4 0 *3~43 g

0 4 0 4w ~ 4 .04 - 0044.. ~ 4.4 ~4. *~

3> * 043040443 4. -,*~ - 0444 4404404 44> * 04343.4 44 44~0 4 4.4A~ 04 - 4..... 0444 04-. ~ ~fl4*3q

4.44~4.4.404 ' 04543, ~ ~ - 33 ~ 404.~.4

>54 4404 04Z~ ~ 33 = .53 - 3.44 - 4 >~ 0

0 ~ asfl . 04.4...A A.~4.~3 >

4 ~ in.. '4.-~- 0 - - - - 47044

0

- - 43~440433 430304

00404 ~ ~ .S~.44 3 ~4 .04*~

- A 5.- 4.

*0 040404 '4 4 -.4.4 44 3~

4 *0484440404 4 0444h. ~ 4 4- - 043 4 ~ 044. ,~ ~ 044 ,3, 455 >5-43~4.4~4.44~04 4.4*404 .4~5. -4 3;..*4404...50404 ~ ' 4*0 54 A

4 ~34.304 *4~ *~0404*g004*04 03333

04>5~4 0404.4 0440.5. 4.043~3304.4 Ii - U.~04 .. ~04.48..4S

04 041304 -- - ~4.4 04 ;3,..4.4 8

- 43*.40404

3.38.. - ~ .43404443.44*.404~

504*4.4 ~g3 .404Pp .4,.4.4 0 3 .. 4~44fl04 04

04 0 04304 4.9 044~.40.04 '4. ~40044..s8344 ~ ~ 044'4.44. 8 43*3.44.4.404 4fl 040 3 4* 304 04 4 ~ ~>-~ 3oO A *044. - 04 0443 84 >~ *:* ~*4. ~ ~ 04- 30.3 - 4.4 >04 04~*4 3~04 4~ ~ .44.4043 ~ -~ ~ 40404.40. 4 ~ ~ ~S3=~4~4 ~ ~ >043~~04~3~.4

* 3! A 043.4 4 .4 4040430404 ~~0404 0404 044*4.3.. 04~0483. 354 04 t ,-~ 44404 - 04 *~ .48041 3*3~3 0434 ~043A q .43.043404 .=,~ 4 45 430404 o. 33 04

* 34 ~ 3*. .4 *S~4~ 04~~04 33.4 83 3 ~4 04 4* - 5. -304 3 443 .4 30434 *A 04 s 3*0434 04.404.43~.4 3.33 .4 .54 4.3

. ~ .440 41 0403. ..J 0453

4. 44044.44.043~ 5.0408 30~4..54 0404 34304~04 -. ~ 040.4404 3 - - 04 ~ 0404.44

* . 4. 0404 -- .3.44 048

-- . - 004304044.404.4>03 0000 ~

~ ~*0443304~5 404 *~.404.404.4040404= 30440404.4> ~404.4 .404 .43.4304 0404~434. 4.A 04*0433..p4.... 04 4 >..4.~3 34..8 4..4 0444.s04 0 4044

0.443 0440404 04 ~4--4044* 4 040334..40404>~ *404

4* .-. 4.4404340..,.. 04 044 34444s S 30430434 *4*3 -~ 0404*~.4~4>.304

4.4

8.4------3 3430404.....3 3 ~~~434G3333.4 **434433-.4.*0.* -..- t .*... ,4*

I I THERMAL STABILITY OF PRECIPITATES IN A

RAPIDLY QUENCHED Ti-Al-Si ALLOY

S.fl. Whang, Y.Z. Lu, and B.C. Giessen

Proc. MRS Sym, Vol. 28, 1984 pp. 367-373

20

0A is 32 20' 0k - k 0 In*-

AA 0 Gn 0 - AA.

