W7 AMP lf - DTIC · W7 AMP .1IP lf.-2_/ ••"DOCUMENT IDENTIFICATION-•k- doccult kc bei= civowd for puobl roleme me and I* i s u , __ DISTRIBUTION STATEMENT ... stresses in the

PHOTOGRAPH THIS SHEET

LEVEL

W7 AMP .1IP lf

.-2_/ ••"DOCUMENT IDENTIFICATION

-•k- doccult kc bei= civowd

for puobl roleme and me I*i s u , __

DISTRIBUTION STATEMENT

ACCESSION FORNTIS GRAM~IDTIC TAB [J p.

UNANNOUNCED__ _D IJUSTIFICATION

,\MAY 261982•,BY CD. DISTRIBUTION / Z :

AVAILABILITY CODES LF. _ _D1ST AVAIL AND/OR SPECIAL DATE ACCESSIONED

DISTRIBUTION STAMP 2

82 05 25 053

DATE RECEIVED IN DTIC

PHOTOGRAPH THIS SHEET AND RETURN TO DTIC-DDA-2

DTIC 70A DOCUMENT PR3CESSING SHEETOCT 79

1--6,

__I q

ANMat1UERHFUDTO F1,1A-

-v ~_ _ _ _ _ _ _ _A ll,

________________'- V,

_____ ___ __ .A ~ OI ~N0, ~

THEOR OF -PAS-C.E -ltT0

1-.1

X. M, Parikh

------- arch 23, 1961

Watertown Arsenal -

Watertown 72, M~assachusetts

Contract No. lAý-11-O212-5O5-OiR-3092OMJ Project No. TBh-002

D/A Project. No. 5B93-32-002TE-CHNICAL REPORT NO. WAL 372/32

-ncl~iss-_ fled

BEST AVAILABLE COPY

ASTIA DOCUMENT NO. _______FILING SUBJECTS-. C-

alloys, refractory .alloys, cemented

DEVELOPMENT AND APPLICATION OF A

THEORY OF PLASTIC DEFORMATION

OF CEMENTE ALLOYS

By •-

N.M. Parikh

March 23, 1961

Watertown Arsenal V• Watertown 72, Massac:husetts

Contract No. DA-11-O22-505-ORD-3092

tip", ORD Project No. TB4-002

D/A Project No. 5B93.3,3-002a TECHNICAL REPORT NO. WAL 372/32

!V3 Unclassified

Qualified requesters may obtain copies o' i-eport from theArmed Services Technical Information Agencj, Arlington HallStation, Arlington 12, Virginia.

R1ARMOUR RES5EARCHN FOUNDATION OF: ILLINOIS I NSTIT"UTE OP TCI4:HNOLOGY

|

DEVELOPMNEr AND APPLICATION

OF A THEORY CF PIASTIC DEFORM4ATION

OF CEMENTED ALLOYS

ABSTRACT

I An investigation was undert~aken to atudyr the frundamental rmasons ~

for room temperature ductility, and strength of the W-Ni-Fe alloys. Phase I

of this program involved the development of the theory of plastic deformation aof such alloys andt Phase II will be concerned with the applicatidon of this

-II

theory to the development of new alloys.In the cemented W-Ni-Fe alloys, the chemistry of the matrix alloy

__ was establi'shed by the use of the electron probe analyser. The sources of

strength and auotilitV of cemented structures in the W-.Ni.-Fe system vweie ea-

tablished by careful detennination of yield strength - grain size relationship

in -the matr-ix alloy as well as tungsten and extrapolating the plots to thejgrain sizes of the components in the ceme~nted structures. The work hardeningI characteristics of the components were also taken into account. The residual

stresses in the cemented W-Ni-Fe alloyF3 were calculated from precise measure-

I ments of the lattice parameters of matrix and tungsten as a function of gross

composition. A cursory application of the principles derived to the develop~-

I ment of new alloys based on tungsten and on iron showed the soundness -.f theseprinciples and resulted in a few new Fe-Cu base alloys.

II

'I

R AM0UR C SFA RC10 ONtDATION VI I INSTITUT OF 1- CNOLOGY

J. J_ - ~ AIW 2182-12

-Interim Technical Report

TABLE OF CONTENTS ISection j]

I. INTRODUCTION . . . . . . . . . . .. . . . . . .

_1 THEORECTICAL CONSIDERATIONS ......... . .. 2

III. DEVELOPMENT OF TIEORY ......... . .......... 6 .1

-IV. EXPERIMENTALTECHNIQUES. 9A. Preparation of Sintered Bars . . ............... 9B. Processing of Matrix Alloy. . . . . •.. . -... . ... •C. Processing rf •ure Tung:.tien irs. ................. * .iD. Mechanical Property J.Masurielant • • ... .... .. 11 IE. Metallogr.iphy and Microlmrdne.o ....... ................ 12 2F. Electron ilicroprobe Analy:3is.. ....... .............. . . ..12G. Residual Stresses by X-ray Mlethod .. ............. 13

V. RESULTS ....... ........................... .4.A. Processing Variables in Sintered Compositions ..... ..... . .14

1. Sintering Tine and Temperature . ........ ..... 152. Initial Particle Size of Tungsten .............. . . .15"_3. Tungsten Content ................... . 15

B. Properties of Sintcred Compositions. ...... .............. .19I. Influence of Tungsten Content ............ ... 192. Temperature Dependonce of Yield Strength of

Sinterud Alloys....................... . .193. Temperature Dependence of I)eflection of

Sintered Alloys .......... ................... . . . .23""4. Discontinuous Yield in Ifigh-Tungsten Bars ...... .23 1

C. Matrix Alloy ............ ... ..................... 9.23D. Pure Tungstn ............... ......................... ..26E. Work-l-larcdening Characteristics... .. ................. .32F. Electron Microprobe Anmlysis................ .32"G. Rosidual Stresses from X-ray Work ...................... 36IH. Metallographic and Microhardneso Work. . ... . . .41

SVI. DISCUSSION OF RESULTS NND APPLICATION.. . .47

VII• CONCLUSIONS. .......................... 57

VIII. LOGBOOKS AN\D CONTIIBUTING "ITRNSON4EL..................... . ....

IX. REREH.,E.5 ..... ................... 60

A RMOUR• R t I A PACiI I k; 1 A 1)A ( I N 1 1 INOI:. IN5 FITU ) OF TECHNOLOGY

-. -- API, 2132-.1.2Interim Technical Report

LIST CF ILLUSTRATIONS

Figure Pg

1 Yield Strength of Steel vs. Ferrite Mean PathIn a Cementite Alloy. ................ .... . 5

2 Schematic Representation of Stress-Grain SizeRelati6nship in Components of Cemented Alloys ........... . 8

3 Influence of Sintering Time and Tc.mperature on theYield Strength of 9OW-IO(7Ni-3Fe) Alloys. ......... 16

Ila Microstructure of 90W-lO(7Ni-3Fe) Alloy, Sintered at11k 1350OC-lhr-Vacuum, "Undersintered" Structure .......... . . . 17

bb Microstracture of 9OW-IO(7Ni-3Fe) Alloy, Sintered at.t l)4750OC-lhr-Vacuum. Typical Li.q~Lid Phase Sintered

Structure .. ... . 175 Effect of Initial Particle Si-,e of Tungsten on Yield

Strength of 90W-IO(7Ni-3Fe) Alloys Sintered at1475 0C-lhr-Vacuum. . . . . . . 18

SOptimum Sintering Temperatures for One Hour of Sintering 2• -Time for Alloys of Various Tungs•ten Contents ......... 20 '

7 Effect of Tungsten Content on Yield Strength andDeflection of Sintnred Bars .... ................. . 21

8 Temperature Dependence of Yield Strength ofSintered W-Ni-Fe Alloys ...... ................... 22

9 Temperatuie Dependence of Deflection of SinteredW-Ni-Fe Alloys of Nickel to Iron lhItio of 7:3. 24.........2)

10 Discontinuous Yield in 95Wv-(7N:i-3Fe) Alloys atElevated Temperatures. Schematic Representation ofStress-Strain Curves 25...................

11 Grain Size Dependence of Yield Strength of Single.•,. Phase Matrix Alloy of Different Tungsten Contents... . 27

4 12 Grain Size Dep-rndence of Yield Strength ofNi-28W Alloy and Copper. 28

4 13 Temperature Dependence of Yield Strength - Grain SizeRelationship ofingl' Phase Matrix Alloy of Composition2)4W-6(7 Ni--3Fný)...) .............. ........................ 29

1)4 Effect of Cold Work on Yield St•rnith and Microlardness of11itrix Alloy of Coiliow;ition lJW-82(7Ni-3Fe) . . . ....... 30

15 Temporaiitu-v and Grajifn Sjize D]opr,ndcore of Yield Strength of

ARMOUR RLSL.APCH -OU AiO f ILF INOI, INSTITUTU O4 A I 1 -TECHNOLOGY

Siv - All . 2].(-2-12Interim Tec! iical Report

LIST OF ILIDSTRATIONS (Coit 'd)I

16 Electron Microprobe Scan of Tungsten Content in theGrains and Matrix of Cemented W-Ni-Fe Alloys... . . 35

17 Residual Stresses in Matrix of Cemented W-Ni-Fe AlloysCalculated from X-ray Measurements . . . . . . . . . 9 . 38

18 Residual Stresses in Tungsten Grains of CementedW-Ni-Fe Alloys. Calculated from X-ray Measurements ..... 39

"10a Relationship Between Weight and Volume Per CentTungsten in Cemented Tungsten Nickel-Iron Alloy . . . . . . 40

19 Microstructures of Cemented W-Ni-Fe AlloysContaining 80 to 93 Per Cent Tungsten...... h2

"20 Sintered Tungsten Grain Size and Interparticle

Separation as a Function of Total Tungsten Contentin Cemented W-Ni-Fe Alloys . . . . . . . 43

20a Interparticle Separation (matrix film thickness)as a Function of Volume Per Cent Tungsten inCemented W-Ni-Fe Alloys. . ... .. .. ..* ...... .* 44

21 Variation in Microhardness of Matrix and TungstenGrains as a Function of Total Tungsten Contentin Cemented W-Ni-Fe Alloys. . . . ... ........ 46

22 Temperature and Grain Size Dependence of YieldStrength of Components of Cemented W-Ni-Fe Alloys . . . .. 8

•>23 Yield Strength of Cemented W-Ni-Fe Alloys and Single

Phase Matrix Alloy as a Function of "Effective"Grain Size ........................ 49

24 The Ftent of Variation in Microstresses in Miatrixof W-Ni-Fe Alloys as Calculated from a Range of

• ~2@ Values* #.. . . . . . . . . . .s. 52""25 Schematic Representation of Yield Stress-Grain Size

Relationship in Iron and Copper. . . . . . . . . . . . .55

A ARMOUR RESEARCH FOUNDATION OF ILLINOIS INSTITUTF OF TECMNOLOGY

-- - v - ARF 2182-12Interim Technical Report

I

I LIST OF TABLES

Table Pae

1. Chemical Analyses of Tungsten Powders . ... . ....... 0

II. Work Hardening Coefficients of Pure Tungsten andMatrix Alloy 24W-76(7Ni-3Fe) at Different Temperatures ..... 33

III. Chemistry of Matrix Alloy as Determined by ElectronMicroprobe Analyzer.............. ..... . 34

IV. Lattice Parameters of Matrix Alloy and Tungsten GrainsV As a Function of Composition in the Cemented Structures. 37

V. Properties of Some Fe-Cu Base Alloys .............. 56 IA- RFOL 1

. AT,21

I Tl

4- ARMOUR R A5S•ARCN FOUNDATION OF I [LItNOI S INSTITUTE Of: TI:CHNOLOGY

1 .- vi. - AIU• 2182-12Interim Technical Report

:1.:DEVELOPMENT AND APPLICATION

OF A THEORY OF i'LASTIC DEFORMATION

OF CEhMENTED ALLOYS

I. INTRODUCT ION

The "cermet" type dispersed systems, wherein strong, hard particles

are embedded in a more or less continuous phase of a ductile matrix, have been

j subjected to extensive investigations in the past decade. Systems in this

category have found widespread applications in the carbide tool industry, in

bearings, and in some structural parts. In spite of -'.:e high strengths of

alloys such as WO-Co, TIC-Ni-Mo, W-Cu-Ni, etc., none of them have exhibited

any degree of ductility at any temperature. One exccption to this widespread

rule was reported by Green et al. They found that some compositions In the

Si-Ni-Fe system with about 90 weight per cent tungsten exhibited good room-

4 •temperature ductility and their strength characteristics were those of pure

tungsten. For an extension of these attributes in the development of new

systems, it must be understood quantitatively why the W-Ni-Fe system exhibits

good room-temperature ductility and formability while its sister systems,

"as W-Cu or W-Ni, do not.

