80 A NICKEL-ALUMINUM-MOLYBDENUM CREEP RESISTANT ALLOYI Bv H. V. Krr.rspy2 aNtr M. T. SrBwenr3 Abstract llir pup."l describes a preliminary study of alloys of nickel and aluminum modified with molybdenum. The purpose-of this work is to develop an alloy for use under conditions of stress-at temperatures of 815.C. (1500.F.) and over. The room temperature mechanical properties of alloys of nickel and alumilum, .and_the influence of molybdenum on these properties, have been investigated.. Certain combinations 'of nickel, aluminum, ind molybdenum have been shown to possess tensile strengths well over 100,600 lb. per sq. in. at room temperature, and it has been demonstrated that certain charaCteristic microstructures,-dependent upon the ratio of nickel to aluminum, are essential for therealizationofthesehigh sirengths. Creep-rupture tests at 815.C. (1500.F.) have been carried out on typical nickel-aluminum-molybdenum alloys. Th; results have shown that cert,in of these allovs are suoerior in manv resDecrs ro exist_ing high tempsrattrrc alloys, when tesied trnder crcep-rupture conditions at 815' C. (1500' F.). The same characteristics of miciostiucture that are essential for high room temperature strengths were also found to be necessary to obtain good creep-ruptuie characteristics at 815" C. (1500. F.). Introduction So much has been written on the subject of alloys for use in gas turbine blades that the more important requirements for such materials are now widely recognized. They must have chemical stability to ensure resistance to corro- sion by hot combustion gases. A second requirement is stability of micro- structure at anticipated operating temperatures, to minimize the deterioration of mechanical properties during long exposure to elevated temperatures. Also, a high strength-weight ratio is desirable in alloys under consideration for gas turbine blades, since the stress causing creep in these parts is generated by centrifugal force. The high melting points and relatively low densities of the nickel-rich nickel-aluminum alloys, specifically of the intermetallic compound NiAl, suggested that an investigation of the creep properties of allo-vs based on the nickel-aluminum system might be worth while. The nickel-aluminum phase diagram (1) is presented in Figs. 1 and 2. All the alloys to be dealt r,vith in this paper rvere melted by high frequencv induction in magnesia crucibles. Test bars were centrifugally cast into molds produced by the "lost $'ax" precision casting technique. All room temperature tensile tests were made on the Hounsfield tensometer, using the test-bar design shown in Fig. 3. All creep and stress-rupture tests I lfanuscr,ipt receiaed, September 22, 1918. - Published by pernission oJ the Director, Mines, Forestrs ond Scienl'if.c Seraices Bronch, Department oJ Mines-and Rcsnurie.r, Ottawa, Canada. Alelallurgical Engincer; Head, lligh Tentpernlurt AIetals Laboratory, physiral ,lletollurgy Resrorclr, Laboralories, Dit,ision of llIinrrnl Dre5i;ag nnd Aletallurgy, Buriou "j ltlinos, OIlaw"i. :t Junior Research Offter, Associate Comnittee on High Temperalure Metals, National Research Council of Canodi, Ottowa. Canoda.. Can. J. Res. Downloaded from www.nrcresearchpress.com by DALHOUSIE UNIVER on 10/29/12 For personal use only.
Transcript
80
A NICKEL-ALUMINUM-MOLYBDENUM CREEP RESISTANTALLOYI
Bv H. V. Krr.rspy2 aNtr M. T. SrBwenr3
Abstract
llir pup."l describes a preliminary study of alloys of nickel and aluminummodified with molybdenum. The purpose-of this work is to develop an alloyfor use under conditions of stress-at temperatures of 815.C. (1500.F.) andover. The room temperature mechanical properties of alloys of nickel andalumilum, .and_the influence of molybdenum on these properties, have beeninvestigated.. Certain combinations
'of nickel, aluminum, ind molybdenum
have been shown to possess tensile strengths well over 100,600 lb. per sq. in. atroom temperature, and it has been demonstrated that certain charaCteristicmicrostructures,-dependent upon the ratio of nickel to aluminum, are essential fortherealizationofthesehigh sirengths. Creep-rupture tests at 815.C. (1500.F.)have been carried out on typical nickel-aluminum-molybdenum alloys. Th;results have shown that cert,in of these allovs are suoerior in manv resDecrs roexist_ing high tempsrattrrc alloys, when tesied trnder crcep-rupture conditionsat 815' C. (1500' F.). The same characteristics of miciostiucture that areessential for high room temperature strengths were also found to be necessaryto obtain good creep-ruptuie characteristics at 815" C. (1500. F.).