V 1o .c0000 6 0 C ~IN .o0" ~ ~

Ua 00 00~0 ~ -

-~~~- a0 AkA A 6A0.U

aS A

AC 0 -

2, - a .A C

uo' a ~ .2 - _

'a Ok g

z 0 Aw 93 .0 >.' n- - o. m

;D v 0' ccAly 00

o 4; oo

O'WO a~ .0 cc0 0 w

0! 0 *-*'c

a- --d' z 0 C u P-

~~~tC~~ 0 0 0 ~ ~ ~

00 Z so 'a "V A'a CIL~'

w 00 0.

.40 "5 Cc; u .'. 0 W) f U..C.4... 0 C-O .A4 0 3 9 a .0' o .w.0 a0.00c J

-, 3 aaoa AAA.0~ A 2

2~ A. ~ -4o

.4aae..4.4 AO , 0 - -~ . >" ~ U. .. 4A3

'.4., > Ao

a. ca~ - - 40 -t

*a nb r.f a

ta2 4 -21a~~. -a 0.. 3a.

0 0-. ~ ~ a ..i - .

4;. A .... a * U

a. '0 a.-E

.,~,6.

a A>4- 'A-

aa- 14tE

a0 0. .4; >W- a

0. a

-~ ,0

0* 0 0 0

I - . 0 00 1

C 5 0

04~~A C0 u ~

- 000C Cc, .

-If -Ioa NUP (0. 0 0

0.. 00 0 Ad) o 0

m- 00 1. Cc CD - 0*

eW 0.

0~6 0.00 ao C 0 0

1'. 0 > .oa 0 d .0

A~ ~6O3 all~~qo AC6 04 4' 0 '0 -0.0 .

.0 L@.. If0

C2 0 0

0 0

COO 604

-4. 0 B

0 1.

o~~~ 0 4O

CL 0NC 0.00

g 0 Cu-d .0 4; T

g t- go

0* 0.

. -- --c

-. 9 00 .

I ;:- 0 I --

0. .- 0L. a 4 0.

3 ED a 0 'A~

C3~v m -00-0 0.-aa 1.- -e 0

a-" .d 0 . .;l00 1a 0 13C

m 0.c co c a*.

. . . . .W r ...

z e . -C - -C U.-C.

.4 U W'C '! .

soaa -- eeaca I- 019 as C. t

ea.U a

-0 I-.. .a, 0.-. s

z10 10 z a ~uu~

0 -0 1.-:-U

aaa4 oaa .g *

-oA 0 a- 00-aa.

O a 0

a00r A-A a aa 0 0 0 0 -a- 4;

NICROSTRUCTURAL CHARACTERISTICS OF RAPIDLY QUENCHED

alpha-Ti ALLOYS CONTAINING LA


Submitted to Met. Trans. A. Dec., 1984

Barnett Institute, Materials Science Division,Northeastern University, Boston, MA 02115

28

.' °°..* q

MICROSTRUCTURAL CHARACTERISTICS OF RAPIDLY QUENCHED a-Ti ALLOYS CONTAINING La


Rapidly quenched Ti-La alloys show that the maximum extended solubilty of

La in Ti at the cooling rate of -106 K/sec is -0.3 at. %. An excess amount of 0

La (0 0.3 at. %) deposits in the form of clusters ((50 A diameter) along the

subgrain boundary. Grain refinement resulting from rapid quenching is

effectively aided by alloying elements (A1,Sn), but no apparent effect,

however, on grain refinements by La was found. Fine precipitates in Ti-5AI-3La

alloy were identified as a mixture of binary La compounds: AI4La (Hex., DO1 9)

and AI3La (Orth., D13) while those in Ti-5Sn-3La alloy appeared to be .

La2Sn(Hex., B82 ). So far, no ternary compound has been identified in these

alloy systems. An EELS study clearly shows that the precipitates of these

alloys treated in a high vacuum contain La, Al and Sn, but not Ti. Various -

substructures resulting from rapid quenching often remain in the matrix even

after high temperature annealing, probably due to the pinring by fine

precipitates. These precipitates appear to be stable and corasening resistant .

at high teperatures. Strong age hardening accompanied by fine dispersoids .

exists in the Ti alloys containing Al (or Sn) and La. Microhardnesses of the

alloy ribbons annealed isothermally or isochronally are systematically lower . ..

than those of splat quenched foils, consistent with coarser precipitates of -"..-'

alloy ribbon.

S- . 9I. INTRODUCTION

The development of high temperature Ti alloys through precipitation or

dispersion hardening dates back more than two decades. Since then,

......... ,........ . . ,.....

7.".

. .. ..

strengthening of Ti alloy matrix has been achieved by a small addition of such

compound forming elements as rare earth metals (Ce, La, Gd, etc.) [1-4], 0.

actinide (Th) [4] and metalloids (Si, B, C) [4, 5, 6]. Of these additives, Si,

has been systematically incorporated into Ti alloys through conventional ingot

casting and subsequent solution treatment to yield fine silicide precipitates, 0

since the maximum solid solubility of Si in p-Ti reaches at a maximum of

-5 at. % at 13000C. These silicide hardened alloys show good creep resistance

at or near 5000C, due to thermally stable precipitates at these temperatures . .

(7]. Nevertheless, the coarsening rate of silicide particles in the Ti-matrix

at or above 7000C significantly high, resulting in a major loss in the strength -

gained by the Orowan mechanism (8]. -

In contrast, rare earth metals have a solid solubility range of less or

equal to 1 at. %, even at high temperatures. As a result, not only is an .

adequate solution treatment within the solubility range not possible, but also

a larger addition of these elements than 1 at % results in coarse particles,

ranging from one to several microns in diameter through conventional ingot .

casting [4]. Therefore, in some earlier works, it was reported that the -.

addition of rare earth metals did not produce an increase in strength [2], but

did improve creep strength [4] and creep rupture time [3].

One recent method of more effectively producing dispersion strengthening

of Ti alloy is through rapid solidfication processing, through which -

significant amounts of rare earth elements are dissolved into solid solution

or precipitate uniform clusters in the matrix. The succeding heat-treatment""""

results in fine rare earth dispersoids in the matrix. Such dispersoids in

Ti-5Sn-4.5La alloy [9] demonstrate that they are far more coarsening resistant

2

.. '. .,'. . ,' .'.--...-...........,%,............... -. . . -.. .-. .. ..... ........- "- --% -.;,"°-,.-

electro-polishing contains perchloric acid 4%, 2-butoxyethanol 35% and methanol

1 61%. The polishing was carried out at a voltage range of 20 V and a current of S

less than 12 mA/cu 2 . The annealed samples containing coarse particles were

. thinned by electropolishing followed by the ion mill technique. In order to

study the structure of dispersoids in RS Ti alloys, carbon extraction replicas S

were prepared from the annealed samples containing coarse particles.

* Microstructure was studied by a Philips EM300, a JOEL 100CX and a vacuum

generator HB5 field emission STEM with EELS and EDXS capabilities. The S

structure of dispersoids was determined by diffraction ring patterns while, the

composition was confirmed mainly by EELS and EDX studies. Microhardness was

r" measured on heat treated samples by a Vickers diamond pyramid tester with 50

- and 100 g loads.

III. RESULTS

A. Extended Solid Solubility

Extended solubility of solutes by rapid liquid quenching is a well-known

* phenomenon. During rapid solidification, phase separation and phase formation .

are restricted within one per hundred thousand second (- 106k/sec) so that

- excess solutes are trapped in the matrix. Therefore, the degree of

supersaturation in rapid liquid quenching (-10 6deg/sec) far exceeds that in the S

equilibrium phase. The resulting effects include solid solubility increase,

lattice parameter change, strength increase and microstructural refinement.

Table 1 shows that in general, the extended solid solubility is much larger P

*'-" than the equilibrium solubility at room temperature where the equilibrium solid "

solubility of the solutes becomes very small. In particular, upon rapid

I quenching, the increase in solubility is large in phase diagrams with a

terminal eutectic, i.e., Ti-metalloid systems (C, B, Si) with eutectic show a

large extended solubility. From a structural point of view, solid solution may

be a substitutional type (Si) or an interstitial type (C), and at least in part S

4

Microstructures of Ti-5AI-3La alloy may be better understood by studying

microstructures of binary alloys: Ti-SAI and Ti-3La (Figs. la and lb). The 0

addition of Al to Ti results in grain refinement and twin structure during .-. ,..

rapid quenching. In contrast, grain refinement is less apparent and no twin

structure can be observed in the Ti-3La system. Instead, very fine clusters of 0

La are dispersed in the matrix. The microstructures of as-quenched Ti-SAl-3La

alloy appear to be a combination of these two binary microstructures, as shown

in Fig. Ic, i.e., a strong grain refinement with uniform clusters. The excess S

La beyond the extended solid solubility (Table 1) deposits preferentially along

sub-grain boundaries (Fig. Ib). When the alloy is heat treated at 7000C, 2h,

precipitate coarsening initially occurs along the sub-grain boundary (Fig. 2). N

In particular, large particles are observed at the triple junctions of high

angle grain boundary. After the Ti-5AI-3La alloy was annealed at 700c, lh

(Fig. 3a), an EELS study of the precipitates (Fig. 3b) showed very high La

concentrations in the precipitate while no apparent oxygen peak was detected in

these spectra. In contrast, those spectra from the matrix consist of a strong

Ti peak with the absence of La peak (Fig. 3c). . -.

Since oxygen levels in both the precipitate and the matrix are not

distinct, no comparison between the two levels can not be made based on the

spectra alone. Further, the precipitates isolated by the extraction replica 5

technique (Figure 4a) yield the diffraction patterns in Fig. 4b. Listed in

Table 2 are those d spacings of the ring patterns, AI4La (Orth., D13) and .

A13La (Hex., DO19). Other d spacings from La203, TiO2, Ti 203.03 , A12.12Lao.g 8

(FCC), Al 2La and pure La were checked against the obtained d spacings and no

similarity found between them. Neither the diffraction patterns of Al3La nor

those of AI4La alone can satisfy all the experimental data. However, the d -

spacings of both Al4La and AI3La together match the experimental results.

6•.-. .. ,% ,

d spacings of LaSh 3 (Cu3Au), Ti-Sn binary compounds, La203, TiO 2 and found to

disagree with the. EELS of a precipitate in the Ti-9.5Sn-5.3La alloy exhibits

a strong La peak and a weak Ti peaks which probably comes from the background

(Fig. 9b). Another spectrum of the matrix (Fig. 9c) shows a strong Ti peak, .*.

but no La peak can be identified. In both figures, the intensities of oxygen

and Sn peaks cannot be evaluated, since the intensities appear to overlap with

Ti peak intensity.