Thus, the basic purpose of this .investigation was to study the

fundamental. reasons for room-temperature ductility and strength of the,

W-Ni-Fe alloys and to attempt the application of these fundamentals to the

development of new alloys. Phase I of this program involved the development

of the theory of plastic defornmation of such oemented alloys, and Phaoc II was

to be concerned with the application of this theory to the development of new

alloys.

In the current program, the objectives set forth in Phase I have been

completed successfully. The chemistry of the alloys in the W-Ni--Fe system has

"Delivered by Armour Research Foundation of Illinois Institute of Technologypursuant to Contract No. DA-II-O22-505--ORD-3O92. Govenment's use controlledby the provisions of Articles 26 and 27 of Title II (General Provisions)."

ARMOUR RE S E A PC F4 F OUN DATION 0P ILLINOIS INS T"ITUtI O) TFCMNOLOGY ]

- 1- ARF 2182-12

Interim Technical Report

be en e stah~la shttid l-ýy the I31t tltil) O'~oh 111 .t1" t ~i 1', T!1i1 ii i'lp (l j k!: (tahlI) S sh

1mg the exaict ('omlJO i t) n o y ' 1 ti. :1ll'1, /.- ..I i Wy 1,1 tip [: .Jiiltt'd C,, ]'I]jcturi2 The

soulrces of stegt and duct ii 1<!C!IC:il3 i ty of' itJ~ý, o 1,) -Ni-Y -1"ey slw were

est~abll shed by careful (IL! tclrai1i; 1.i: of £~h )i, ia .'ult 1 vo; graJin si cC

relationships int thie mAtrix al 17:18VI 'Ii a~ )s ai~:;C lad extrapolating, the1

plots to the` grain 2 siFe of -tic 111u,111 S I i thC -ITl'Alted structures. Thec

work-hardening cactfi'(2c- t:, of theo cutiouua 1k a measiured. The residual

stresses in the abovt-m-~ent U i tod Como 0 1 IIc )I" v Vt) Ln1k u is t Le (I'J renm precis e

measurements of the. wlattice, paran;: Lu s -)I' thec'.ŽmcmeI a., a Iunctz on. of com-

position. A cursory appicloLion of' the- ;u'in,: ilei dur d tenud to the development

of new alloys bacied onl tunigs ll(,i aii on ivu ;e howed theo s3oilfdnf5s, of these

principles and resulted in i low lu~w aM-:;' ,'fn tin'Sves-tigati on has resullt-

ed in the establishmtrent of' aý quli tti'ep i n ewoeen the primary

sources of st~rongl-,.h and due Li .1 i y of~e .rLa te'Lrs

II. THEORYETICAL1 CONS l INIZAl I D3

Duo to thu rather large toiAirl n of theý di snerse-d phase,

(greater than 65 voluime pe~r cezit. Jo1 de o ie I cic sparatioen between

-the dispersed phase: partaclesý (rrloa-.er tha~n 1 1.11'~, I cemetnted structures,

the latter cannot be anal y-rd :iii terse:; ci tlhe ofe~us disape rsion hardenling.

Therefore, a new approachi for thco ann"Lvsi., of ;0211 ceme;nte-d structures was

proposed in this prol'ain. A.;, :sate efor( Lh* oi.tcinentiual ls(as , for

example, in the W.-Ni -*Fc 0y!em hijýJ go llu-gIten Couicen LrTAl ion ITr. oharacter-

ized by the di scontinuoous ori- iciSnc.,d pr t of spheri cal tungsten

gl'ail. till Hr a Uiu1,ljiuous OT' SUIT] coii.i ni ou:-ý PINtni 1'34111(11 z of the orde r greater

than 1. micron) of the Ni *Fc rP h phnase. Ne(w,, in-stead ef consi Befýring the matrix

aas a "fjilm" whidch is not. subj,jc otn i ie.a ss-a has, bee-n thenormal practice to datet: - if' we cunl;ii due the maii ,'x. to he a second phase whose

reffetive grain u~ nis the -,ari 30j tltŽ! il-F Vi ~f be fte matrix filmu

in the cemented a~lto> tie prob.rn kmLco::tos)( iLy act su~sceptible to

arialysi s3

"IDelivered b-y Armour ]kusun3-A ½ ouoie U ]i l ii, Lti ei,,e of Te(ulnologyp ursuant to Contract No. IDA-] -2-CK(ld-lO c'a;"uI s oto eby the p~rovi sions C? Arti c -C) 2t:3w 2! of ", I 1] St :1ra Provi si Ons.

A R A Cp U P1 I P I ' I A P~ti 1 1 U N 1) A It II C) I I I I I N U I N I I T t3 j rI o r I I C 14 N C0 I C) G

¶ interim Technical Report

First, let us examine the stresn-strain rcspornie characteristics of

two-phase alloys using Taylorts(2) pestulates. In view of the fact, that the

alloys under investigation were two-phase, they may be compared with other

two-phase systems which do not necessarily exhibit a cemented structure. The

deformation of such alloys can take place by either of the two mechanisms:

maintaining the stress continuity or maintaining the strain continuity.

Although numerous investigations have been made on the plastic behavior of

two-phase alloys, the data reported do not conclusively obey either of the

two conditions exclusively. It may well be that the deformation of polyphase

alloys is partly isostrain and partly isostress because the two conditions of

equilibrium of stresses and continuity of strains across the two-phase boundary

must be satisfied simultaneously.

"The work of Honeycomb and Boas on o< -A brass supports the iso-

stress hypothesis for the deformation of a two-phase system. Clareboughp(5)

on the other hand, in his work on Ag-Mg alloys and brass reports that the

barder,3phase was more extensively deformed near the boundary than in the

center of the AO cry!Ual, and the deformation of the softer o< phase was

smaller in the vicinity of the P crystal.. Moreover, the harder /A phase

deformed more at the same stress level in the presence of increasing amounts

of softer c. phase. This effect is partially in agreement with the continuity-

of-strain hypothesis. A survey of the past literature indicates that the main

difficulty in clear analysis of the deformation of two-phase alloys has been

"- in the design of meaningful experiments to obtain unambiguous results with the

use of the available experimental techniques.

There is considerable information in the literature on the effect of

microstructure on the mechanical properties of two-phase alloys. Because the

cemented alloys of the present investigation are somewhat like the Fe 3C dis-

persions in a continuous matrix of ferrite inasmuch as the matrix film thick-

ness is considerably larger (greater than 1 micron) than called for in classi-

cally dispersion-hardened alloys (interparticle distance - 0.01 micron), the

"Delivered by Armour Research Foundation of Illinois Institute of Technologypursuant to Contract No. DA-II-022-5(O-ORD-3092. Government'cI use controlledby the prcvisionb uf Articles 26 and 27 or Title II (General Provisions)."

A PMO0U QR ~S EA PC rU N DA T 10 0V LF JL IN 1S IN St I TU TF O F T C C " N0 L0G Y

-3- MY 2182-12Interim Technical Report

KAconcepts developed for the dispersion-hardened alloys would no longer be

tenable for the cemented alloys. A new approach has to be evolved for the

analysis of such cemented alloys. The first systematic study of the effect(6)of dispersions of carbides in ferrite was reported by Gensamer and his

colleagues. Essentially, their results show a reasonable correlation between

the effective ferrite grain size (or Fe C-Fe C interparticle separation) and

the yield strength of the composite alloy. Their data, within a broad scatter

band, reveal that the yield strength of alloys with carbide dispersions in

ferritli steels decreases linearly with the logarithm of the mean ferrite path,

indepr-.dent of carbon content or the shape of the dispersions. Their results

are reproduced in Fig. 1. Thus the strength is purely a function of effective

size of the ferrite grain. Of course, one must keep in mind that the effective

size of these ferrite grains can be much smaller than the one which is observed

metallographically if the interface between the carbide particles and ferrite

matrix is in a state of high stress.

To date, there has been no quantitative investigation of the sources

of strength and ductility of the cemented alloys. On the other hand, the high

strengths in alloys such as WC-Co and TiC-Ni-Mo have been attributed to the

thin matrix film, but no attempt has been made to analyze the systems on a

quantitative basis.

The residual thermal stresses do play a part in the behavior of such

cemented alloys, but there has been no simple correlation between these stresses

and mechanical properties of such alloys.(7) These stresses are readily measured

by X-ray diffraction analysis. I

Lastly, assuming that during the deformation of such cemented alloys

as in the W-Ni-Fe system the microscopic grains and matrix are in a state of

triaxial stress, the effect of such a constraint must be considered. Unfortu- 4nately, there is no easy way of even qualitatively evaluating this effect.

"Delivered by Armour Research Foundation of Illinois Institute of Technologypursuant to Contract No. DA-11-O22-505-ORD-3092. Government's use controlledby the provisions of Articles 26 and 27 of Title II (General Provisions)."

ARMOUR RESE ARCH F OUNDATION O U ILINO1S INSTITUIEC oF TECNOLOGY

- 14 - ARF 2182-12-•I °Interim Technical Report

; -W

iI

120

ii100

80

60

404

20

I I , I ,I

4.0 4.5 5.0 5.5 6.0

Log mean free ferrite path, Angstroms

ig-, I Iffect of the distance between cementite particles 9Z theyl.eld stress of the ferrit-e matrix (after Gensamer(O))

ARMOU RESAH roumATON OF- ILLINOINS YITE TICHNOL oGY

-5 - ARF 2182-12Inte rim Technical Report

77 -. 7 , :j - --~I~- - .;.& -- i

II

III. DEVELOPMENT OF THEORY

It is clear that the cemented alloys are quite different in their

characteristics from the classical dispersion-hardened materials because the

load-bearing capacity of the latter is derived from a critical dispersion of

hard particles at such spacings as to offer the greatest resistance to propaga-

tion of slip. In the cemented alloys, however, the spacings between dispersed

particles are too far apart to offer any appreciable resistance to slip propa-

gation. Secondly, the dispersed phase is not really hard and brittle. And

lastly we are trying to determine the sources of strength and ductility of such

cemented alloys. Since the concept of dispersion hardening are not applicable

to cemented alloys, a new approach is necessary to properly analyze the character-

istic behavior of these alloys. A simple way of correlating the strength and

ductility of cemented alloys (such as W-Ni--Fe) with the individual component

properties would be to determine the strength-grain size relationship for the

-< components. If A denotes the dispersed grains and B denotes the matrix~then

the strength-grain size relationship for the components is given by

= + k 1/

y o y

for both the components and

0 F .d

where 7 = yield strength

" fracture stressF

. stress necessary to initiate motion of dislocations

d grain diameter

and k magnitude of the barrier opposing slip propagation.y

"Delivered by Armour Research Foundation of Illinois Institute of Technologypurzuant to Contract No. DA-l!-022.-•fl-TOt-3092. Government's use controlledby the provisions of Articles 26 and 27 of Title II (General Provisions)."