IntroductionSo much has been written on the subject of alloys for use in gas turbine
blades that the more important requirements for such materials are now widelyrecognized. They must have chemical stability to ensure resistance to corro-sion by hot combustion gases. A second requirement is stability of micro-structure at anticipated operating temperatures, to minimize the deteriorationof mechanical properties during long exposure to elevated temperatures.Also, a high strength-weight ratio is desirable in alloys under considerationfor gas turbine blades, since the stress causing creep in these parts is generatedby centrifugal force.
The high melting points and relatively low densities of the nickel-richnickel-aluminum alloys, specifically of the intermetallic compound NiAl,suggested that an investigation of the creep properties of allo-vs based on thenickel-aluminum system might be worth while. The nickel-aluminum phasediagram (1) is presented in Figs. 1 and 2.
All the alloys to be dealt r,vith in this paper rvere melted by high frequencvinduction in magnesia crucibles. Test bars were centrifugally cast into moldsproduced by the "lost $'ax" precision casting technique.
All room temperature tensile tests were made on the Hounsfield tensometer,using the test-bar design shown in Fig. 3. All creep and stress-rupture tests
I lfanuscr,ipt receiaed, September 22, 1918.
- Published by pernission oJ the Director, Mines, Forestrs ond Scienl'if.c Seraices Bronch,
Resrorclr, Laboralories, Dit,ision of llIinrrnl Dre5i;ag nnd Aletallurgy, Buriou "j ltlinos, OIlaw"i.:t Junior Research Offter, Associate Comnittee on High Temperalure Metals, NationalResearch Council of Canodi, Ottowa. Canoda..
Can
. J. R
es. D
ownl
oade
d fr
om w
ww
.nrc
rese
arch
pres
s.co
m b
y D
AL
HO
USI
E U
NIV
ER
on
10/2
9/12
For
pers
onal
use
onl
y.
KirSEv AND STEIITART: A Ni-AI Mo CREEP RESISTANT ALLOY 61
were obtained with the test bar clesign shou'n in Fig. 4. Note that one end
of this bar is hollon,, to protnote directional solidification.The metal temperatures of the melts cast into the large test bars (see
Fig. 4) were measured with an immersion platinum-rhodium thermocoupleused in conjunction with a high speed potentiometer-type temperaturerecorder.
The metals used in this work are listed in Table I, together with the analysesobtained from the suppliers.
phase diagram oter cttmplete range. (AJter Alexander and
Can
. J. R
es. D
ownl
oade
d fr
om w
ww
.nrc
rese
arch
pres
s.co
m b
y D
AL
HO
USI
E U
NIV
ER
on
10/2
9/12
For
pers
onal
use
onl
y.
82 CANADIAN JOLIRNAI. OF RESEARCH. VOL.27. SEC. F.
r500
1400
75 80 a5 90 95 IOO
NtcKeu, Wercxt PeeCer.rrphase diagram for the range oJ 75 lo 100/s nickel. (After
Frc. 3. Tensile bar Jor llounsf.eld, tensometer.
Frc. 4. Precis.ion cost / in. diameter creep test bar.
System of Alloy NomenclatureWith the exccptiorl of thc prelirnirrarl, groLlp of f<tur birrarl. alloys of nickel
and alurninum, all the allo-vs lo tre rliscussed wcrc nraclc up of nickcl, aluminum,and n'iolybdcnum. In ordcr to facilitatc the discussion of these alloys, thefollowing nomenclature \r/-as devisecl.