Microstructures of as-quenched alloy ribbons of 50m were studied by TEN

after the ribbon specimens were thinned by mechanical polishing and later the

ion mill technique. A micrograph of the alloy ribbon (Fig.10) shows relatively

coarse precipitates (average -50=a diameter) within large grains without -.

substructures. This indicates that the low cooling rate is responsible for

such coarse microstructures.

) Ti-5AI-2.5Sn-3La and Ti-SAl-2.5Sn-4Zr-3La alloys.

These alloy systems are made of conventional a-Ti alloys and La addition.

Microstructural characteristics of as-quenched Ti-5Al-2.5Sn-3La alloy are

predictably similar to those of ternary alloy systems, Ti-Al-La or Ti-Sn-La.

These similarities include the formation of La origin cluster and the degree of

grain refinement. Although grain refinement in the Ti-SA1-2.SSn-3La alloy is

more extensive than in the ternary alloys (Fig. lla). Diffraction patterns of

the Ti-SAl-2.5Sn-3La alloy indicate that no compound exists at the as-quenched

state. (Fig. lib). Upon aging, uniform precipitates start to emerge in the

matrix while slight coarse precipitates already appear along grain boundary

(Figure 12a). As a result, the diffraction patterns (Fig. 12b) exhibit new

ring patterns which do not belong to Ti, presumably resulting from the

precipitates or dispersoids. After high temperature annealing of this alloy at

9000C, 2h, particle coarsening accompanied by substructure disappearance

99- % . . . . . . . .***%* **.*.*.*~.*.*.**~ . . . . . . . . . . . .

(CU3Au), Ti-Sn binary compounds, La203, Ti02 and found to disagree with them.

EELS of a precipitate in the Ti-9.SSn-5.3La alloy exhibits a strong La peak and .

a weak Ti peaks which probably comes from the background (Fig. 9b). Another

spectrum of the matrix (Fig. 9c) shows a strong Ti peak, but no La peak can be

identified. In both figures, the intensities of oxygen and Sn peaks cannot be

evaluated, since the intensities appear to overlap with Ti peak intensity.

Microstructures of as-quenched alloy ribbons of 50m were studied by TEN

after the ribbon specimens were thinned by mechanical polishing and later the 6

ion mill technique. A micrograph of the alloy ribbon (Fig.lO) shows relatively

coarse precipitates (average -50nm diameter) within large grains without

substructures. This indicates that the low cooling rate is responsible for

such coarse microstructures.

c) Ti-5A1-2.5Sn-3La and Ti-5AI-2.5Su-4Zr-3La alloys.

These alloy systems are made of conventional a-Ti alloys and La addition. . .t

Microstructural characteristics of as-quenched Ti-5A1-2.5Sn-3La alloy are

predictably similar to those of ternary alloy systems, Ti-Al-La or Ti-Sn-La.

These similarities include the formation of La origin cluster and the degree of

grain refinement. Although grain refinement in the Ti-SAl-2.5Sn-3La alloy is

more extensive than in the ternary alloys (Fig. lla). Diffraction patterns of

the Ti-5AI-2.5Sn-3La alloy indicate that no compound exists at the as-quenched ...

state. (Fig. lib). Upon aging, uniform precipitates start to emerge in the

matrix while slight coarse precipitates already appear along grain boundary

(Figure 12a). As a result, the diffraction patterns (Fig. 12b) exhibit new .

ring patterns which do not belong to Ti, presumably resulting from the

precipitates or dispersoids. After high temperature annealing of this alloy at

9000C, 2h, particle coarsening accompanied by substructure disappearance

occurs in some grains, whereas in other grains, fine precipitates and

8

the calculated Nusselt number shows that splat quenching covers over a cooling

range from the Newtonian to the intermediate regimes. Presently, no direct S

measurment of cooling rate for Ti alloy ribbons or splat foils is P.vailable.

Nevertheless, the estimation of the cooling rate for Ti-6.3Si alloy is

possible, using the measured heat transfer coefficient [311 at the interface -

between the molten alloy and the cold copper crucible. From a one dimensional

heat flow relation, the cooling rate at the liquid state may be written as

dT h.AT

dt d.p.Cp 0

Where AT=(Ti-T s ), T i , T s are superheat temperature and substrate

temperature; p, C and d are density, heat capacity, and half thickness of

splat foil, respectively. The calculated cooling rate for Ti-6.3Si shown in

Table 4 is an order of magnitude smaller that those values measured by other

" authors (Table 4). On the other hand, two dimensional models with the aid of

computer calculations yield systematically lower cooling rates, e.g., 3.2-6.4 x

105 OC/sec for Fe-25Ni splat of 40pm thick (d) [29] and -105-106 OC/sec for Ta-

45 at. %Ir, Nb-45 at. %Rh and Cu-34 at. %Zr [301. It is important to note that

the outcome of the cooling rate calculations depends significantly on what

" value of the heat transfer coefficient is being used among the large selections

available. Therefore, it is not reasonable to compare these cooling rate

values to each other without checking into the heat transfer values employed.

It is interesting to note, however, that the Nusselt numbers from our present

work and the work of Predecki et al are close to Newtonian cooling whereas

" fHarbur et al have found numbers large enough to fall into the intermediate

cooling regime. Such discrepancies may be rationalized by the fact that in the

former two experiments, the melt was not quenched under the high pressure

11

.-.

temperature. Nevertheless, kinetic conditions such as kinetic undercooling and

interface attachment kinetics may allow only certain compositions to undergo 0

partitionless solidification below T0 138]. From these arguments, alloy

compositions near the eutectic, where high undercooling can be achieved through

rapid quenching, are expected to show high extended solubility. By contrast,

monotectic compositions of Ti-rare earth metal systems are expected to lead to

lean solid solution, since the magnitude of reduced undercooling for the

critical cooling may be too large to achieve in these systems (22]. Table 1

demonstrates that a large solubility extension is achieved in eutectic systems,

whereas little extension is possible in monotectic systems.

Another phenomenological approach by fundamental parameters may be 0

considered in the following. The solubility ratio (between the extended and

the maximum equilibrium solid solubility, a-phase) calculated from Table 1 is

listed in Table 4 together with atomic size ratio and electronegativity of

additive elements. These data were plotted in Fig. 16a and b. Both Figures

indicate that the solubility extension is correlated with atomic size and

electronegativity. Since the pioneer works by Hume-Rothery and his coworkers ..

have demonstrated that the solubility of solutes is governed by atomic size

ratio and relative valence electron value (391, many other parameters such as

electronegativity [40], heat solution [41,421 have been employed to predict the

solubility of solutes in equilibrium alloy systems using a two-dimensional

map. These electronic parameters are basically related to each other, but more

effective prediction has been made by a combination of heat of formation and

atomic size for a two dimesional map [42]. A similar approach has been

successfully applied to quenched alloy systems: ion implanted alloys [43] and

RS Al alloys [44]. Such success in RS alloys indicates that the parameters - _

used for the equilibrium alloy systems are equally effective in solubility

13

- , ~~~~~~~. ... ....................... ..... . . oo ... .. ",o-

;.-.-.',.'..-.-.'.. -.'-... ..- '..' .'2. i'-. -. -. - ', '. ' .''... ',..-' - -. . - '..- .- ... ". '..,. .,---..'-: -.,-', . ,, -.i ]. 2.' . ,,.."," , ".-.,

1. At the cooling rate of -106 deg/sec, a large solubility extension of

additivesis observed in the eutectic type alloy systems whereas a 0

small extension is noted in the monotectic type alloy systems. This

phenomenon may be analyzed with fundamental parameters such as atomic

size, heat of mixing, and electronegativity. "

2. Ternary alloy Ti-SA1-3La yields uniform precipitates (average size

-200 A dia.) upon aging at 700 0 C under high vacuum. These precipitates, •

prepared by the extraction replica technique, were identified from the

diffraction ring patterns as a mixture of particles

of Al3La (Hex., D019 ) and A14La (Orth., D13 ).

3. Precipitates of annealed Ti-9.SSn-5.3La alloy (8000) yield diffraction

ring patterns similar to those of La2Sn (Hex., B82 ).

4. No rare earth oxide was identified from the diffraction ring pattern in

the aged Ti-5AI-3AI and Ti-9.SSn-5.3La alloys. This indicates that the -

oxide formation under vacuum is insignificant compared to non-oxide

formation.

5. Coarser precipitates wey-er found in the as-quenched ribbon of Ti-SAl .

-2.SSn-3La alloy, in comparison with fine precipitates of splat foil

with the same thickness. This indicates that the average cooling rate _

of the spun ribbon appears much lower that that of the splat foil(~106

deg/sec). The cause of the lower cooling appears to originate in the

poor contact between the melt and the substrate, as evidenced by air

pockets and wetting patterns.

16 0

ACKNOWLEDGEMENT

The authors gratefully acknowledge the support of the Office of Naval

Research for this work (Contract N00014-82-K-0597). In particular, we thank

Dr. Donald E. Polk for his thoughtful suggestion for rapidly solidfied Ti alloy

research at Northeastern University. Contribution No. 202 from the Barnett

Institute of Chemical Analysis and Materials Science.

REFERENCES

1. E.J. Chapin and R. Liss, Report of NRL Program, December 1955.

2. B. Love, WADC Tech. Report, No. TR57-666, Part II, March 1959.

3. N.J. Grant, U.S. Patent 3,070,468, 1962.