ARMOU P V LAP CI H F ONDATION OF I LLINOI S NII O II T I F "0FC tOLOGY

6 -- AFT 2182-12 KIntorim Technical Report

let us plot ' vs d relationship for the hypothetical components

A and B as in Fig. 2. Here we see that the C and C for A areF y

both greater than those for B at all grain sizes. What is more important,

the 0 "F for B is less than C for A. So in a cemented alloy of A

and B, application of stress would result in fracture because -F - B is

lower than 5-y - A and fracture would propagate through B. If, however,y

- A were lower than 0- - B, the cemented alloy would be brittle andF y

fracture would propagate through A. Thus, it appears reasonable that for

ductile A - B cemented alloys the conditions to be fulfilled aret

y y

0 -A and -B > and

In the hypothetical case discussed abovethis can be accomplished by

manipulating the O-- and ky values for B by solid solution hardening ory

age hardening of the matrix B. It should be kept in mind that formation of

any hard intermetallic compound in the cemented alloys would impart brittleness.

Thus the grain-matrix interface in cemented alloys should be free of any barrier

to propagation of slip. And slip would only propagate continuously if the flow

stresses are reasonably well matched.

The basic tenet of the theory to be evaluated in the current program

is that the strength and ductility of the cemented alloys in the W-Ni-Fe system

are derived from the close matching of the flow stresses of the matrix and the

S..dispersed tungsten grains. Since these flow stresses would have to be correlat-

ed in the size ranges of the components which are characteristic of the cemented

alloys, a simple means of analyzing them would be to determine the yield strength-

grain size relationships in the components and extrapolate them to the sizes of

the components which are characteristic of the cemented alloys.

To accomplish this, it is essential to know the precise chemistry of

the components in the sintered state. This can be accomplished by electron

microprobe analysis of the components.

"Delivered by Armour Research FouridaLiuln of lllinoi3 In"tituti of Technologypursuant to Contract No. DA-I1-O22-505-0RD-3092. Government's use controlledby the provisions of Articles 26 and 27 of Title II (General Provisions)."

ARMOUR 1F5EARC" FOUNDATION OF ILLINOIS INSTITUTE OF 0 TCHNOLOGY

- 7 - ARF 2182-12S. Interim Technical Report

ii

!IL OF'A •I

-A

"-B

10)

Grain diýmaev'r d-I/2

Fig. 2 nchematic representation of stress-grain size i'elntJonsh±p"in components of a cemented alloy. A graina; B -> Matrix

APMOUR RI U 5 A P C( H 1 . N A T 1 . N 0 F I1LINOIN I NSTITUT[ OF 1 f- C.H NOI. OGY

-' - ARF 2182-12ITterim Technical Report

The significance of the principal residual thermal stresses should Iialso be assessed. This can be accomplished by precise measurement of the

changes in lattice parameters of the components with composition and calculating

the stresses therefrom.

Besides all these factors, the importance of work-hardening character-

istics of the components in the cemented structures cannot, be ignored. These

can be evaluated from the stress-strain curves.

The experimental results should give some idea of the relative

importance of the various parameters discussed above in relation to the mechan-

ical behavior of such cemented alloys.

IV. EXPERIMENTAL TECHNIQUES

A. Preparation of Sintered Bars

Sintered W-Ni-Fe bars were prepared by normal powder metallurgy techni-

ques. The raw materials used were all powders of high purity. Tungsten powders

of graded particle sizes were supplied by Sylvania, the nickel powder by Inter-

national Nickel Co., and the iron powder by the National Radiator Co. The

chemical analyses of the tungsten powders are summarized in Table I.

The required amounts of powders were weighed on a triple-beam balance

and blended in glass jars for four hours on a ball mill rack. Transverse-rup-

ture bars were pressed from these powders at about 6 tsi. The dimensions of

the pressed bars were 2.00 in. x 0.25 in. x 0.15 in.

These bars were then placed in pure alumina trays(Morganite) and

covered with high-purity alumina grit. The trays, which were 3 iri. in diam-

eter, were placed in a vacuum induction furnace which bad a hot zone of

3 in. in diameter x 3 in. long. After the vacuum chamber was pumped down to

about 50-micron pressure, the temperature of the heating chamber was increased

to about 1450*C in one hour. Depending upon the tungsten content of the bars,

the final sintering temperature ranged from l)435% to 1530°C. This temperature

"Delivered by Armour Research Foundation of Illinois Institute of Technologypursuant to Contract No. DA-1I-022-OS-fRDf-3O92. flnvPrnmpnt,'q usp onn+r.rolPdby the provisions of Articles 26 and 27 of Title 11 (General Provisions)."

AP P OUP Q U EA RCI FOUNDATION C.) IIt INOIS IIPS1 T1UTI O1 - T cI WNO( OcY

- 9 - AIRF 2152-12Interim Technical Report

77-1

TABLE I

CHEMISTRY OF TUNGSTEN POWDERS

Analysis Tungsten Powder No. iI II III IV V

Average Particle Size,/( 1 3 5 7 9

-Per cent Oxygen 0.11 0.08 0-05 0.03 0.03

Per cent Nonvolatiles 0.01 0.01 0.01 0.03 0.01

Per cent Molybdenum 0.0046 0.0017 0.001 0.001 0.001

Per cent Iron 0.001 0.001 0.001 0.001 0.001 1 -

4

I

"Delivered by Armouw Research Foundation of Illinois Institute of Technolog>ypursuant to Conit-au No. DA-11-022- 0 -•o0ff-3o92, Government's use controlledby the provisions of Articles 26 and 27 of Title II (General Provisions)."

ARMOUR R• S CA C.I-4 FOUN DATION Or ILLI NOI1S INSTITUT L OP 1Y C-INOLOGY

T -10- AUF 2182-12

.i Interim Technic al Report

was maintained for one hour, and then the power was turned off. The samples

cooled down to a few hundred degrees in about two hours. They were then

unloaded and finished for testing.

B. Processing of Matrix Alloy

Several steps were involved in preparation and processing the matrix

alloy. The raw materials used were Armco ingot iron, electrolytic nickel. and

ferrotungsten. These materials, in the right proportions, were arc melted in

an Argon arc furnace in order to obtain 3 in. diameter buttons weighing about

"L OO gm. These buttons were impossible to cold work due to their chilled micro-

structure, so they were vacuum melted in an alumina crucible and furnace cooled.

This slow cooling improved their microstructure considerably. The buttons thus

produced were surface ground to obtain flat discs, which were then cold rolled

to various reductions and recrystallized at 1600' to 1650'F for different times.

Thus it was possible to obtain test bars (2.00 in. x 0.25 in. x 0.125 in.) with

varying grain sizes from ASTM-1 to ASTM-8.

C. Procesin f PureTunten Bars

Pure tungsten strips were obtained from Fansteel Metallurgical Corp.

These were of the X-ray target grades. The sLrips were cut to the transverse-

Srupture bar size and warm rolled at 50°C in air from 15 to 30 per cent reduc-

tion. These were then recrystallized at 1950°C to 2100*C in vacuum from 1 to

3 hours. The bars thus processed had uniform grain size in a range of ASTM-14

to ASTM-8. Production of larger grain sizes would have entailed a detailed

investigation of recrystallization behavior of pure tungsten, which was outside

* the scope of the current program.

1D. Mechanical Property Measurement

All the mechanical proporty measurements wore made using transverse

rupture bars. Mont of the room-temperature measurements were performed on a

1Hounsfield Tensometer using a three-point loading device. The elevated tem-

perature (up to 1200'F) testing was done on the "Instron" Universal tester

with a four-point loading device. In both casus the sample sizes were main-

"Delivered by Armour flesearch Foundation of Illinois Institute of Technologypursuant to Contract No. DA-lI-.022-50O-OFD-3092. Government's use controlledby the provisions of Artjcles 26 and 2/ of Title 11 (General Provisions)."

A PMOUR R I fAR Il IFUUN[0ATI 0 II IN I. , 1 lH TITUTI ()f T (LCIINOtOGY

- JL. AlF 218-2-1.2Intorim Technical Report

7-!

tained close to the uniform dimension of 1.75 in. x 0.25 in. x 0.125 in.

Due to the nature of the tests, the elongations cannot be reported in con-

"ventional terms; only the deflections are reported. However, by experience ,

it was found that for alloys in this system a deflection of 0.25 in. corre-

- .. sponded to cold rollability of about 60 per cent. -

E. Metallography and MicrohardnessThe faces of the samples to be examined were mounted in bakelite

and polished on an automatic polishing wheel through 120, 2h0 and 4OO grit

emery paper. Deep scratches were then removed by using a 9-micron diamond

V paste on a felt. Final polish was then given to the samples by using Al 03

Linde "All suspension in water, again on the automatic polisher.

The etching of the matrix alloy was accomplished with a dilute

solution of alkaline ferrioyanide. The tungsten grains in pure tungsten

were also etched with the same solution.

Microhardness measurements were made on a Leitz microhardness

: tester using a Vickers indenter. The load was kept uniformly at 15 gm for

all samples to produce small impressions. Larger impressions gave erroneous

results due to the small. sizes of the matrix alloy and tungsten grains in the

cemented structures.

F. Electron Microprobe Analysis

In the technique of electron microprobe analysis, an electron beam I

focuised to a diameter of about one micron is allowed to bombard the region

of a polished specimen from which a chemical analysis is desired. 1he inter-

action between the electron beam and the specimen results in the production

of the characteristic X-rays of the elements present in the bombarded region.

Since the depth of penetration of the electrons is approximately one micron,

the resulting X-rays originate within a few millionths of a microgram of

material. An analysis of this X-ray spectrum from the region of unknown

"Delivered by Armour Research Foundation of Illinois Institute of Technologypursuant to Contract No. DA-II-022-505-ORD-3092. Government's use controlledby the provibions. of Articles 26 nnd 27 of Title II (General Provisions)." p

![

AUMOUA A CA14C$ (-OUNDATION O II lrogs IN TIT TF OU T7[CHINOIOGY t"~~~T 2132-12 . L

. e - 12 - AReF Report1•. Interim Technical Report

chemical composition is obtained with the aid of an X-ray spectrograph and

compared with the spectrum from specimens of known chemical composition. The

"machine used in the present investigation is owned by the Research Laboratories

of International Nickel Co., Bayonne, New Jersey, and employs a flow counter in

addition to a proportional counter for analyzing the X-ray spectrum. The electron

beam in this equipment could be focused down to a one micron diameter spot on the

* -specimen surface, and the intensity and location of the spot could be retained

without drift for long periods of time. The counters had an accuracy of measure-

ment of about 0.1 per cent. The lateral motion of the specimen was made possible

in two directions by a micrometer stage which was free of backlash. The electron

beam was perfectly aligned with the optical axis of the microscope, and the born-

barded spot could be easily seen because of localized deposition of the decom-

"position products of the diffusion pump oil present in the system. The use ofthe flow counter made it possible to analyze for light elements down to magnesium

in the periodic table.