4. M.B. Vordahl, U.S. Patent 3,622,406, 1971.

5. K.C. Antone, Trans. AIME, 1968, Vol. 242, pp. 14 54-1 4 56 .

6. H.M. Flower, P.R. Swann and D.R.F. West, Met. Trans., 1971, Vol. 2,

p.3289.

7. N.W. Paton and M.W. Mahoney, Met. Trans. A, 1976, Vol. 7A, p.1685.

8. S.H. Whang, Y.Z. Lu and B.C. Giessen, Mat. Res. Soc. Sym. Proc.1984, Vol. 28, 367-373.

9. Y.Z. Lu, C.S. Chi and S.H. Whang, Proc. 5th Int. Conf. Rapidly SolidifiedMetals, Sept. 2-7, 1984, WVrzburg, W. Germany, in press.

' 10. B.B. Rath, et al., Proc. Fourth Int. Conf. on Titanium, H. Kimura and 0.Izumi, eds, AINE, N.Y., N.Y., 1980, pp. 1185-1196.

• 11. S.M.L. Sastry, T.C. Peng, P.J. Meschter and J.E. O'Neil, Journal of Metals,Vol. 35, No. 9 (1983) pp. 21-28.

12. S.H. Whang, Journal of Metals, 1984, Vol. 36, No. 4, pp. 34-40 .

17

i • S

33. B.P. Bardes and N.C. Flemings, 'Modern Castings,' 1966, Vol. 50, pp.100-106. -

34. W.W. Mullins and R.F. Sekerka, J1. Appl. Phys., 1964, Vol.35, p.44 4

35. S.R. Coriell and R.F. Sekerka. Rapid Solidfication Processing II,R. Mehrabian, et al eds, Claitor's Baton Rouge, LA 1980.

36. S.P. Midson and H. Jones, Proc 4th Int. Conf. Rap. Que. Met. Sendai, 1981pp. 1539-1544.

37. M. Cohen, B.H. Kear and R. Mehrbian, Rapid Solidification Processing II,R. Mehrabian, et al, eds, Claifor's Baton Rouge, LA 1980. pp. 1-23.

38. V.J. Boettinger, Proc. MRS Sym. Vol. 8, B.H. [ear, B.C. Giessen, M. Coheneds, 1982, pp. 15-31.

39. W. Hume-Rothery, G.W. Mabbott and [.M. Channel-Evans, Phil. Trans. R. Soc.,Ser. A, London 1934, Vol. 233, pp. 1-97.

40. Darken-Gurrey, *Physical Chemistry of Metals,' McGraw-Hill, New York 1952pp.86-90.

41. S.R. Chelikowsky, phys. Rev. B, 1979, Vol. 19, pp.68 6-7 01.

42. I.A. Alonso and S. Simozar, Phy. Rev. B, 1980. Vol. 22, pp.5 58 3-5 58 9.

43. D.K. Sood, phys Lett., 1978, Vol. 68A, pp.469-472.

44. H. Jones, Proc. Mat. Res. Soc., 1984, Vol. 28.

45. L.E. Marr: "Electron and Ion Microscopy and Microanalysis,1982."

Marcel Dekker, Inc., pp.78-187.

0

19

0

1.840 16 [232.105

1.8001 10 202

1.774 M1.774 2 301

1.687 16 r2331163

1.668 6 204

1.664 16 220

1.657 4 064

1.584 10 212

1.511 20 311

1.467 w 1.480 16 r1031302

1.442 4 400

1.376 10 401

1.342 w 1.353 30 [203L222

21

TABLE 4: Calculated cooling rate for Ti-6.35i compared with those reported in othermetals.

alloy Half Thermal Heat Nusselt Cooling Substate Techniquethickness conductivity Transfer number rate* materials usedof splat (k Coefficient (20iim thick)foil Wd V/cm-K W/cm2-K OC/seccm

Ti- 2.Ox10-3 0.218 0.6 0.0027 3.3x1O5 CU & Cu Hammer &6.3Si (pure Ti) Anvil

Al 7.OxlO-3 2.20 18.4 0.06 (4.8z106 )[26] Fe & Al Plate &Pb g.8xl0-3 0.33 20.9 0.26 (5.3xl06)[26] Fe & Al Plate

Al 1.0xl10 4 0.28 11.3-28.8 0.00051- (1.5-3x106)[25] Ni/Ag Coupled0.0013

Ag 4.29 56.9-226 0.0013- (1-5xlo7) (25] Ni/Ag Substrate0. 0053

*The cooling rates are adjusted based on 20ijm thick foil by using the relation ofinverse proportionality of cooling rate to the thickness.Nusselt Number =(h-d)/k

TABLE 5: Atomic Size Ratio vs. Relative Solubility Increase

Due to Rapid Quenching

Wigner-SeitzAlloy System Solubility Ratio Atomic Radius Ratio

(Ex. S.S.IM.Eq.S.S.(a) (RS/RTi)

Ti-C -18 0.68

-B -14 0.75

-Si - 9 1.04

-Y - 5 1.23

-Ce -0.6 1.25

-La -0.3 1.28

23

.~~~~~~~~~ .* . .. . . . . . . . . . . . . . . . . . .. .

Fig. 12 Annealed Ti-5Al-2.5Sn-3La alloy foil (7000C, 2h)

a) bright field

b) diffraction pattern

Fig. 13 Annealed Ti-5A1-2.5Sn-3La alloy foil (9000C, 2h)

Fig. 14 Ti-5A1-2.5Sn-4Zr-3La alloy foil

a) as-quenched stateb) annealed at 9000C, 2h

Fig. 15 Microhardness of alloys: Ti-5A1-2.5Sn-3Laand Ti-5A1-4Zr-2 .5Sn-3La

a) isothermal annealing at 6000Cb) isochromal annealing from 5000C to 900 0C for 2hc) comparison between splat foil and spun ribbon

after isochromal annealing

Fig. 16a [Extended solid solubility/maximum equilibrium solidsolubility] vs. [Solute atomic radius/Ti atom radius]

Fig. 16b [Extended solid solubility/maximum equilibrium solidsolubility] vs. Electronegativity

25

.1

1

. . . .

0.2

* Fig. la As-quenched Ti-LA1 alloy .-

foil (B.F.)

Fig. l As-quenched Ti-53Lalo *al foil (B.F.)

0.2)u0

L

Fig. 2 Annealed Ti-5A1-3La allo-.foil (7000C, 2h; B.F.)

Fig. 3 Annealed Ti-5Al-3La allov.:foil (7001c, 1lh)

a) precipitates at the lowangle grain boundary

9 , ~ .Fig. 4 Particles obtained from

faI , annealed Ti-5A1--3La alloyA 1(900 0C;i)

a) bright field

1*.*.. .' 4

'* ' Iim

2.0

r 1.8T-15AIm3BLe

1.6 La

.4 PRECIPITATE

01.0 0

Z 0 .80 Fig. 3b) EELS of a precipitate

0-6. (arrow sign)

0.

0.2- TI 0

0 0

0 0.1 0.2 063 0.4 0.5 0.6 0.? 0.8ENERGY ,KEV

P,0

2.0-

1.86 Tl-5AI-3La0

0 0

Figure 3c

00.6 EELS of the matrix near the*precipitate

0.4-0.2- L

0

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8ENERGY ,KEY -

Si

Fig.4 b) diffraction pattern

Fig. 5 As-quenched Ti-9.5Sn alloyfoil

a) bright field -

Figure Sb

diffraction pattern

II

Fig. 6 Annealed Ti-9.5 Sn-5.3 0La foil alloy

a) bright field

0.5p0

S Figure 6b

higher magnification of a)

'O.25/LLm

0

A Fig. 7 Annealed Ti-5Sn-3La alloy 0'4 foil (7000C, 2h)

O.25)um

A Fig. 8 Annealed Ti-9.5Sn-5.3 Laalloy foil (850'C, lh)

a) Bright field

Fig. 8b) Diffraction pattern

Fig. 9 Annealed Ti-5Sn-4.5La all--,,foil (8000C, l0h)

a) bright field

~ ~ . * *~..*~**~%**** * .500A.,

3.5T I -5 -8 4.5La

3.0 L

2.5PRECIPITATE

.2.0.

1I.5

0.

TI Figure 9b

0.8 p EELS from the precipitate* (arrow sign)

0

0 0.1 0L. 0. 3 0.4 0.5 06 0.7 0.8ENERGY ,KEV

* 3.5T i-5Sn.4.5La -

3.0

2.51

.2.0MATRIX Ti

Sn-z

o1.0,Figure 9c

0.50.5a EELS from the matrix

Lar00

*0 0.1 Q.2 0.3 0.4 0.5 0.6 0.7 0.8

ENERGY ,KEY

0

0

Fig. 11 As-quenched Ti-5Al-2.5Sn

3La alloy foil

* a) bright field

Fi.ure 1 lb

difira1c t on pat tern

I S

S.Fig. 12 Annealed Ti-5A1-2.5Sn-3Laalloy foil (7oC-)C, 2h)

* a) bright field

Figure 12b

diffraction pattern

Fig. 13 Annealed Ti-5A1-2.3Sn-3La

allov foil (900'C, 2h)

0S

.- *

Fig. 14 Ti-5A1-2.5Sn-3La allov foi

a) As-quenched state

b) Annealed at 9000~c, 211

h2 -

v -T

Age Treatment (600 C)

0 70

6.0

z/

< 5.0-00r A Ti1-5A/-4Zr-2.5Sn-3L-c

4.0-A ni-5A1-2.5Sn-3Lo

Specimen thicknessi 2Oum

,AQ Ti-5A1-2.5Sn- - - -- -- - - -- ---- - - - --...............

3.0

A l2 4 10 20 40

TIME, hours

Fig. 15 Microhardness of alloys: Ti-5A1-2.5Sn-3La and Ti-5A1-4Zr-2.5Sl3La a) Isothermal annealing at 600'C

lsochronal Annealing (2h) AT-A-.S-L

A A 7i-5A1-4Zr-2.5Sn-3Lo

* 7.0-~- Specimen thickness: 2,am-

a. A

~6.0 Az

o 5.0r

4.0

%AQ Ti-5AI-2.5Sn

AQ 500 600 700 800 900TEMPERATURE, "C

Figure 15

b) Isochromal annealing from 5000Cto 9000C for 2h

. .. . . .. . . . .

9.0. Isochronal Anneolngi2jJ

(D 8.0.rtkes4u

. 7 i-5Ai-4Zr-2.5Sn-3La

5-0- Figure 1500

S4.0 To - T-5Ai-2.5Sn-3La c) Comparison between splat foil arr'0 Rbbr soun ribbon often isochromal

2 3. -- --- ---- --- --- ---- --- --- annea~ing

A Q 500 600 700 800 900TEMPERATURE, -C

18 %

~16 %~

CO)

12\0 S

W 0

Ce~

0051-0 1.5

Fig. 16a [Extended solid solubility/maximum equilibrium solid solubility]

vs. [Solute atomic radius/Ti atomradius]

~4j

-1

0*- w

qo u

oD (0 IJ D ( ~

000 LC)b3"SS

SECOND PHASE COARSENING IN Ti-5Sn-4.5La SYSTEM

Y.Z. Lu, C.S. Chi, and S.H. Whang

Proc. Fifth Int. Conf. Rapidly Quenched Metals,Sept. 3-7, Wurzburg, W. Germany, the proceedings in press.

67

* ... ... o.

SECOND PHASE COARSENING IN RAPIDLY SOLIDIFIED Ti-SSn-4.SLa SYSTEM

Y.Z. LU, C.S. CHI and S.H. WHANG

Barnett Institute of Chemical Analysis and Materials Science, Northeastern University,Boston. MA 02115, USA

Ostwald ripening of rare earth dispersoids in rapidly solidified Ti-SSn-4.La alloy was -studied by TER at 8000C (u-phase). The results show that 1) coarsening mechanism isbasically governed by lattice diffusion 2) slow coarsening of these dispersoids isattributed to low diffusivity of La in a-Ti, -4 x 10-14 cm2/sec at 800°C.

1. INTRODUCTION coarsening kinetics for these rare earth

The recent developments of rapidly dispersoids in Ti.

solidified (RS)a-Ti alloys with high In this paper, we report on a preliminary

specific strength and good high temperature study of Ostwald ripening of La2Sn

properties demonstrates that the alloy dispersoids [4] in Ti-SSn-4.5 La Alloy at

matrix is systematically strengthened by 800 0C.

various novel d~spersoids. These

dispersoids consist of oxide and non-oxide 2. EXPERIMENTS W

compounds of metalloids or rare earth In order to prepare uniform foil alloys,

metals, created uniformly from the the same quench conditions were applied

supersaturated matrix via heat treatment, during hammer and anvil quenching. Disk

In these dispersion strengthened alloys, the specimens of 20±51m were obtained from the

stability of the dispersoids at elevated middle section of the foils. For heat

temperatures becomes a critical parameter treatments the disks were sealed in quartz ..

for high temperature mechanical properties. tubes under vacuum (10-3 torr) and annealed

A previous study of particle coarsening in at desired temperatures and durations. The

Ti-SAI-2Si alloy has shown that high particle distributions of the annealed

diffusivity of Si is resVLnsible for a specimens were studied by TEM and image

strong propensity to second phase analyzer. The coarsening study was carried

coarsening in this alloy [2]. out using 400-600 particles for each data

Similarly, qualitative observation of point. The precipitate volume is determined

particle coarsening in RS Ti alloys from the TEN micrographs using an

containing carbon or boron shows that the approximate relation [5].similar rapid coarsening of precipitates has_.7

been identified (1.31. In contrast, in Ti- 3. RESULTS AND DISCUSSIONS

Al(Sn)-La alloy systems, microstructural Microstructural characteristics of the.9

study reveals that La dispersoids are as-quenched alloy may be largely divided " .'

coarsening resistant during high temperature into two categories: 1) microstructural

annealing (41. However, so far, no refinements due to high cooling rate, -106

systematic study has been made of particle deg/sec, including acicular shaped grains

% J.....

- -.- 7

and fine twin-like structures; 2) uniform : _Y ,

clusters of La in the matrix. When La .3

concentration exceeds the solubility limit

0.3 at %, the excess La deposits in the form 'N

of clusters. The cluster size is less than

50 A diameter, as shown in Figure 1. These N '-..-

as-quenched microstructures disappear

rapidly during high temperature annealing

(8000 C). Instead, new microstructures

consisting of equiaxed grains and uniform

precipitates replace the as-quenched ones

(Figure 2). During annealing, the

precipitates retain their spherical shape as

they grow into coarse particles. The size FIGURE 1Bright field micrograph of as-quenched Ti-

distribution as a function of reduced 5Sn-4.5La foil alloy.

particle radius remains constant throughout

the entire annealing period. After trial

and error, it was found that a linear

correlation was obtained from the plot of . 6 .

the 3rd power of average particle radius A.D.

() vs. annealing time (t) (Figure 3). ...Therefore, the diffusivity of the solute * •O

concerned may be estimated using LSW 0 6-

equation [6,7]. According to LSW, - " : *0 * 0-

.g W7 % ,. S. *.•, :!:3 o3 8 K.C s.(-..D.¥,y V In ';--:'

r -r 0 9 .C (R). ** (t-t0)

ro" F average particle radii for the onset 0 'O25 m M.'

and final statesK = volume fraction factor

Cs(®) equilibrium solubility of solute in FIGURE 2the matrix with particles of The annealed alloy of Fig. 1 at 8000 C, 60h. "' "infinite size at a given temperature

TD = diffusion coefficient of solute,

cm 2/sec

y fi interfacial free energy of particle, calculations was obtained from one of the .J/m' models [8]. Interfacial free energy of the

Vm = gram-molar volume of precipitate,

cm'/g-mole precipitates (La2 Sn) is arbitrarily assumedto t = onset and final annealing time, sec as 1 Jlm 2 , which is an intermediate value

R = gas constant, 3/mole-KT = absolute annealing temperature, K. among the reported values for incoherent

This equation contains an additional precipitates. In fact, interfacial free

factor K to the original LSW equation (8,9]. energy of precipitates can be determined

The volume fraction factor K for the present with reasonable accuracy from an LSW

.. ... r o C. *........................

2 0 . 8 0 0 6C 8 o o .4.0 _ h. : " . LS W""

too41 :..LSW

! Ti-Sn-4SLo -

E jo

It-.5' TI-SSn-4.5La 2

010 10 20 30 40 50 60

ANNEALING TIME (hr)

FIGURE 3 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 IS 2 0

Plot of the cube of the average particle P

radius vs. annealing time.

FIGURE 4

equation, provided the accurate diffusion Plot of particle size distribution functionvs. reduced particle radius.

coefficient of the solute is known.

Alternatively, interfacial free energy may

be obtained by measuring the solid two solutes, La and Sn, from the

solubility of the solute as a function of constitution of the precipitate (La2Sn). Of

annealing time [10]. Unfortunately, the these two solutes, Sn has almost identical

K former approach is outside of the current diffusivity with that of Ti in A-Ti [12].

experimental context and the latter is not Hence, it is difficult to theorize that the

feasible technically. Using these given diffusivity of Sn becomes suddenly two ." -.

values, the diffusion coefficient of the orders of magnitude smaller than that of Ti

solute was calculated from the modified LSW in a-Ti. Nevertheless, it is clear that

equation as listed in Table 1. The obtained these calculations involve significant

diffusion coefficient is two orders of error, perhaps because of multiplying

magnitude smaller than the corresponding effects between arbitrary interfacial free

diffusivity of Ti at 800 0C. The diffusion energy value, precipitate volume fraction

- coefficient is believed to represent that of factor and some uncertainty in equilibrium ,

La. The reason for that is as follows: solubility measurement.

First of all, the lattice diffusion for When Size distribution function p2 h(p)

" Ostwald ripening of the dispersoids involves was plotted against reduced particle radius

Table I: Diffusion Coefficient of La in Ti-5Sn-4.5La

Temperature Slope in the T- t Total vol. fraction K [8] Cs()[ll] Vm D (La)

3 2°K plot (m3/sec) of precipitate (M) (m /g-mole) (cm /sec)

103-29 5 ~ -14

1073 9.39xlO 1.4 2.35 -0"0046 2.33x10 3.7xlO1, ________________________"___________-

(p), apparent deviation from the LSW model 6. I.M. Lifshitz and V.V. Slyozov, J. Phys.Chem. Solids, 19 (1961) 35.

was noted, i.e., the maximum value of p inC . l 1 9)

these experiments is 2.0 in comparison 7. C. Wagner, Z. Elektrochem, 65(1961) 581.

with 1.5 for the LSW model (Figure 4) and 8. A.J. Ardell, Acta Met., 20 (1972) 61.

the predicted value 1.75 [8]. In fact, 9

volume effects on the ripening were not well Metall., 27 (1979) 489.

understood. 10. A.J. Ardell, Acta Metall, 15 (1967)

1772.

4 CONCLUSIONS 11. E.M. Savitsku and G.S. Burkhanov, 3. 00

1. Particle coarsening in Ti-SSn-4.5La Less-Common Metals. 4 (1962) 301.

alloy at a high temperature is 12. J. Askill and G.B. Gibbs, Phys. Status

controlled by the volume diffusion Solidi, 11 (1965) 557.

mechanism.

2. The slow coarsening rate of La 2 Sn

dispersoids is governed by low

diffusivity of La in a -Ti. This

fact is contrasted to the high

diffusivity of metalloids and the

high coarsening rate of metalloid

dispersoids in a-Ti.

3. A significant deviation of particle

size distribution from the LSW model

is identified in Ti-SSn-4.SLa alloy.

ACKNOWLEDGEMENT

The authors would like to thank the

support of tae Office of Naval Research, VA

(Contract No N00014-82-K-0597). Contribution

# 231 from the Barnett Institute.

REFERENCES

1. S.1. Whang, J. Metals, 36 (1984) 34.

2. S.H. Whang, Y.Z. Lu and B.C. Giessen,Proc. Materials Research Society

Meeting, Nov. 1983, Boston, vol. 8, in

press.

3. C.S. Chi and S.1. Whang, Proc. MaterialsResearch Society Meeting, Nov., 1983,Boston, Vol. 8, in press.

4. C.S. Chi and S.H. Whang, unpublishedwork.

5. E.E Underwood, Quantitative stereology,(Addison-Wesley, 1970) 148.

:I :

2_

PILOT SCALE-ARC PLASMA MELT SPINNING PROCESS

FOR Ti RICH ALLOYS

S.H. Whang, C.S. Chi, and Y.Z. Lu

Proc. Fifth Int. Conf. Rapidly Quenched Metals,Sept. 3-7, 1984, Wurzburg, W. Germany, the proceedings in press.

72

PILOT SCALE ARC HELT SPINNING FOR TI RICH ALLOYS

S.H. WHANG, C.S. CHI AND Y.Z. LU

Barnett Institute of Chemical Analysis and Materials Science, Northeastern University,

Boston. MA 02115. U.S.A.

For pilot arc melt spinning of Ti alloys, thermal behavior of the melt during arc heating has

been studied with respect to arcing conditions. Based on 1 dim. steady state heat flow

analysis, the results suggest that Ti alloy melting requires a large superheat and a steep

temperature gradient while intermediate cooling i!. characteristic at the alloy-crucible

interface in the temperature range of 1200-14000 C Some ribbon and flake Ti alloys were

spun by the pilot arc melt spinning unit.

1. INTRODUCTION and some results of melt spinning experiments

Much past and current research [1-4] using this unit.

suggests that rapidly solidified Ti alloys with

novel additives have distinct potential for 2. EXPERIMIENTS

developing high specific strength and good Current experiments have beer, carried

high temperature properties. In order to using a pilot-scale arc melt spinning unit at

stimulate further research in this direction, Northeastern University. Fig. 1 shows tLy

more flexible processing techniques are front view of this unit. This unit is a:,,-

essential. Although many rapid solidification equipped for other procebsing 'aan:"

processing techniques for Ti alloys have been controls such as arc power, inert gas pressur. p

developed [5-71. the arc mel1t spinning disk speed, electrode rotation speed arnd

2 technique has many advantages over other granule alloy feed rate. The detail

* techniques with respect to contamination-free descriptions of this unit are given elsewher.

processing, high quench rate, and inexpensive [3]. In order to study thermal behavior of Ti-

operation. In this technique, a cold copper 6.3Si alloy button in the cold copper crucibie.

crucible with an orifice, combined with a non- a thermocouple tPt-13Rh) and an optica-

consumable arc electrode for melting and pyrometer (Ircon :000 series) were employed t.

spinning, is employed. This scheme was, for monitor the temperatures of the top an

the first time, applied to Ti alloys on a lab- the bottom suri:'oes. The alloy button has an

- cale (8]. However, some concerns about ellipsoidal shape (a=b=3cm, c=l.3cm) weighin . -.

extensive freeze at the alloy-crucible 50 g. A non-consumable tungsten electrode wa,

interface and effectiveness of high intensity placed above the button with a distance fl) -f

argon arc as a heat source have been raised. 1 cm 11id Ar pressure of 37 Kpa. Temperature

Therefore, it is necessary to investigate readings as a function of current 100, 200,

thermal behavior of the melt with respect to 300, 400, 500 A) have been made at a location

beating conditions and heat flow profiles at between the center and the edge of the surface-

the steady state. This paper presents a when the steady state heat flow conditions have

preliminary study of such behavior of Ti alloy been established.

%

* by increasing superheat arnd total current. At

the interface,. the heat transfer is l imited by

the high interface resistance while at the top

surface, very large superheat occurs, due t J