"G. Residual Stresses by the X-ray Method

Residual thermal stresses in the cemented alloys were calculated from

"the precise measurements of the changes in lattice parameters of the componentsS" with changes in composition.

"Lattice parameter changes in the matrix alloys were calculated from

Sthe averages of two reflections, hkl - 311 and 222. All the measurements were

*- performed on a Norelco Diffractometer using iron radiation at 40 XV and 10 ma.

Previous Debye-Scherrer photographs of these samples indicated that

the back-reflection lines were broadened enough so that the e-I o42 components12

of the characteristic Fe Kradiation were not resolved. This line broadening

Swas undoubtedly due to small crystallite size, lattice strains, etc. In order

to obtain accurate measurements, it is imperative to have samples of optimum

crystal size in the unstrained condition. However, this problem was overcome

by scanning and manually counting the angular range over which the reflections

occur. The counting intervals were 0.1 to 0.2* 29. The counts per second

were then plotted against the angle 2@ . The resulting curves represented

"the combined effects of Fe K..I and Kw. Because of the broadness and"Delivered by Armour Research Foundation of Illinois Institute of Technologypursuant to Contract No. DA-II-O22-5o0-ORD-3092. Government's use controlledby the provisions of Articles 26 and 27 of Title II (General Provisions)."ARMOUR Q 15 FA PCWrc 0 OU N DA T f N 0: rIiiL IOIH01 , IN STTIIU T1 Of- TECH-NOLOGY

- 13 - ARF 2182-12Interim Technical Report

a2ymmetry of the curves, the peak 29 values would not be valid. To obtain

-corrected 2 values, the center of gravity of each curve was calculated. ]From this norrected 29 and the weighted Fe K wavelength, the d values

and subsequently the a values were calculated.

- 20 1 (29)corrected 29 = I (29)1 '(29)

2 X O + + K 2weighted Fe K

3R ~d 1 /2 q

Sin 0

a d h + k + 12

'To illustrate the difficulty in measuring the peak 29 value, the

hkl reflections of the matrix alloy in the cemented structure exhibited inten-

sities of only about 5 counts per second greater than the background radiation.

Measurements were made on the as-sintered samples, polished samplcs,

and etched samples.

SV. RESU LTS

A. Processing Variables in Sintered Compositions

The basis of preparation of sintered W-NI-Fe alloys was liquid-phase

sintering. In this technique, a powder compact of this alloy is heated in

vacuum or hydrogen at a temperature above the melting point of the nickel-rich

alloy. The densification takes place by solution and reprecipitation of

tungsten grains in the liquid. To establish the optimum processing conditions,

a single composition, 90W-1O(7Ni-.3Fe), was selected for investigating the

effect of sintering temperature and time and influence of initial particlesize of' tungsten.

"Delivered by Armour Ptesearch Foundation of' ll]inois Institute of Technologypursuant to Contract No. DA-11-O22-505-01R-3092. Government's use controlledby the provisions of Articles 26 and 27 of Titlc II (General Provisions)."

* ARM?4OUR 0 P P 'VAPCII OU tjI)II F) I IN0 'I, I• 1itI ( oT- T t T I( HI IO S Y

i - ]. - AIUE 21 82-1 2

• in t 'rim T chniiical Roport

77 -. : . .. . . _ . .. I_. " ,

I.. Sintering Time and Temperature

1The influence of sintering time and temperature on the yield strength

of the 90W-lO(7Ni-3Fe) alloys is plotted in Fig. 3. All these alloys were

sintered to full density. It can be seen that the alloys sintered at tempera-

tures below lhL4OC have poor yield strengths. This is due to the fact that the

samples sintered at 13500C and 1h00*C were denoified only in the solid state and

exhibit a highly segregated microstructure. Also, at these temperatures there

is no appreciable solution of tungsten in the nickel-rich matrix -- thereby

resulting in a weak matrix. Figure 4a and 4b show, respectively, the above-

mentioned alloy sintered at 1350C and 1475*C, respectively -- the former repre-

senting the "undersintered" microstructure and the latter a good "liquid-phase

sintered" microstructure. From these results it was apparent that one hour

at temperature was sufficient for good samples. Excessive time or temperature

of sintering above the melting point of the matrix resulted in loss of matrix

by evaporation or migration of the matrix to the surface of the samples. The

microstructures of these alloys indicate that almost all the grain growth of A

tungsten is completed within thp first 15 minutes of sintering although the

elimination of voids takes about one-half hour at lhhO°C and above.

2. Initial Particle Size of Tungsten

The effect of the initial particle size of tungsten on the yield

strength of liquid-phase sintered alloy of composition 90W-IO(7Ni-3Fe) was

also investigated. The results are plotted in Fig. 5. It was found that

alloys with larger initial particle size of tungsten were a little weaker and

slightly porous. The optimum particle size of about 3 microns was selected

as a standard for the remainder of the program.

3. Tungsten Content

Due to the nature of the sintering process, the sintering time and

temperature were expected to be different for different ýoncentrations of

tungsten. Series of tests were mede to establish the optimum sintering

"Delivered by Armour Research Foundation of Illinois Institute of Technologypursuant to Contract No. DA-11-022-505-ORD-3092. Government's use controlled,by the Provisions of Article- 2(1 Fnd 27 of Title II (General Provisions)."

A•MOUR RE5EAR('I CHFOUN DATION Of II LIN I 5S T I I U T OF T: iCqNOLOGY

- Lt; - ARF 2182-12Interim Technical Report

200

iL.oc 140

*100

7 15

04 .0152. .

Sitrngtm, or

Fi. Ifuneofsneigtm-adtmeaueonyedsrntof9W2(M 3e loy.HusiidTnoetrtss

100 epon lodig

AR O R1350C O N AIN Or ILN I N~T T r TC H 1~ OG

Intterim techic, hoport

.. . . ' b, j -. •,X -• t -., •i!

7§7

F•ig 4(a) Neg. 19 89 7

A., -I

Fi .4 .(b) Neg 190

II hor-vcu.Nt5udesneePsrcue

,t9\'k%*"-\i

Fig 4(b) Mirsrutr of 90-07i3e sitee-ai4

'" A) ' ,, ) P

"1 - "\ 2182-12

- Tehia R97-pr

Fig. 4(b) Neg. 199701

i.h~)Mcrsrutr o 0Wl(Ni-Fe '1trdat15"i hor - acuu. Noe "nderintee4• ~cucure

Fig. ~b) Mcrost Fitu. o(b 90-1(Ni3eg. 19900 datih5

1. hour - vacuum. Liquid-phase sintored structure.

- 17 - ARF 2182-f2 I .Interim Technical Report%.

]I

, ~~250 ,

IJ

' II

"ElE

200 1

I /I-4'

150i

4.)

~100

Initial particle size ojf tungsten, microns

Fig. 5 Effect of initial particle size of tungsten on yield strengthof 9nlW-lO(7Ni-3Fe) alloys sintered at 1475°C - 1 hour - vacuum.Hounsfield sintered at lh75°C - 1 hour - vacuum.Tensometer tests. Three-point loading.

ARMOUR RESFARCHI FOUNDATION OF ILLINOIS INSTITU 7 OF TrCHNOLOGY

""-18 - APF 2182-12Interim Technical Report

temperatures for one hour of sintering time for compositions containing

75 to 95 per cent tungsten and balance nickel and iron in a ratio of 7:3.

The results are plotted in Fig. 6.

B. Properties of Sintered Compositions

1. Influence of Tungsten Content

The effect of tungsten content on the yield strength of the sintered

bars is reproduced in Fig. 7. The measurements were made on the Hounsfield

Tensometer using a three-point loading device. It may be noted that 'samples

containing less than 80 per cent tungsten show low yield strengths and a great

deal of scatter. This is because it is impossible to avoid gross segregation

of tungsten in these alloys. Although the data reported in Fig. 7 refer to

measurements made on alloys with a nickel-iron ratio of 7:3, changn-" 4 h1is

ratio to 6:h does not alter the picture measurably.

The influence of tungsten content on the u-tal deflectikn of the

sintered bars is also plotted in Fig. 7. The deficutior. fairly constant

until high tungsten contenus are reachod, whereupon it begins to drop -- most

probably because the alloys with high tungsten content begin to exhibit prop-

erties similar to those of pure tugsten due to the continuity of the tungstc.

grains,

2. Temperature Dependence of Yield Strength of Sintered Alloys

Yield strengths of sintered alloys containing varying percentages

of tungsten were measured in the Jn-vtron Universal tooter using a four-point

loading device. The results are reproduced in Fig. 8. For the bars containing

less than 90 per cent tungsten, the strength drops evenly with increasing

temperatures. However, for the higher tuxigstien bars, there soeins to be no

further druo after hOO°F up to 1.2001. This may be due, again, to the fact

that, for alloys with high tungsten contcnts, there is very Little inatrix

"and we may well be measuring the y-eld strength of pure tungsten. These,

tests were all performed in air.

"Delivered by Armour Research Foundation cf ]l]inoelfi Institutc of Technology,pursuant to Contract No. DA-1I-022-5TY-OGPD-3092. Govvenrmentl'ýi use control] oi

by the provisions of Artictes 26 amd 27 of 1':t.1.c] (1.1 ral ]'Movi •Jun;)."

A R M 0 U U P ,f P C F U 14 Dt ATI N 1, 1 LIr I' N T I I I. C) 1 ( N ()

- ].} -. .(. 21.-1.? ..

'1525

1500

11 475

1451

0))

I Io5 90 95 100

Total. tungsten content

Fig. 6 Optimum sintering temperat,-re for one hour of sintering timefor alloys of variouos tungsten contents and nickel to Aronratto of 7:3.

A P I o J L P r F A P•C' . H F 0(UN OATIO1 O-F ILLINOC)I5 I H II IILJUI OF TICHI N(1I O( Y Y

--20-- A,' 2182-)2

lntor'iz• Technical Roport

]

.,3

225 0.5deflection

200 04 .

Nma,

8,

" 175 0.3 J04-)

150S• 15o - 0.2

Yield strength

125 0.1

- 4 - - - I . .. I80 9o 95 100

Per cent total tfungsten

Fig. 7 Effect of tunynteii content on yield strength and deflectionof cementpd W-tii-Fe 31loy- .4th rickel to jron ratio of, 7:3.Hounsfield Tersometer tcets. Three-point liadin.,.