the poor conductivity of this alloy. Since r. e

arc -e lt thermal conductivity (k) of Ti-6.3Si alloy in

mel-spn poces i- 240C Allhoy Think. 13 mmcoma r men c chamber U iEectrode

44

S800-

F'ront view of pilot~ ar elt spinning unit. .

C 2 343. RESULTS AND DISCUSSIONS NCMINAL '4P',- PC~ - A

The temperatures of the alloy button were FIGURE 2-measured by increasing nominal input power Plot of alloy temperature vs. nominal inpUL -

(Fig. 2). The temperature of the top surface power.

increases rapidly in a non-linear manner,

whereas the temperature of the bottom surface

increases very slowly in a more or less linear E - 0 Ti 6.3 SiE .

fashion. As a result a large temperature 10 P.' Par allIelI To ArcPla sm a

* gradient alIong the c axis, parallel to the 971,

* incoming rc plasma, has developed as shown in z 80Fig . 3 . When the surface temperature (T') was 70

plotted aga inst the square-root of the totala:6

current (1) (Fig. .4), an empirical relat ion was 5* found from the lineAr ccorrelajtiun elct ode*.gr '

6u 40 S et a oe M

o 0 reSiure Afr 3~ (POwhere a anJ b are temperature intercept and -

S20conversicn factor,

In order to understand the situation more 0

* clearl y, a schema tic drawing i s g iven in F ig. I) 2 3 4 5 6 7 j 9

S. By assuming that the heat f low only occurs N4OM IN AL I N P 1T PO E R 'KWIalong the c axis or in the direction parallel

to te icomng ac pasm, te soidiicaionFIGURE 3toth icoin ac lama te oldiictinPlot of average temperature gradient vs

* front moves toward the slloy-crucible interface nominal Input power.

interface, this problem may be simplified into

2800

a plane wall case with the maximum temperature

2600 - at the top surface.0/2400 k(Ts-T I )

6 hI (TI-To ) = (3)

2200

Where:.000 ?!ihI = heat transfer coefficient at the alloy-800_.

1800 crucible interface

INSOO T0 = temperature of the alloy bottom surface

1400 , To = room temperature10 '2 1 16 IS O 20 22 24 260 1 4 1 8 2 22 24 26 X = alloy thickness (c axis)

Current , I (A) Using equation (3), the heat transfer

coefficient at the interface for the molten pFIGURE 4 alloy is estimated as -0.6 W/cm2 -[ at T1 =

Plot of the top surface temperature vs.

total input current. 14100 C. The value indicates that high

interface resistance is characteristic in the

cold hearth melting. On the other hand, at the D

Alloy-Crucible - top surface, heat transfer occurs in the formInterface - Coiction Plasma

of radiation and Ar gas convection. TheTin Radiation radiation can be expressed as:

-r Qr= e T-

SEAtraction Where: Qr heat loss through radiation- - ~Front"-""

Copper Crucible Top Alloy = total emissivityCopItCr~lll -- X . SurfaceI TS.uf c = Stefan-Boltzman constant

Distanco Based on the previous assumptions, the

radiation heat loss at T. = 25000 C is -211

W/cm 2 . The calculation of the heat transferFIGURE 5

Schematic drawing of 1-D temperature profile coefficient by the convection requiresalong the thickness direction, information about the ambient temperature near -

the surface.

the range of 1200-20000 C is nearly constant Current 1-D heat flow analysis is not

(91, the heat flow of the steady state may be correct in a strict sense, since the

expressed using a standard 1-D poisson temperature gradient and heat flux exist along

equation, neglecting heat flow in the radial the radial direction. Future study will

direction (a and b axes) include the solution of the 3-D poisson

d2 T Q equation by a numerical method.-~+ 0(2) Many a-Ti alloys were spun by this unit and

dx2

Where Qi: heat generated by arc plasma some of them are shown in Fig. 6(a) and 6(b).

The solution of eq. (2) would provide The ribbon alloy varies in thickness from 20roi .

temperature distribution as a function of x. to 80pm depending upon superheat and disk

In order to estimate the heat transfer speed. Similar arguments can be applicable to

coefficient (h) at the alloy-crucible flake materials.

-2:. -. . .. ... .....-".-* " f •° ' % a % % % ." *." .jb . • • . . *- .. •••• b. . . . . * . . -• . • = . •

3. The alloy surface temperature increases

linearlv with the square root of total current.

provided other pararetcrs such as Ar ;a

pressure, arc electrode distance are constant.

4. Heat loss by radiat.on fzc tie :o; urf,

reaches ai much as 1 10 ', of the total ; J:

loss at 2500 0 C.

ACKNOWLEDGENENT

The authors would like to thank Mr. Y.'.

Z5 MW Ning for his assistance in carrying :ut tCe~e

FiGURE 6 (a) experiments. The work has been upported

the office of ,aval Research contract -

N00014-82-K-05797) . Contribution -4 u frcM

Barnett Institute.

REFERE:,CES

i . J. Grant, U.S. Patent 30"U4fS,. ,ec.25, 1962.

2. Milton D. Vordahl, U.S. Patent 362Z406,Nov. 23, 1971.

3 .S.H. Whang, J. Metals, 36 (1984 34. ---

4. S. M. L. Sastry. T.C. Peng P. J.2 1m Mtschter, and J. E. O'Neal, J. Metals

35, (1983) 21.

FIGURE 6 (b)-' 5. R. Maringer and C. Mobley, J. Vac. Sc. .

Technology, 11 (,1974) 1ub7.

Figure 6. Spun materials by the pilot arc11elt spinning unit (a) ribbon alloy (b) 6. P.R. Roberts and P. Loe',enstein. in:lake aJoy, Powder Metallurgy oI T-tan Am Aljs. I o,. s

F.H. Froes and J. E. Smugeie k, eds,(TS-AIME 1980) Z1.

7. D.G. Konitzer. K.W. Wa iters, E.L.4. CONCLUSIONS Heiser, and H.L. Fraser, %.et. rrans. 3,

I Melting and spinning of Ti-6.3Si alloy in 1SB (1984, 149.

a cold copper crucible require large superheat 8. S. H. Whang and B.C. Giessen, Proc.Third Conf. Rapid Soliditication

and temperature gradient in the direction Prc ng, Rober MebrainProcessing, Robert .Mehrabian ed. NBS i-

parallel to the incoming arc plasma due to the 1982) 439.

poor conductivity of this alloy. 9. C.Y. Ho, P.W. Powell, and P.E. Liley. J"

2. The heat transfer coefficient of the melt- Phy. and Chem. Ref. Data, vol. 3, 19.4

crucible interface is -0.6 W/cm2K at 14100C supplement No 1, Am. Chem. Soc. and Am.Inst. Phy. for Nat. Bur. Stand.

(above melting temperature).C .

%% %.7

* ~ ~........

MICROSTRUCTURES AND AGE HARDENING OF

RAPIDLY QUENCHED Ti-Zr-Si ALLOYS

S.H. Whanga, Y.Z. LUa, and Y.W. Kimb

Submitted to J. Mat. Sci. Lett.

aBarnett Institute, Material Science Division, Northeastern UniversityBoston, MA 02115

bMet. Cut. Materials Research Group, P.O. Box 33511,Wright- Patterson AFB, OH 45433

77

sS

MICROSTRUCTURES AND AGE HARDENING OF RAPIDLY QUENCHED Ti-Zr-SiALLOYS

S.H. WHANG and Y.Z. LUBarnett Institute of Chemical Analysis and Materials Science,Northeastern University, Boston, MA 02115

Y-W. KIMMetcut Materials Research Group, P.O. Box 33511, Wright PattersonAir Force Base, OH 45433

Ti-Zr-Si alloys (0-8 at.% Si) rapidly solidified (RS) by arc

melt spinning and the hammer-and-anvil techniques were studied to

characterize microstructures and annealing behavior. TEM analyses

show that extensive refinement of microstructures and extended

solubility of Si (up to -6at %) in Ti alloys result from rapid

quenching. All the alloys exhibit a strong age-hardening at 500-

6000C by forming uniformly distributed fine, spheroidal

precipitates. At higher annealing temperatures, the precipitates

become ellipsoidal, which is in contrast to rod or needle shaped

silicides in conventional Ti alloys containing Si. A significant

strength loss occurs after high temperature annealing (850-

950°C/1-2 h) due largely to precipitate coarsening and in part to

grain growth.

1. INTRODUCTION p

Ti-Zr-Si alloys and their modified alloys have been studied

by a number of authors [1-5] with respect to their strength, age .,

hardening and creep strength. These alloys are strengthened by

both solid solution of Zr and precipitates (silicides). The

martensite structure that results from Zr addition provides not

only an additional strength but also numerous nucleation sites for

." o. • - ° -,- o o . . ' .." °-° • - ° . ..% . . . .." . -. • .. - . . % . ° . .• . .' '.. ..". - - -. - . '- - .- ' - o . . ..- " • .•

• '. . 'r " "-.-." " ', - -' ' ' .' ' '-' '° .' _' . " ;, ' ' _ ''.. ..'. . . ..". ., ' ", , " -.. ". . ." =

the silicide precipitation. The precipitates [(Ti,Zr)5Si3 [2]] in

these alloys are more stable than Ti5 Si3 precipitates in Ti-Si

binary alloys, improving high temperature properties such as creep

strength [6].

Recent alloy synthesis through rapid solidification

processing yields extensive microstructural refinement [7,8] and

extended solid solubility of solutes [9]. This, in many cases,

leads to beneficial morphology and sizes of precipitates, aging

response and mechanical properties that are not obtainable by

conventional solidification techniques. In this paper, these

effects are reported for rapidly solidified Ti-Zr-Si alloys.

2. EXPERIMENTS

Ti-Z 1 0-Si 0_ 8 alloys were processed into ribbon (20-40 gm

thick, 2 mm wide) by an arc melt spinning technique and foils (20-

30 gm thick) by the hammer and anvil splat quenching technique.

The injection pressure for this particular run was -8.8 x 104 Pa

and the speed of the disk surface was -18 m/sec. Samples of

similar thickness (-20gm) were used for the experiments in order

to avoid experimental variation resulting from different cooling

rates. The surface topology of ribbons and foils was examined by

SEN and microstructures of foil samples were studied by TEM. For

annealing, samples were wrapped with Ta foil and sealed in quartz

tubings in vacuum. Annealing experiments were conducted in the

temperature range of 350 - 850 0 C for 1-60 h. Microhardnesses of

as-quenched and annealed samples were measured using a Vickers

diamond pyramid tester with loads of 50 and 100 g.

-,-.. ........ ..

3. RESULTS AND DISCUSSIONS

3.1 As-Quenched State

The free surface of ribbon processed by an arc melt spinner

shows characteristic dimples (Fig. la) and the chill side has

wetting patterns and relaxed air pockets (Fig. lb and ic). TEN

observations show that as-quenched microstructures of all the

alloys studied consist of equiaxed dendritic cells containing

extremely fine martensite structure, as is shown in alloy

Ti 4Zr10 Si6 (Fig. 2a). The dark field image (Fig. 2c) of bright

spots of the ring pattern (Fig. 2b) shows that the ring patterns

consist of reflections from the fine cells. Up to the Si content

of 6%, no precipitates were detected indicating that the extended

solid solubility of Si in the Ti-Zr alloy is at least 6 at %.

As-quenched microhardness values are shown as a function of

Si content (Fig. 3). Hardness increase in the Ti-Zr-Si alloys is Ai

a linear function of Si concentration up to 8% Si see (Fig. 3)

which is consistent with TEM observations. The strength

increment of Ti-Zr alloys is somewhat less with Si addition than

with B addition [10].

3.2 Annealed State

Isothermal annealing at 550 0C resulted in gradual

disappearance of martensite structure, nucleation of precipitate,

formation of better defined sub-grain structure (0.1-0.3 pm

dia.), and elimination of dendritic cells, as shown in Ti84Zr1 0 Si6

alloy (Fig. 4a). Fine, spherical precipitates (-30nm) appeAr

along prior dendritic cell boundaries. The corresponding

-.S

diffraction patterns (Fig. 4b) show tiny diffraction spots inside

the first ring, which is in agreement with the appearance of the

precipitates. The precipitates were not identified, but they are

likely to be silicides. A high temperature annealing (8000C-

9000 C) resulted in coarsening of the precipitates (to 0.5-1pm)

and grain growth (to l-3pm) as shown in Fig. 5a. The

corresponding diffraction patterns (Fig. 5b) exhibit many

distinctive spots from the large precipitates (Fig. 5c). It is

not clear whether these large silicides have the same structure as S

that of fine precipitates in Fig. 4. Such significant coarsening

of silicide particles at a high temperature (Fig. 5a) may be

explained by observing particle coarsening in a Ti-5A1-2Si alloy .

[11]. The coarsening rate of the silicides in this alloy system

shows that the cube of particle radius increase per second is -3.4

x 10-26 m3 /sec at 8000 C. The estimated diffusivity of Si in the S

Ti-5A1 matrix is -1.2 x 10- 1 1 cm2 /sec. It appears that the high

coarsening rate of silicides in Ti is attributable to the high

diffusivity of Si in Ti alloys. 0

Microhardness of Ti-Zr-Si alloys increases at 550 0 C with time

up to 20h and decreases with further annealing (Fig. 6). Similar

trends are observed in the isochronal (2h) annealing curve (Fig.7) 5

where Ti-Zrl0 -Si 8 age-hardens with temperature up to 6000 C and

softens significantly above 7000 C. The hardness change is in

accordance with precipitating as well as other microstructural

coarsening. The strong age hardening observed in the rapidly

S-2 '

solidified Ti-Zr-Si foil alloys (Fig. 6 and Fig. 7) is in contrast

to mild age hardening in the solution treated Ti-Zr-Si alloys [2].

This is because RS alloys contain a higher volume fraction of fine

precipitates than their counterparts, conventional ingot alloys.

Another distinctive feature observed in the RS Ti alloys is

remarkably refined grain sizes and martensite structures. It is

noted that the grain size remains within a sub-micron range even

after annealing at high temperatures (850-950°C/1-2h) [7].

ACKNOWLEDGEMENT

We thank the Office of Naval Research for the support of this

work. (Contract ONR N00014-82-K-0597) Contribution #202 from the

Barnett Institute.

REFERENCES

1. K.C. Antone, Trans AIME, 242 (1968) 1454.

2. H.M. Flower, P.R. Swann, and D.R.F. West, Met. Trans. 2(1971) 3289.

3. N.W. Paton and M.W. Mahoney, Met. Trans A 7A (1976) 1685.

4. S.M. Tuominen, G.W. Franti, and D.A. Koss, Met Trans. A, &A(1977) 457.

5. C. Ramachandra and Vakil Singh, Met. Trans A, 13A (1982) 771.

6. N.M. Flower, P.R. Swann and D.R.F. West, in Titanium Scienceand Technology, R.J. Jaffee and N.M. Burte, eds., PlenumPress, New York, NY (1973), vol. 2, pp. 1143-54.

7. C.S. Chi and S.H. Whang, Proc. Materials Research Society,Nov., 1983, Boston, in press.

8. S.H. Whang, J. Metals, 36 (1984) 34-40.

9. S.H. Whang and C.S. Chi, unpublished results (1984).

. .q

10. Y.Z. Lu and S.H. Whang, unpublished results (1984).

K 11. S.H. Whang, Y.Z. Lu and B.C. Giessen, Proc. Materials . -

Research Society Symposium, Boston MA (1983), in press.

FIGURE CAPTIONS

Figure 1. SEX micrographs of as-quenched Ti,,Zr,.Si, ribbonalloy

a. The free surface (xS0)b. The substrate side surface (xSO)c. Detailed micrograph of b (W200)

Figure 2. TEM micrograph of as-quenched Ti4 Zr6oSi foil

alloya. Bright field aicrograph Sb. Selected area diffractionC. Dark field micrograph

Figure 3. Microhardness of Ti-Zr,,-Si alloy system as afunction of Si Content. L

Figure 4. TEN micrograph of aged Ti, 4Zr,,Si, at 5500C, 20 h

a. BFMb. SAD

Figure 5. TEN micrograph of annealed Tio4zr1OSi6 alloy at9500 C, 2 h

a. BFMb. SADC. DFM

Figure 6. Microhardness of Ti-Zr o-Si alloys as a function oftime at 550 0C 1 losa ucino

Figure 7. Microhardness of Ti,,ZroSi, alloy as a function of 0annealing temperature (2h each)

r ... -.. . .

0. 2,yn in 0. 2.n 7

Fiur

Figur 1ur a

Figgure I b

am - ------

Fiqur 2 aFigue 2

b 4S

Figure 2c

Ti90 _xZri0 Si X

5.0-AS-QUENCHED STATE5.0 (Foil Thicknesu: 2Opjm)

4Z-

C,

z04.0

0

3C5

4- - . i Z

001

3.15 1 .0

SI I O ,a

Fi ur/

Figure 4 a

0.5 L0

Figure 4 b

. .

Figure 5 a

O.5/.Lm

Figure 5 D

Figure 5 c

'Um

MICROHARDNESS (DPN),GPa

00

cj,

01 - x

~o

- Cl,

00 9t

M ICR OHARDNESS(DPH),GPa

00

an 0

CD)

z

-,4

0no-n

0 10

oA N

PRECIPITATION HARDENING OF RAPIDLY QUENCHED Ti-Zr-B ALLOYS

Y.Z. Lu and S.H. Whang

Submitted to J. Mat. Sci., Nov. 1984

9 2

PRECIPITATION HARDENING OF RAPIDLY QUENCHED Ti-Zr-B ALLOYS

Y.Z. Lu and S.H. WhangBarnett Institute of Chemical Analysis and Materials Science

" " Northeastern University, Boston, MA 02115

Extensive solid solubility of boron (up to 6 at.%) in Ti can be

6achieved through rapid solidification (the cooling rate of "10 k/sec).

A small increase in lattice parameters is noted with increasing boron

* concentration in Ti. Transmission electron microscope observations show

that age hardening in this alloy is associated with very fine boride

Sr precipitates that are preferentially formed along prior-beta grain

boundary. These precipitates also grow directionally into rod shapes

*. . after annealing at elevated temperatures. A strong age hardening

* occurs from 500OC to 6000C, whereas a significant softening due to micro-

structural coarsening exists above 8000 C.

1. INTRODUCTION

In recent years, research on rapidly solidified Ti alloys has been

focused on the formation of stable dispersion by novel additives in the

Ti matrix [1-5]. In the past, the dispersion strengthening of Ti alloys

has been investigated with Si addition through solution treatment [6-8].

. -By contrast, the fact that boron has little solubility in Ti at high

". temperature makes it difficult to utilize solution treatment. Alterna-

tively, homogeneous Ti-B alloys can be prepared through rapid solid-

ification processing. The succeeding heat treatment would result in

uniform titanium borides (TiB, TiB2), which are attractive as dispersions

because of their good stability at high temperatures.

Previously, the effects of Ti borides on microstructure and mechanical

i'! 1

properties in Ti-5AI-2.5Sn-lB alloy were reported [2,9]. In this paper,

a study on the effects of B addition to Ti-Zr10 alloys will be discussed.

2. EXPERIMENTS

Small Ti -Zr -B and Ti B alloy buttons (1-2g) were pre-SalT90-x-Z10-Bx Tl00_x x

pared by melting pure Ti (99.9%), Zr (99.9%) and B (99.5%) in an arc

furnace under Ar gas atmosphere where x = 0-8 at %. The buttons were

remelted repeatedly to insure compositional homogeneity. Splat foils and

thin ribbons were made by the hammer and anvil technique and by an arc

melt spinning technique [10-11], respectively (Fig. 1). In the arc

melt spinning technique, an alloy button (0.5-1g) is melted in a cold

copper crucible and injected through an orifice (0.5-1 mm dia) centered

in the crucible by applying a gas pressure differential above and beneath

the crucible. The spinning speed for the ribbon processing and the

pressure differential for the injection were '18m/sec and %,8.8 x 104

Pa, respectively. Surface topology of as-quenched samples was examined

by scanning electron microscopy (SEM) and microstructural changes with

heat treatment were examined by transmission electron microscopy (TEM).

Thermal properties of alloy ribbon were studied by differential scanning

calorimetry (DSC) at a temperature range of 293 K to 1000 K with a heating

rate of 200/min. Microhardness of samples (the substrate side surface)

was measured by the Vickers diamond pyramid tester with loads of 50g and

100g.

3. RESULTS AND DISCULSIONS

3.1 Microstructures

Fig. 2 shows SEM micrographs of the free surface (a) and the sub-

strate side surface (b) of as-spun Ti8Zr B alloy ribbons. The freeT8 4 10 6

2

2 .. . . . . . ..

4'- . -.Z

surface is characterized by longitudinal wavy patterns with finely dis-

tributed dimples. The substrate-side contains large air pockets and fine

*" wetting patterns.

As-quenched Ti B alloy did not contain any precipitates, according96B4

to the bright field micrograph (Fig. 3a) and the diffraction patterns

(Fig. 3b). Microstructures of as-quenched ternary alloy Ti8 4 Zr1 oB6

(Fig. 4a) consisted of fine martensite structure, prior-beta grain (1-3 um)

and cellular structure along the prior-beta grain boundaries. The dif-

fraction pattern (Fig. 4b) from Fig. 4a does not show the presence of any

precipitates.