ARMOUR RA S GAARC W FOUN DATION OF ILLINOCS 1 N ST ITL) I C OF T"CWNO(OGY

--21- ARF 2182-12Interim Technical Report

. . . . . . .. . -L: . . . . " " . -

.. 0 95 tungsten

S••• @Q 92 tungsten -

11

S•200 -

-GO 20 o0 6,00 S0O1O 0

"" 'r~~~~~est temperature 'F ':--------•

Fig. 8 Temperature cependen(e of yield strength of Cemented W-Nit-Funten

• 11Lloy with jiicke]. to ironi ratio of 7:3. Instron tests.. ~F'ur-pocint I oadiii., •

,ARMIO UQ A[ WISFAIRCL F:OUJN IAYION OFý it I, IM4OI INSTIIUjT[ OF r'[C)IM OkOGYtI

, ~~~Lur~im Tecimical. Report ,-

I ", " 1I

"3. Temperature Dependence of Deflection of Sintered Alloys

The effect of temperature on the deflection of sintered alloys is

plotted in Fig. 9. It can be seen that there is no measurable change in Ideflection with temperature of alloys containing less than 90 per cent tung-

sten. However, the alloys with higher tungsten content actually exhibit an

increase in deflection (or ductility) with increasing temperatures. This

phenomenon again corroborates the view that properties of high-tungsten alloys

(> 90 per cent) are determined by properties of tungsten, and the increase in

deflection can be explained by the fact that, with increasing temperatures,

the brittle-ductile transition zone of tungsten is entered upon. In the initial

experiments the limit of deflection of these bars on the Instron was 0.3 inch

due to the geometry of the knife edges. The knife edges were later modified

for measuring higher deflections. However, the results are unambiguous andreproducible. It may be recalled that for alloys in this system a deflection

of about 0.25 inch corresponds to cold rollability of about 50 - 60 per cent.

4. Discontinuous Yield in High-Tungsten Bars

It was observed that samples with greater than 90 per cent tungsten

exhibited a di_,.uiitinuous yield, the magnitude of which increased with increas-.

ing temperature. The results are reproduced schematically in Fig, 10.

C. Matrix Alloy

According to the concepts outlined in Section III and procedures

described in Section IV• the grain size dependence of yield strength of the

matrix alloy was measured in the Instron tester. The electron microprobe

analysis of the matrix alloy showed a composition of 2hW-76(7Ni.,3Fe). How-,

ever, arc-cast and vacuum-melted ingots of this alloy were almost impossible

to cold work more than 15 per cent without cracking. The alloy buttons which

were cold wor'ked less than 15 per cent recrystallized only to large grain

sizes. This problem was overcome by dropping the tungsten content of the

single-phase matrix alloy to 18 and 15 and measuring the yield strength-grain

size relationship in these alloys. These results, combined with the same

"Delivered by Armour Research Foundation of Illinois Institute of Technologypursuant to Contract No. DA-1I-022-.505..ORD.-3092. Government's use controlledby the provisions of Articles 26 and 27 of Title II (General Provision-)."

A Q ?O. 0U R E 5[EARCH FOUN DA 7ION OF ILLINOIS IN 5TITUT E OF TECHNOLOGY

- 23 - ARF 2182-12,'1Interim Technical Report

INN'

J 44

0.,

0 £ig2l

20 40-i 8010

Tetteprtue3

onei 90 Tunenia eo

C 4)H CO4-'

0 r

0

0

I~V "A UC

Str4e;'

tol

Q).

A~~t1OUA~ 0)~AC HONAIN O LIOS IS~~L O FHOO,

ARF 248-1

----------------------------------------------- -> 0nei ehnclRpr

relationship for large grain sizes for the matrix alloy containing 24W,

enabled the results for the matrix alloy of 24W-76(7Ni-3Fe) to be extra-

polated with confidence. The room-temperature grain size dependence of

yield strength for matrix alloys of three different tungsten contents is

plctted in Fig. 11. Extrapolation of the plot for 24W-76(7N!-3Fe) provide

a value for the yield strength of this alloy at a grain size comparable to

that found in the cemented structure (i.e., the matrix film thickness).

A comparative study, the same relationship was determined for a

binary Ni-28W alloy and also for copper. The results are plotted in Fig. 12

and will be discussed later.

The temperature dependence of the yield strength - grain size re-

lationship in the matrix alloy of composition 24W-76(7Ni-3Fe) was investigated

and is plotted in Fig. 13.

During the investigation of the recrystallization characteristics

of the matrix alloy, a systematic study was made of the properties of the

cold-worked alloy. The effect of degree of cold work on the strength and

microhardness of the matrix alloy of composition 18W-82(TNi-3Fe) is plotted

in Fig. 14.

D. Pure Tungsten

Commercially pure tungsten strips were obtained from Fansteel

Metallurgical Corporation. These strips were warm worked at about 550@C

and recrystallized at various times from 1950°C to 2100°C in vacuum. It

was difficult to obtain a very wide range of grain sizes for pure tungsten

and especially difficult to grow grains of sizes larger than ASTM-14. Thbe

temperature and grain size dependence of yield strength for pure tungsten

is plotted in Fig. 15. If these results are compared with those reported

for the matrix alloy in Fig. 13, a close matching of the respective yield

strengths of the two materials is found for the grain sizes which are

characteristic of the cemented structures.

"IDelivered by Armour Research Foundation of Illinois Institute of Technologypursuant to Contract No. DA-1I-O22-5O0-ORD-3092. Government's use controlledby the provisions of Articles 26 and 27 of Title II (General Provisions)."

ARMOUR RF• S ARC H FOUN DATIOT O1 ILNINOIS LI 'INT I T1 ST I OS T II C TE HO L. GC Y

- 26 - ARF 2182-12

Interim Technical Report

ItsA I

200 A A

* 166 o

0

"A ,9, /

s 120

. 0 14 tungsten

hO O 15 tungsten

C, AA0

1--.

10 15 20 25

SAverage gair. diameter, d-1/2 -1/2Averae grar2 ,mm

Fig. 11 Grain size dependence of yield strength of single phasematrix . of different tv1n3ten contents arld nicKelto Jroln rati O 73 . . Teometr tests,Three-point loadinig.

ARMLUR PES:ARCH FOUNDATION 01: ILLINOIS IN STIT U.TV" F. TCCWNOLOGY

--2'1- ARF 2182-12Inter-,mTecniaiL

200 /

/ i

160

H/

80 120 /S320 /

g /a)4

i N 80

* Ni-2d aloy i

ho i

10 15 U25

Average grain diameter, d1/2, mm"/2

Fig. 12 Grain size dependence of' yield sitrength of sinrlc phaseNi-28W alloy and copper. Hounsfield- Tensometer .ests.,TI-ree-point loading.

AIMOUR R I S E A RCP4 FOUNDATION OF ILLINOIS INSTIIUTUFr OF TECHNOLOGY

28- ARF 2182-12Inte• im Technical Report

--• • . ... . • •• • . . . • . .. " " •" -. - " '' ' - ' - -i i

7

200 7

T Z""o,

/ <--

75 O 75°

0

-4ý

4 125 2 i

.rIn

1000 E3 Room temperature

0 75cO0F

01000-F

48 12 16 20

.,rae grain diameter, d32,r&

KF g. 13 Temperature dependenen of yteld strength-grain siz~erelationship of matrix alloy of composition 24LW-76(7Ni-3Fe).Instron tests. Four-point loading:

ARMOUR RIE5ARCW FOUNDATION O ILLINOIS INSTITUTE OF TECNNOLO.'Y

-29- ARF 2182-12Interim Tochnical Report

250 500

200 -00

150 300

H

ID4

00

100 200

50 -100o0 o

I !.I20 4o 60 8o

Per cent cold work

Fig. 14 Effect of cold work on yield strength and microhardness ofsingle phase matrix alloy of composition 18W-,2(7Ni-3Fe).Hounsfield Tensometer tests. Three-point loading.

ARMOUR RE SFARC1- FOUNDATION OF ILLINOIS INSTITUTE OF TECHNOLOGY

-30- ARF 2182-12Interim Technical Report

2200

= -

175

o o 0.15

.~125

A

(00-100 0 40OF v

0 750°F O

75c 1000-F

4 i I I I1 12 16 20

Average grain diameter, d-/2 m"I/2

Fig. 15 Temperature and grain size dependence of yield strength ofpure tungsten. Instron tests. Four-point loading.

ARMOUR QFSCAQC W FOUNDATION OV ILLINOIS INSTITUTE OF TECINOLOGY

31- ARF 2182-12Ir.terim Technical Report

"9. Work-hardening Characteristics

It was postulated in Section III that, in addition to the flow

stresses of the components, their work-hardening characteristics may also

play an important role in the plastic deformation process of the cemented

alloys. Therefore, their work-hardening coefficient h was calculated

from i

II

using the plastic region of the stress-strain curve for a few grain sizes.

The results for pure tungsten and for the matrix alloy of composition

2hW-76(7Ni-3Fe) are summarized in Table II. These values are only approxi-Smate due to the nature of the tests, but they give a rough indication of

the work-hardening characteristics of the components.

F. Electron Microprobe Analysis

The results of the electron microprobe analysis of the cemented

alloys are summarized in Table III. The results are self-consistent in

that there is no appreciable difference in the solid solubility of tungsten

in the nickel-iron rich matrix regardless of the over-all tungsten content

of the alloy. Also, this limit is 24 per cent tungsten, as against 18 per

cent reported in the first investigation. The nickel-to-iron ratio remained

fairly close to 7:3 in the cemented alloys.

In a separate experiment, a careful scan was made of the variation

of tungsten content across a tung:tai grain-matrix interface. The readings

were taken at one micron intervals, and these intervals were decreased to

1/3 micron near the interface. The results indicated a sudden drop in tung-

sten content from the grain to the matrix at the interface and no gradual

change in composition. The results are plotted in Fig. 16.

When the nickel-iron ratio in the cemented alloy was changed to

6:h, small traces of a new phase, Fe 3 W2 , were observed with this instrument.

'Oelivered by Armour Research Foundation of Illinois Institute of Technologypursuant to Contract No. DA-11-O22-5-OS-ORD-3092. Government's use controlledby the provisions of Articles 26 and 27 of Title II (General Provisions)."

APMOUPR CSEAPC R FOUNDATION OF L S L I INSTIIUTE OF 7ECHNOLOGY

- 32 - ARF 2182-12Interim Technical Report

TT

TABKE II

Work-hardening Coefficients of Pure Tungsten

and of he Alloy 2LW-76(7Ni-3Fe)•- ~at D_. , Temperatures

Temperature 'F Work hardening Coefficient, h, In psi

Pure Tungsten Matrix

Room brittle 11.8 x 102

400 brittle 10.2 x 102

• .. 1 2 1 ~ ~102'-'150 10.) X 1 10.5 x 1

2' 21000 88X 1 9.2 X 102

V'Delivered by Armour Research Foundabion of Illlnols Insti.tuto of Technologypui-t]unrit to Contract No. DA-II-022.50.-ORD-3OP?. Cover'nmenb's use controlled.,y thn provisions of Articlen 26 and 26 of litleh 1. (GonerTal Provisions)."

AP. OUR I? A RL4 F COUN4 AT1ON OF ILLINC0 f NT ITUT OF T CHH 0LOGY

- 33 - ARY 2182-12Interim Technical Peport

.................................................

TABLE TIT

Chemistry of' Matrix Allo~y

aE, D(eeiminined by E~lectron Microprobe Analyzer

TungstenNominal Composition Structure in Matrix Alloy

lOW-90(MN-3Fe) Single phaqe 10.1

l5W-85(7Ni-3Fe) Single pl: -se 1J4 .9

18W-82(7Ni-3Fe) Single phase 17.8

20W-80(7Ni-3Fe) Sj 1g le pjhLaL! 20.2

22W-7 8(7i-3Fzi) -Sjng le 1)ha.-,,e 21.8

24W-7 6(71-.3Fe) S j r 1. phafne 24.1

26W-7h('7Ni-3'Ye,) S.J.n glo plaie 2.

~I.72~i-7 )Sing le phase 23.9

8OW..207We ie Cw' t ed 24.3

90W- .()U( 7Nj-3'e) C(,niuted 2h.2

91.W- 9(,/Ii-.3F(,) Celw~n~tcd 2-1.9

9(W- 7 (7N'- -31-e' G : Coot ed 24 . 2

.5 -5 T ---.7T.- 2h , 3.. ~ .- -~ - -. ~.-

I'Drelivev'ocd( by A rwmour Hco-ýjo J I,:: Folndaiti on of .1 1.inol .3 ln~jtitulie of Technologypu rmuint, to (Gnrct i.D-i02-Q;-~?-3V?.covcrnrnoot,;J. use controilea

by o p() uVi n fAtcV 2 rid I-/ of' T)i i! 11. (Gonor.,],. Provisaio~rzs).

X nto-riin Tec hnica.el W;,per t

iii100 -- 0---00

, Tungsten Tungsten-grain grain

- MaLrix film -

" 60.Ht

SI I I

200

10i 15 20 25"

Distanca In microns

Fig. 16 Electron microprobe scan of tungsten ,optcnt of grains andmnatrix in a typi.ic.l ýamarl oal ceienLed alloy of composition90W-l0(7V7i--3Fe).

A.p M 0 uP P 5 LAQ 11 F Ut AT I N O I I I P I I.I I TI TI Or ' L.•C, 1 P4O L O GY

3rl, AF 2?I2-L2Int(ernim T•jchnical Uopozt

This phase could hardly be observed with a metallograph. It appeared as a

discontinuous light green phase and had no embrittling effect on the alloys.

G. Residual Stresses from X-Ray Work

By the techniques described in Section IV, the lattice parameters

of the matrix alloy were measured for a variety of compositions. The param-

eter measurements were made on "as-sintered" samples, polished samples, and

polished, etched samples. The results are summarized in Table IV. In all -4

cases, the matrix alloy 24W-76(7Ni-3Fe) was used as a standard of lattice

parameter in the unstrained condition. Due to the small percentages of the '

matrix alloy in the cemented structures, the intensities of reflections were

low. However, the technique described in Section IV gave reproducible results.

The residual stresses on the matrix were calculated from the simple

relationship:

d

where -I and are principal residual stresses

E is modulus of elasticity

.is Poisson's ratio

d is lattice parameter of theS!matrix in cemented alloy

d is lattice parameter of single-phaseU matrix alloy

The residual stresses in the matrix of cemented alloys are plotted

in Fig. 17.as a function of volume per cent matrix. The results are repro-

ducible and, therefore, quite reliable.

The residual stresses in tungsten grains of the cemented alloy3 are

plotted in Fig. 18. They are much smaller in magnitude, probably because of

the large grain size of tungsten. The relationship between weight and volume

per cent tungsten in these cemented alloys is shown in Fig. 18a.

"Delivered by Armour Research Foundation of Illinois Institute of Technologypursuant to Contract No. DA-11-O22-5O0-ORD-3092. Government's use controlledby the provisions of Articles 26 and 27 of Title II (General Provisions)."

ARMOUR RF5 1ARCH FOUNDATION OFS LLIN0I1 N STITUT- OF TFCHNOLOGY

- 36 - AU' 21.82-12Interim Technical Report

"AL'~

TABLE IV

Lattice Parameters of Matrix Alloy and Tungsten Grains

as a Function Composition in the Cemented Structure.

Samples were Polished and Etched.I

Nominal Composition Structure a (A) Matrix a (A) Tungsten0 0

2hW-76(7Ni-3Fe) Single Phase 3.594

80W-20(7MN-3Fe) Cemented 3.602 3•164

85W-15(7Ni-3Fe) Cemented 3.601 3.165

90W-10(7Ni-3Fe) Cemented 3.600 3.167

93W- 7(7Ni-3Fe) Cemented 3.592 3.167

96W- 4(7Ni-3Fe) Cemented 3.592 3.166

Pure Tungsten 3.165

"Delivered by Armour Research Foundation of Illinois institute of Technologypursuant to Contract No. l)A-11-022-%O%-ORD-3092. Government's use controlledIby U16L prQIS1.01i-1.5~ of± JAtic 2J4S and '3'7 TtbJ

ARt4OU 0 RFU P A RC 1l4 VQUN ATION O1 H ILI1.1NC I SI5 TIT(j T OF IF.CHHNOLOGY

- 3y - ARF 2182-12

II t,urim ... c, , I Report

- 0 0

+o

,-4CV

.41 T1

SeJ 40

o mj

S.- I (I.

4.1 )

P;J 43 4

o ]oo'-.-

0

+)

C1. I- -43

Compressive 'esnaile

Priiicipal residual stresses on matrix X 1000 psi

ARMOUR RESEARCH FOUNDATION OF I LLINOIS IN 5t ITUTE OF TECHNOLOGY

"ARF 2182-12I IInterim Technical Report

Obg

"-li

orl

0

' "A

tt J

b 4J

r-4 0

C,

4a3

SI 1 I ! . .. I IV

4-1

coSo o o o4

-Compressiv e TensileH

Principal residual stresses on tungsten grains X 1000 psi

*•'A R•MO U R R ESE •ARC H F•O U N tDATIO N O F ILLINO IS IN ST)ITU T E: OF T EC H-iNO LO G Y

Interim Technical Report,

IE

P.-I

C• ', I

c~j

o N•

- 0

uae IOunj ýuao jd ovznoI

ARMOUR RESEARCH FOUNDATION OF ILLINOIS INSTITUTE OF TECHNOLOGY

- ho - AJF 2182-12Interim Technical Renort

H. Metallographic and Microhardness Work

Typical microstructures of W-Ni-Fe cemented alloys containing 80 to

96 per cent tungsten and balance nickel and iron in a ratio of 7 to 3 are shown -

in Fig. 19.4.

Attempts were made to investigate the modes of dofornation of these

alloys metallographically. The transverse-rupture bars of these alloys were .

strained to varying degrees prior to fracture. The test pieces were polished

prior to testing. It was found that large "pools" of the matrix in these

alloys showed a fair number of slip bands prior to reaching the elastic limit.

This may well be a simple grain size effect wherein the deformation takes placeon a microscopic localized scale. However, once the stress levels increased

beyond the elastic limit, there was extensive deformation of the tungsten grains

as well as the matrix and the fracture path was equally through the matrix and

the grains.

Samples of binary alloy compositions- which had similar microstruc-tures but exhibited very little ductility -- such as 80W-2ONi anid 8OW-2OCu were

also tested inma similar manner. In the Ni--W system, deformation seemed to be

more predominant in the tungsten grains, and the fracture path was also mostly

"through the grains. In the Cu-W system, however, deformation was only slight

through the copper phase (the tungsten deformed not at all), and the fracture

path was almost entir-ely through the copper phase. It was not possible toobtain good photomicrographs because of the displacement of large amounts of

material from the plane of the noliL3hcd surface.

For the series of W-Ni-Fe alloys containing varying percentages of

tungsten, the metallographically observab~e parameters are summarized in

Fig. 20. Here the average sintered tungsten grain size and average inter-particle separation (matrix film thickness) is plotted as a function of per

cent total tungsten.

"Delivered by Armour Research Foundation of 1]linois Institute of Technologypursuant to Contract No. DA--II-022-05-CRD-3092. Governrment's use controlledby the provisions of Artinleps 26 arnd 27 of Title TI (6nnornl Prnvi.5lnnn) "

APMOUR R7-SFARCt4 F-O.U N DATION .)- IIt lIN.IS iN'TIT UT} I W [(IIN 1O GY

S- hl - AR],' 21.82-12 -.J Interim I'echnical ite.port

5; ''-

A ¾ (2• 1

5oOX 50ox

Fig. 19(a) Neg. 19896 Fig. 19(b) Neg. 1989485 per cent tungsten 80 per cent tungsten

• ... .. 4 , ' -- I • '

90 p r cet t ngst n 9 per cent tungsten

A

A1-

V. W.

Fig. 19 Microatructures of cemented allays containing 80to 93 per cent tungsten and balance nickel and .iron in a ratio of 7 to 3.

42 ARF 2182-12Inerim Technical Report 41-

IA

! -5o -- 1.0 !

0

74

Ni

320 T

"I I

@05

1&4 00

C-I

100

75 60, 85 90

Per cent total tungsten

Fig. 20 Sintered tungsten Frain ,;ize and Interparticle separation(matrix film thiclne,•n) at, a function of total tungs•tencontent in cemented W-Ni.-Fn a].oyv with C-Irkn] to _iron

ratio of 7-3.

APRMOUR Q F.S C ARCt4 FOUN DA71ON QP ItL I N01IS I NST I TUT, C OF T FC V4NOLOGY

•' ,-•ARF 2182-1?

J 03 1-4 1.•(! " I '• p r

The separations between adjoining particles were measured by using

a Hurlbut counter in conjunction with a micrometer "Filar" eyepiece. A plot

of interparticle separation (matrix film thickness) versus volume per cent

tungsten is reproduced in Fig. 20a.

10,

0

F T O

. - . 6U1)

i0)

SInterim TcnclR-pr

.r'1

S6o 70 80 90 3.oo

,Fig. 20a INTERPARTICLE SEPARATTON (MATRIX FILM4 THICKNESS) AS A

S~FUNCTION OF VOLUME PER CENT TUNGSTENIN CEMENTED

TUNGSTEN-NIC1 IKE I,-]N A IIOYS

AR•MOUR• R S:ARCr-Al• FOUNDATION OF ILLINOIS INSTrTUIF OF: T'lCRlNOLOGY

TI

--- 4-, AllF 2182-2..2ntrmTechnical Report

The density of 24W.-76(7Ni:3Fe) matrix alloy is 11.1 gins/cc.

The variation is 1nicrocharness of the tungsten grains, as well as

the matrix, with varying composition is plotted in Fig. 21. H

Several attempts were made to elucidate the state of stress of the

tungsten grains in the 90W-l0(7Ni-3Fe) alloys. These bars were cold rolled

various degrees and polished. The tungsten grains deformed from spherical

shape to oval shape after rolling. Careful microhardness measurements of

large deformed tungsten grains from the center of the grain to the edge Ufailed to reveal any measurable difference in microhardness. This lack ofii

indication of variation in stress state does not necessarily mean that it iA

is not present. More likely, attempts to measure the stress variation by i

microhardness failed because the tungsten grains are too small (- 50 microns) iito enable any more than four readings to be made across a grain. The impres- H

sions with lowest load (25 gni) range from 9.5 to 10 microns across. Thus thesensivity of such a measurement on the rather small tungsten grains is poor. i

In determining the hardness of the matrix in cemented alloy micro--

structures, we had to restrict ourselves mostly to interstitial areas. For

a 25 gm load the minimum matrix film width required for accurate measurements

would be about 10-12 microns for alloys studied.

Attempts were also made to study microhardness variations with a

diamond scratch hardness indenter. In this method, the pyramid shaped indenter U

is replaced by a sharp pointed indenter and after a 25 gm load is applLed on I

the selected area of the sample, the micrometer stage is slowly moved to make II

a long scratch on the surface of the sample. The load is then removed and

the scratch width is measured over the different phases exposed on the surface. ½Initial experiments on polished surfaces exhibited a difference in scratch

hardness between matrix and tungsten grains. Further experiments showed that

this anomaly was primarily due to the unevenly polished surface. When the ii

samples were electropolished and tested for scratch width, the latter was i!uniform throughout the two phases. This method is not too satisfactory for

measuring hardness variations in two-phase alloys for another reason. TheIi

"Delivered by Armour Research Foundation of Illinois Institute of Technologypursuant to Contract No. DA .ll-022-505-0RD-3092° Government's use controlledby the provisions of Articles 26 and 27 of Title II (General Provisions)."