Upon aging at 550 0C for 20 h, the fine martensite structure changed

into lenticular shaped structure while the prior-beta grain boundary

produced G.P. zone-like precipitates (Fig. 5a), which are identified as

spot patterns inside the first ring pattern of the diffraction patterns 77

0(Fig. 5b). The size of G.P. zone-like precipitates ranges up to 100 A in

diameter. When heat treated at 950 0C for 2 h, the precipitates grow

into rod-shaped structures with high aspect ratios (Fig. 6a and b). The

growth of grains is sluggish and their average size is generally smaller

than 1 jim after annealing at 9500 C for 2 h.

3.2 Extended Solid Solubility

Equilibrium solid solubility of B in a-Ti is very small (0.43 at %,

886 C), whereas rapidly solidified Ti-B alloys (cooling rate 105-106 deg/

sec [3]) show homogeneous matrices up to 6 at %B based on the TEM study

(Figs. 3a and 4a). The lattice parameters of the as-quenched Ti-B alloys

measured by x-ray diffractometer are shown in Fig. 7. The x-ray patterns

exhibit both significant line broadening and peak amplitude decrease with

3

increasing B concentration, indicative of lattice strain, probably

caused by boron solid solution. Both a and c axes increase with

increasing boron concentration. In addition, a large increase in the

c axis and a small increase in the a axis results in the increase in

c/a ratio. This trend is in agreement with the case of carbor in Ti - P

as shown in Table 1. Despite boron's having a larger atomic size than

carbon, the lattice expansion by boron is rather smaller than that of

carbon. This discrepancy raises a question--does boron form the

interstitial solid solution, as does the carbon in Ti? First, the

0atomic radius of boron (0.82 A) is larger than the octahedral hole

radius (0.61 A) by about 34.4%; !n comparison, there is a 27% difference

0between carbon (0.77 A) and the hole radius. Therefore, from this

simple argument, it is not likely that all boron atoms reside at the

interstitial site. The absence of B atoms from the interstitial sites

may be attributed to one of the following circumstances: 1) The boron

segregates to the prior beta grain boundary ("-2pm dia). The calculated

average diffusion length is 0.8 - 1.5)im, using alloy cooling rate i0 5

k/sec and diffusivity of carbon in Ti at 1100 0C (representative tem-

perature). This figure indicates a real possibility of a certain degree

of such segregation to the prior grain boundary, though the actual

diffusivity of B in Ti might be lower than that of carbon. A distinct

dark line separating the prior beta grain boundaries, however, was

observed (Fig. 4a). It is not known how much boron segregates to the

boundaries. 2) Another plausible explanation for the absence of boron from

the interstitial site may be that the boron atom occupies the substitutional

site (1.32 A radius) which is 60% larger than the boron atom radius (12n.n).

LEvidently, such replacement leads to the collapse of the substitutional site and

forms a strong compression field center. 3) As a hypothetical situation (since

4

17 F

sufficient evidence does not exist to support this argument, two boron atoms

occupy one substitutional site and one octahedral site (or one tetrahedral)

depending upon energy minimization. It is interesting to see whether or not

- this is a similar situation to the di-substitutional solid solution in rapid-

quenched Gd-Fe [11. 4) As the last possible explanation, the rest of the

atoms may form clusters which can't be identified by ordinary x-ray diffrac-

tion. From the evidences given, this possibility also can't be ruled out.

In order to resolve the site question, it is desireable to measure den-

sity and to study electrical resistivity, Mossbauer spectra and EXAFS of as-

". quenched alloy. In addition, the sample quenched at a cooling rate of l0'

deg/sec of larger should be used in order to prevent significant segration to

the sub-grain boundary.

Thermal behavior of Ti 84 Zr10 B6 alloy foils has been studied by DSC.

* . No visible exthothermic peak was observed up to 1000 K. This indicates that

fine boride formation is not related to the strong exothermic reaction.

3.3 Mechanical Properties

Microhardness increase in Ti-Zr o-B alloys is linearly proportional to B

concentration (Fig. 8), which is in agreement with that in Ti90 x Zr1 0 Six alloys

where x-0-6 at.% [15]. The degree of hardness increment is more

distinct in the B containing alloy than in the Si conatianing alloy. The Ti86

ZrloB4 ,foil alloy shows a strong age hardening resnonse at 5500C while the

Ti82 Zr10 B8 foil alloy exhibits a slight age hardening (Fig. 9). The

"Ti84Zrlo ribbon alloy has lower hardeness than the foil alloys (Ti Zr B8i4Zr10 b 82 10 8

Ti 8 6 ZrloB4 ) and its aging response is relatively weak; Isochronal anneal-

ing of Ti84Zr 6 foil alloy (Fig. 10) demonstrates that a significant

1 softening occurs above 800 C after 2 h due to the precipitate coarsening,

as is shown in Fig. 6a.

* Microstructural features such as early precipitation anld dendritic

5 o[ 5 -.. . . . . . . . . . . . . . . . .

°"* * .........

-. . -°, . -- .

structure along a prior-beta grain boundary suggest that this boundary

is rich with B and, therefore, solidified at the last minute. In the

isothermal annealing curve, the microhardnesses of alloy ribbons is sys-

tematically lower than those of splat alloys. This is because low

0population of fine precipitates ('l00 A dia) are present in the ribbon

alloys, as shown by microstructural study. That is, coarse precipitates

already exist in the ribbon materials at the as-quenched state. This

low cooling of the ribbon alloy for a given thickness results primarily

from the one-sided substrate quench and, secondarily, from the fact

that the ribbon contact with the substrate is not ideal at the cooling

stages, a situation evidenced by air pockets and wetting patterns [3].

ACKNOWLEDGEENT

We would like to express our gratitude for the support by the Office

of Naval Research (Contract ONR N00014-82-K-0597). Many thanks are due

to Dr. Y-W Kim for his help in the preparation of TEM micrographs and

manuscripts. Contribution #203 from the Barnett Institute.

REFERENCES

1. S.H. Whang, J. Metals, 36 (1984) pp 34-40.

2. C.S. Chi and S.H. Whang, Proc. Materials Research Society Symposium,Vol, 28 (1984) pp 353-360.

3. C.S. Chi and S.H. Whang, submitted to Met. Trans, A, Nov., 1984

4. S.M.L. Sastry, T.C. Peng, P.J. Meschter and J.E. O'Neal, J. Metals,35 (1983) pp 21-28.

5. S.M.L. Sastry, P.J. Meschter and J.E. 0'Neal, Met. Trans A, 15A(1984) pp 1451-1474.

6. H.M. Flower, P.R. Swann and D.R.F. West, Met. Trans. 2, (1971) 3289.

7. S.M. Tuominen, G.W. Franti and D.A. Koss, Met. Trans. A, 8A (1977) 457.

8. C. Ramachandra and Vakil Singh, Met. Trans. A, 13A (1982) 771.

6

--. 2

9. C.S. Chi and S.H. Whang, TMS-AIME paper Selection F83-14, September,5 1983.

10. D.G. Konitzer, B.C. Muddle and H.L. Fraser, Met. Trans. A, 14A (1983)p p 1979-1988.

-11. S.H. Whang and B.C. Geissen, Proc. Third Conf. Rapid SolidificationProcessing, R. Mehrabian, ed; NBS, 1982, p 439.

*12. 1. Cadoff and J.P. Nielson, Trans. AIHE 197 (1953) 248.

-13. W.L. Finley and J.A. Snyder, Trans AIME, 188 (1950) 277.

* 14. R. Ray, M. Segnini and B.C. Giessen, Solid State Communication, Vol.10 (1972) 163-167.

15. S.H. Whang, Y.Z. Lu and Y-W Kim, submitted to J. Mat. Sci. lett.,Nov. 1984.

[ 7 1

FIGURE CAPTIONS

Figure 1. Schematic figure of an arc melt spinning unit.

Figure 2. SEM micrographs of surfaces of alloy ribbon.

(a) The free surface exposed to Ar atmosphere.

(b) The substrate side surface contacted with a copper disk.

Figure 3. TEM micrographs of as-quenched Ti96B4 alloy.

Figure 4. TEM micrographs of as-quenched Ti84Zr10B6 alloy.

Figure 5. TEM micrographs of Ti84ZrloB6 alloy aged at 550'C, 10 h.

Figure 6. TEM micrographs of TiB 4ZrloB6 alloy annealed at 950'C, 2 h.

(a) BFM

(b) SAD

Figure 7. Lattice parameters of Ti-B (0-6 at %B).

Figure 8. Plot of microhardness vs. B concentration in Ti-Zr-B alloys

Figure 9. Plot of age hardening of Ti-Zr-B alloys at 5500C.

Figure 10. Plot of isochronal annealing of Ti-Zr-B alloys from 350C

to 8500C.

8

. * -. °

. . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . .

* TABLE 1

*All1oy Lattice Expansion Max. Equil.

per Unit at % Alpha Solubility

AC(A) 9 Aa(A)

Ti 99 B1 (as quenched) 0.0096 0.0020 0.0013 0.00044 0.43

Ti 99C 1(Equilibrium) 0.0126 0.0027 0.0027 0.00091 2 [12] [13]

Ac - a= _

co a,

0

Where Ac and Aa are lattice parameter increments for 1 at %B; c, 4.686 A and a0 = 2.950 A.

9

L ZA

D.C. Powr Suppl

Pressure Gag

In Gas

................................................*..**~** Volvo,*~ c ...........

~.. **' .*~.*%****4

mo

rhYn~

Fig.~- 2- Fig 2b

.0 0 .2 k - - . - -

- ._z: - *~-.- T 3Dr,~ -~- ...-

It P ,p0.2

Fig. 6a

Fig. 6b

Li... O

110 ~~

0 10

00

N ~N0

PL AS-QUENCHED

5.0-

a-~4.5-

z Ti 9 0 xZr,0 BX ~

wo4.0-

Ti Zr Si90-X 10 XIL cI

90 -r

3.51 I

0 2 46 8X (Si or B), at %

Fig. 8

DEVITRIFIED TITANIUM..(U) NORTHEASTERN UNIV BOSTON MR

U NCR BRNETTINST,0F CHEMICL NLYSIS.. S H WHAN DEC/84

UNCLASSIFIED NOee14- 2-K- 597 RP/G i1/6 NL

W26 2.2

WA ___

1.8IIII1.25 fLA~~111 .

MICROCOPY RESOLUTION TEST CHART-' NATIONAL BUREAU Of STANDARDS- 1963-A

* S S 6 6 6 6 S S S S 5%.. '. .. .. -. . .- .. . .. 5%.

.. . . . . .. .%* %..-.- * --- . .S.-. .- * -**.-** . . . . . . . . . . 5.%

o 0o~ 0 0 .

aN O W

x 00

I _

w

0%

41-mcv

Io

I- N 2

%

0~S"Ha SNaH81

0za

04wN

ECc

0 0

zd"H~)S3aV03I

. . . .. ....

FILMED

2-85

DTIC

Documents

DEVELOPMENT PROCESSING AND I DEVITRIFIED TITANIUM..(U) … · 2017. 3. 17. · l d-a149 699 alloy development processing and characterization of 1/2 i devitrified titanium..(u) northeastern