ARMOU RLSEAPCN FOUN DATION OP ILLINOIS I NTITUTE OP TECHNOLOGY

- 45 - ARF 2182-12Interim Technical Report

S50)0- Tungsten grains

. 1300 --

q44

200[

100 -

• e00 et unstengis

0 o

I f,

C)i

Fig. 21 Variation in mi~crohardness of matrix and tungsten grains

o 0,

as a function of to,,al tungston content in cemented W-Ni.-Fealloys. Mearsurements made by us•ing a 21ý gin load.

A R MOU PR C S VA RC H F U ND A 7IO0N Or I L L I N01 ,I N S TI T JT r Or 1EC 1 OtOGY :

4 6 -ARIP 2182-12 f

Tint.•17-ii Tv"ehn1ieal Report [

a " . .tV.,

scratch width is too large for the cross section of the matrix. The scratch

widths for 25 g7i load were about 10 microns in both phases whereas the matrix

film thickness was much smaller.

VIo DTSCUSSTION OF RESULTS AND APPLICATION

The yield strength-grain size relationship in pure tungsten and

matrix alloy has provided valuable data which can be correlated in two

different ways. Figure 22 consolidates the results of Figs. 13 and l1.

When the results for pure tungsten and matrix alloy are extrapolated to

the micron size range, it is found that, within a broad scatter band, the

flow stresses of the two components match reasonably well. A more accept- 4able way of plotting these results might be to correlate the "effective" ',

grain size of the matrix and the yield strength. In this case (Fig. 23),

the yield strength of the cemented alloys is plotted as a function of matrix A

film thickness (termed "effective" matrix grain size) and yield strength of

the single phase matrix alloy as a function of recrystallized grain size. JFrom many previous investigations, this extrapolation is justified; therefore,

such a "size" effect becomes all-important in determining the mechanical

properties of two-phase "cemented" alloys. At elevated temipcratures, the

alloys containing more than 90 per cent tungsten exhibit the mechanical

properties of pure tungsten in that their yield strength is essentially

constant from 4O0O to 1000*F and their deflection increases. This behavior

is characteristic of pure tungsten in the brittle-ductile transition tempera-

ture region. These alloys also e.,hibit poorer room-temperature ductility as

tungsten content increases.

The moagcr data on work-hqrdeninw coefficients (Table II) also

indicate that the matrix alloy and tungster: work harden to about the same

degree for equal strains, and this would certainly enhance the ductility

of these alloys. The plastic region of the stress-strain curves for the

matri- Tloy and for tungsten over a large range of grain sizes exhibits

slopes ,nich are almost identical for equivalent temperatures.

"Delivered by Armour Research Foundation of Illinois Institute of Technologypursuant to Contract No. DA-II-022-50O-ORD-3092. Government's use controlledby the provisions of Articles 26 and 27 of Title II (General Provisions)."

.4

AP P1OUUP PF A PACH 0 -OUN DATION O ILLINOIb INSTITUT E Of T'C!4-NOLOGY

- h7 - ARF 2182-12Interim Technical Report,

S... . . . ,• . . • . . . . . . . . .. . . . . . . " _ 2 _• •. . . _ . . . . _•. . . .. ... . . . : : , . . ,• . . .. . . . , . ',. -

L ~225

175P

125 7T-!-.H 1

U) - -,f

162

Gri daetr -1? - /2

s-I 7tots Fou -poni -m~rp ()f 13n

Af~~li(-)UP~~ -LL PC1 -A-D~ ( , (f 1 1 1( I, p" 1 1i )s T ( 1 ~ [

Aid*0po,

200 -A

In irigie phase, recrystallizedmatrix alloy ?UW-76(711-3f-e)

1-75

C',0.

Roo

50OF

100O

ft >2.5

D"Y ~ ~ ~ - 0Ifo~jl"v,ý, j fi

a d ,,rn

-ý 2 ýI 9ýoft o foclvsz( ~,jo,~y cm o, ternjdO U 125H- 3 r)- jjliat 7 ]jyi

f()a

P r f- t I AP tncementecA lThI y II t

Tuhnf-01 100evrt

o-m" "at

Microhardness measurements on matrix and grains afford "screening"

"techniques for developing new alloys. However, for these alloys, precise

measurement of microhardness values of matrix becomes a problem because of

.° the very small thickness of the matrix film.

Some conclusion can be drawn from the metallographic observations

in relation to the mechanical behavior of these alloys. In the binary alloys

of the W-Ni system, in which the flow stress of the Ni--28W matrix alloy is

higher than that of the pure tungsten, the deformation in the two-phase alloy

above the elastic limit is predominantly restricted to the tungsten grains.

Due perhaps to the higher flow stresses of the matrix film, cracks would be

nucleated in the deformed tungsten grains at the grain-matrix interface by

dislocation pile-up, and a predominantly transgranular fracture through the

t " tungsten grain results. The same argument is valid for the alloys in the

W-Cu system except in reverse because of the low flow stresses of the copper

matrix relative to the tungsten grains.

In the W--Ni-Fe cemented alloy system, however, the flow stresses of

the matrix and tungsten grains are quite similar. The deformation must start

in the weaker of the two phases; but since the flow stresses of the two phases

are similar, there would be no major barrier to propagation of slip from one

phase to another, This means that there would be no rajor source of crack

nucleation and the two phases would deform simultaneously. As discussed in

the previous section, attempts to detect any microhsrdness gradations in the

deformed tungsten grains were unsuccessful,

The electron microprobe analysis helped in establishing the chem.

istry of the matri-x alloy. It also helped in establishing the fact that there

is no composition gradient in the matrix nor any measurable solubility of

nickel or iron in tungsten.

The residual stress measurements on the matrix of the cemented alloys

indicate that there i2 apparently no dJ rect correlation between residual stress

and the mechanical properties of these ;lloy•s,. That the as-sintered alloys -ihow

ive iavered by Armour Rn,-ariPuund-Lun uf .......•.1i..o U,. .T .,.

pursuant to Contract No, -overnment's u:.c controlledby the provi sions of Articl.es 26 andt 27 of Title 1.1 (Gieneral Provas5ions)o'i

- Ari POU L•P M AP [ CII FOUND, ATI() U 1 I II NQI I IJSTITIJTI OF TI¢cIINOLOGY *

t ;. 50 .- AT' 21? 2-22V Interim Technical hteport

7I

a much higher value of residual stress than the polished samples may be connect-

ed with the observation that in all such liquid-phase sintered alloys there is a

very slight "flow-.out" of the matrix to the surface. The true residual stresses

in the matrix, as indicated by the measurements on the polished samples, are

rather low in magnitude. The measured stresses on tungsten grains are very

low; this is because the grains are rather large and the X-rays can only detect

the average changes in lattice parameters. i

It must be noted that in the method of measuring the residual stresses,

the change in 20 value is measured over a wide area of the sample covering the I

matrix in the interstices as well as in the films. As the residual stresses

reproduced in Fig. 17 were calculated from the average value of 20, it was felt

that similar calculations from the limiting values of 20 obtained for each

sample may also establish the stresses more representative of the matrix film.

These results are plotted in Fig. 24. Curve A is the same as that plotted in

Fig. 17. Curves B and C represent the stresses calculated from the maximum

and minimum values of 20 from the 20 vs intensity scan. It appears that the

stresses calculated from the mean 20 values ;re not appreciably different from

those calculated from the limiting 20 values. This, of course, does not

necessarily mean that the matrix film stresses are of the same order of mag-

nitude as the interstitial matrix stresses. It only points out the difficul- m

ties in measurem(.nt and interpretation of the results.

The reasons for the reversal of the sign in stresses at certain

volume fractions of the matrix are not clear. Theoretically, if one con-,

siderc a spherical inclusion in an infinite matrix. it can be shown that

the radial and tangential stresses in the matrix are of opposite signs and

decrease according to the cube power of the distance from the inclusion,

However, their analysis will apply only to relatively dilute alloys in which

the overlapping of the stress fields of neighboring particles is negligible.

The interference of stress fields of an array of inclusions has not been Ianalysed except by inference. Sternberg and Sadowsky(g) showed that when

"Delivered by Armour Research Foundation of Illinois Institute of Technology

pursuant to Contract No. DA-II-O22-505-O¶D-3092. Governmentts use controlledby the provisions of Articles 26 and 2'( of Title i(ii.. . Prov n-o,3),"

ARMOUR QCSEARC OUNDATION OF ILLINOIS INSTITUI0 O1- TECHNOLOGY

51- ARF 2162-12Interim Technical Report -

F ;>

!7

f

• •, I o 8 "

0 0 0 IKM

k I•3 -W

a) 0

Q)) U

Cd csj Cd c'i J

Q0 400Irdw~~iCHO C.U2 t>0)

to 0) 0) r

00 o~00(

0 1

•e~eZ~0. I~nTe ~lu-

-- 2 1 -

0-I T

SGSS.JW TVnPTS9J t~d~ou1:ad

ARMOUR P E5GA. PC H FOU ND A TIO N 0F I LL INOI5 N S TI TU TE O F TECHNOLOGY

- 52 -AHF 2182-12Interin Technical Report

the distance between two adjacent cavities (which may be regarded as

inclusions) is less than two diameters, the computed stresses deviate

drastically from the limiting values. Thus a model based upon spherical

inclusions begins to break down as the degree of dispersion of the in-SGurlan(7)

clusions decreases. urland has concluded by inference that as the

degree of dispersion of particles decreases, the stresses in the components

become complex and possibly a component of stress of opposite sign is

introduced in such a system, thereby causing a sign reversal. This

behavior in dispersed alloys makes it almost impossible to evaluate

the relationship between residual microstresses and the gross mechanical

behavior of two-phase alloys.

The high values of residual microstresses (see Fig. 17) in the

"as-sintered samples were probably due to the slight "flow out" of the

matrix during sintering. This excess concentration of matrix on surface

of sintered samples is observed in all samples to a greater or lesser

degree. Why this surface in the "as-sintered" condition is so highly

stressed is not clear. However, the polished and etched samples show a

much lower magnitude of residual thermal stresses--of the order of 50,000psi in samples containing less than 10 volume per cent matrix--and one would

think that these stresses represent truly the internal stress condition in

the samples. The mechanical properties of "as-sintered" and polished and

etched samples were identical.

The question of "constraint" of the matrix due to triaxial stresses

during deformation is still unanswered, and to what degree this phenomenon

affects the mechanical behavior of these alloys is not clear. At the moment

it seems difficult--if not impossible--to design meaningful experiments which

could help in unambiguously elucidating this point.

The concepts developed in this investigation were used for a cursory

look at iron-copper alloys with a view to developing high-strength, ductile

Fe-Cu alloys. These alloys have a typical cemented alloy structure, but they

"Delivered by Armour Research Foundation of Illinois Institute of Technologypursuant to Contract No. DA-11-G22-c0-ORD-3092. Government's use controlledby the provisions of Articles 26 and 27 of Title II (General Provisions)."

APMOUR PESEARCH OUN DATION OF I LLI NOIS I N STITUT OF TECHNOLOGY

- ARF- A 2182-12Interim Technical Report

.......................................... ................ .[" ' ... - ?"

W

exhibit no ductility, and their strengths in bending are rather low--about•.- • - 8 0 0 0 0 p s i .

The yield strength-grain size relation for pure iron and copper

are reproduced schematically in Fig. 25. The curves represent a simple

relation:

y a y -where ry is yield stress

o is stress necessary to initiate0 motion of unlocked locations,

"k is magnitude of resistance to•" Y motion of dislocations and

d is grain diameter.

Both the values of and k arc lower for copper than for0y -

iron. Hence. the conclusion must be that there is no chance for ductility

in this system. However, there are simple ways of increasing both the0

and k of copper. One method is by solid solution hardening and the other

is age hardening. For solid solution hardening of copper, additions ofaluminum, palladium, and manganese were made to Fe-Cu alloys; and for age

hardening, small beryllium additions were made. The results of the pre-

liminary investigation of the effects of such additions on the mechanical

properties of Fe-Cu alloys are summarized in Table V. it can be seen that

most of the additions substantially improved the properties of the Fe-Cu

alloys. It was also noted that the manganese additions substantially refined

the grain size of the iron particles.

The work on Fe-Cu alloys is by no means conclusive as this phase

of the investigation was only cursory. Although comparable properties were

"Delivered by Armour Research Foundation of Illinois Institute of Technologypursuant to Contract No,, DA-ll-O22-5O5-ORD-3092. Government's use controlledby the provisions of Articiess 26 and 2^7 of Title; II. (Goneral ProvisinrYs.".

ARMOUR •[SFAPCR FOUNDATION 0E 1LLINOIS I N TIUT OT U T C TECHNOL.OGY

-" 5h - APF 2182-12Interim Technical Report

Tro

iI

Avrg gri dimtr d'122m-/

& il-

r I

A IS~J

I I I !,' Ave~rage grain diameter, d"I/2• m-/2

SFig. 2• Schematic representation of yield stress-.grain sizeI

• ~~relationship i.n Jrcnl' and copper.

• ARMOUR RI S ARIC.W FOUNOATION OF- ILLINOIS IN511tJTF O1 "i £.4 NOLOGY .

-55 - ARF 2182-12Interim Technicnl Rleport

TABLE V

Properties of Some Fe-Cu Base Aloys

Yield Strength, Deflection,Composition Sinte ring Treatment 1000 psi inches

8OFe-2OCu I150"C - 1 hr - H2 60 - 70 None

9OFe-lOCu I150OC - I hr - H2 85 None

9OFe-I0(9Cu-IAl) 1150*C - 1 hr - H 130 O.062

90Fe-lO(7Cu-3Pd) 1150*C - 1 hr - H2 175 O.14

2S8OFe-20(7Cu-3Mn) 950°C - 1 hr - H2 70 - 82 None ••

8OFe-20(7Cu-3Mn) I140*C - 1 hr - H2 180 - 1911 0.1

8OFe-20(7Cu-3Mn) 1200*0 - I hr - H2 110 None

80Fe-20(Cu-2Be) Infiltrated 130O0C 160 - 175 0.1Aged 300*C - 1 hr

i

"Delivered by Armour Research Foundation of Illinois Institute of Technologypursuant to GontracL No. r, n ovrnment's u controlledby the provisions ef Articlps 2?( and 27 of Title IT (General Provisions)."

ARMOUR RF5FARCI4 rOUH DATION OF IILLINO1S INSTITUTE OF T IF CCWNOLOGY

- 56-- ARF 2182-12Int-: rim TecohiiiiiL RPport1

not realized in the few Fe-Cu alloys investigated, the trend seemed reason-

ably clear. The matrix-strengthening additions were selected with care so

that their diffusion in the particulate phase was very slow at sintering

temperatures. If the alloy additions dissolved to any appreciable extent

in the iron particles, this dissolution or diffusion was not reflected in

* the microhardness measurements.

VII. CONCLUSIONS

1- I. The high strength and ductility of W-Ni-Fe alloys with typical IIcemented microstructures is primarily due to the similarity in

the magnitudes of the flow stresses of the individual components.

The particulate phase is almost pure tungsten and the matrix is ia nickel base alloy with substantial amount of tungsten dissolved

in it during the sintering process. The characteristic flow Istresses of the grains and the matrix can be extrapolated from"the yield strength-grain size relationship.

2. If the flow stresses of the individual components in cemented

alloys are too dissimilar, as in the case of W-Ni or W-Cu

binary alloys (with identical cemented structures as in the

W-Ni-Fe system), then not only are these alloys brittle, but

Wthey also exhibit lower strength values. Thus, in such alloys,

the inhomogeneous deformation (or preferential yielding) of

the low flow stress component, will build up a stress discon-

tinuities, nucleate cracks, and on further application of

stress propagate these cracks to failure at relatively lcw

stress levels. In the W-Ni-Fe system, on the other hand,

the response to application of the stress in both components 4

is uniform so that there exists no "barrier" to slip propa-

"gation in either of the two phases.

"Delivered by Armour Research Foundation of Illinois Institute of Technologypursuant to Contract No. DA-II-022-505-OTD-3092. Government's use controlledby tue pi'ovibivas of A..Liles 26 and 27 of Title II "Geeal Pv Wi5)

ARMOUR R E S FAARCH FOUNDATION OF ILLINOI5 INST"ITUTE OF TECI-IOLOGY

57 - ARF 2182-12

Interim Technical Report

ii

3. The work hardening coefficients of the two components in

the W-Ni-Fe alloys are fairly similar and this condition

is presumed to be Important in establishing the degree of

ductility. This conclusion is similar to (2) above, but

is based upon measurements of work hardening coefficients.

4. The chemistry of the matrix alloy in cemented W-Ni..Fe alloys

was established by electron probe measurements. The solu-.

bility of tungsten in nickel-iron matrix (793 ratio) was

found to be about 24 weight per cent at room temperature.

"Solubility of either nickel or iron in tungsten grains was

"negligible. Scanning of the phases showed that each phase

; was of uniform composition.

-5. The residual thenral stresses as measured on the components

-- in this system are rather low and, on this basis, are pre-

sumed to play a secondary role in determining the mechanical

properties of the W-Ni-Fe alloys. As shown in Fig. 24, even

if the maximum value of 29 is used for calculating the d

value from which the stresses are calculated, the stresses

- are calculated, the stresses are rather low and reverse the

sign at about 15-20 volume per cent matrix. This does not,

necessarily mean that matrix stresses are as low as calcu-•

lated. It demonstrates the difficulty in measuring the

matrix stresses in alloys containing such low percentages

of the matrix.

6. Metallographic study of fracture behavior supports the

developed theory that the two components possess flow

.- stresses of the equivalent flow stresses, as discussed

* in conclusions (1) and (2) above.

'!elivered by Armour Research Foundation of illinois Institute of Technologypursuant to Contract No. DA-II-022--505-ORD-3092. Government's use controlledby the provisions of Articles 26 and 27 of Title II (General Provisions)J'

ARMOUR RESEARCI FOUNDATION OF- ILLIN INSTITUIT OL TECHNOLOGY

58 AEF 2182.-2Interim Technical Report,

PA

7. The concepts developed here were successfully applied to a

cursory investigation of the alloys in the Fe-Cu system---

resulting in high-strength al].oys with some ductility.

VIII. lOGBOOKS AND CONTRIBUTING PERSONNEL

The data are all recorded in Logbooks Nos. C-9399, C-1O89, and

C-1180.

N. M. Parikh, J. R. Walenga (Technician), and E. Heckenbach

(Technician) contributed to this program.

Respectfully submitted,

ARMOUR RESEARCH FOUNDATION OFILLINOIS INSTITUTE OF TECHNOLOGY

N. M. Parikh, Senior Metallurgist

W. Rostoker, a-t DirectorMetals Research

NMP:cs

"Delivered by Armour Research Foundation of Illinois Institute of Technologypursuant to Contract No. DA--_•o22o_ uM_. .ont.r1 1 edby the provisions of Articles 26 and ?7 of Title IT (General Provisions)."

ARMOUR A IS CAA CI14 FOU N DA TIO N 0OF I L LI NOIS INS T IT U TE O F TrýC HN C LOC1GY

- 59- AIiF 2182-12Interim Technica] fleport,

REFERENCES

l. Green, E. C., U. J. Jones, and W. R. Pitkin,SyO m on Powder NetallurgyLondonIron" Steel Inst., pp. 1-h, (Dec. 1954)

2. Taylor, G. I., "Plastic Strain in Metals"

J. Inst. Metals, 62, 307 (1938).

3. Soldau, P., Z. Anorg. Allgem. Chem., 194, 1 (1930).

h. Honeycomb, R. W. K., and W. Boas,Australian J. Sci. 1es., 1, 70 (1948).

5. Clarebough, L. M., Australian J. Sci. 11es.,3, 72 (195o).

6. Gensamer, M., "Strength and Ductility"Trans. ASM, 3h, 30 (19h6).

7. Gurland, J., Trans. ASM, L0, 1063 (1958).8. Laszlo, F., Jour. Iron and Steel Inst.,

14719 173, (1946).

9. Sternberg, F., and Sadowsky, Jour. Apple Mech.,

4,, 57 190S!i

"Delivered by Armour Research Youndation of llinorioL .. i..... 'L, .....o E, 3

pursuant to Contract No. DA-11-022-505-ORD-3092. Government's use ccntrolledby provisions of Articles 26 and 27 of Title II (General Provisions).'

ARMOUR RrS.ARCN FOUNDATION 01r ILLINOIS INSTITUT0 OF TECHNOLOGY

60- ARF 2182-12 I

Interim Technical ReportS... .. "-" . .. . " .. ..... .. .. '--"'• • • --••'•- -.... . -• • " •-"r•'z•' q #.-.rle ,- 0,

4 - - -a- -ir -- ill I ---No.

13

~~ * ~ '4) 0 4) gr3

~~to-HI

PC 01 H' I

*HC4 E;9~ 4 rV 1

0 0 15

El (n4 9 01" A

z 0 *r"4'T5 1 4 -

04 U) W

Cy. m 0 wIn 4 -19 H I w j) 4) ) 4 .t

441- 4 -1~ gC - 0 CkW ,o

A *0 (143. pu Vau 424 V)1L~ 0.

"4) +3 w ) -,c

w V\ n4 I0wwP'n1 rl

s-.ca~ mtol 0C''

Or WW Cuf 400) 4> ) 44H*~~~ (44 tO i I .), "' H 4'qj 'a 43a

.c 0,90 o npI0 A- 0u 0o CO co .--q w O0NC

H v -Z U7- ,- ., - - - - -- - - 1 Ž fi -

9 t~..u~m.

re L SL LGS St St A.

SA.D. Accession No.Watertown Arsenal laboratories, Watertown 72, #4 =Massachusetts. DEVEIOPN AND APPLICATIONOF A THEORY OF PLASTIC DEFORMATION OF CEMENT-ED ALWOS, by N. M. Parikh. Report No. VALS372/32. March 1961. pp. 60 - illus. 25tables 5, DA Project ll-022-505-ORD-3092,ORD Project No. TB4-002.

Unclassified report

In the cemented W-Ni-Fe alloys, the sourcesof strength and ductility were found to bedue to the close matching of the yieldstresses of the matrix and tungsten parti-

i. This was proved by measuring yield't•iength as a function of grain size of the

UL'CLASSJIW ID

UNCLASSIFIEDtrix and tungsten individually. The chem-

istry of the matrix was established byelectron probe analysis. The residualstresses in these cemented alloys were -measured by the X-ray method. Applicationsof these principles resulted in a few newFe-Cu base alloys of high strength and I

Iiductility.

4I I

UNC--ASSIFI

7i