Jf thc Ni: Al ratio ancl thr: arriount of r-nol-vbclcnurn presL.nt arer stal-ccl ,
then the alloy is fully clescribecl. F or exarnple, let t1.re Ni : Al ral.io be 10 . 3 : 1
and the amount of mol;-bdenunr present lte 14.7o/r. 'Ihen the allo-v is identi-fied by the expression 1031't147. 'lhe letter "l'I" is inclucle,:l to indicate thatthe third element prcscnt is moh,bclenLtrr. If ther Ni : Al ratio is 7. 6 : 1 ancl themoly'bdenum content is8.2o7e, ttren the cxpression i6M82 identifies the allo1'.
UouldlkdUJo
#
Frc. 2. Nickel-alu.mintLntAlexond.er and Voughon (1).)
l<z+
z
Can
. J. R
es. D
ownl
oade
d fr
om w
ww
.nrc
rese
arch
pres
s.co
m b
y D
AL
HO
USI
E U
NIV
ER
on
10/2
9/12
For
pers
onal
use
onl
y.
KINSEY AND STEWART: A Ni-AI.MO CREEP RESISTANT ALLOY
Nickel-Aluminum Alloys
Table II lists the phase zone boundar-v limits in the nickel-richthe nickel-aluminum phase diagram (see Figs. 1 and 2). Theselimits are stated in terms of the ratio of nickel to aluminum.
Four trinary alloys of nickel and aluminum, characteristic of the first fourphase zones listed in Table II, were chosen as a starting point for this work.These allovs are listed in Table lII.
TABLE IIIBrNe,nv NTcKEL-ALUMTNUM ALLoYS
Anticipatedphases
NiAlNiAI + NisAINi:A1Ni:Al * alphasolid solution
Hounsfield tensometer test bars were cast and the room temperature mech-
anical properties given in Table IV were determined.
TABLE IV
Roou mlrpBnATURE MECHANIcAL PRoPERTTES oF NICKEL-ALUMTNUM ALLoYS
Tensile strength, p.s.i. Averageelongation,
%
Vickershardness value(30 kgm. load)
Alloy No.MaximumAverage
Nil28,50060,0006+,667
Nil30 ,00062,00077 ,000
Nil
18 3
389404265346
Chemical analysis, /6
Aluminum
Can
. J. R
es. D
ownl
oade
d fr
om w
ww
.nrc
rese
arch
pres
s.co
m b
y D
AL
HO
USI
E U
NIV
ER
on
10/2
9/12
For
pers
onal
use
onl
y.
84 ?ANADTAN J)IRNAL oF RESDARCE. voL. 27, sEc. F.
'fhe properties listed in Table IV are shown plotted against the Ni : Alratio in Fig. 5. It will be noted that Allo-vs 3 and 4 above, rvhich on the basisof their Ni : Al ratio .w'ould not be expcctcd to show any of the NiAl con-stituent, possess the most promising room temperature mechanical properties.
Frc. 5. Mechanical, properties as. Ni : Al ratio.
M etallo gr aphy of lVi.ckel-A luminunt, Attoy sFigs. 6 to 11, inclusive, are photomicrographs sholving the microstrr-rcture
of Alloys 1,2,3, and 4, respectively. Alloy 1,lvith a Ni : Al ratio of 2.7 : l,has only one phase, as would be expected. Alloys 2 and 3 have two phases.Alloy 2 has a Ni : Al ratio of 4.9: 1, r,vhich indicates the presence of bothNiAl and NiBAl. Allo-v 3 has a Ni: Al ratio of 5.6: 1. Under equilibriumconditions this alloy should consist structurally of a single phase, NiBAl.However, since all the rnicrostructures uncler consideration are of "as cast"alloys, thc prcsence of tu.o phases in A11o1' 3 is not surprising. These trvophascs would be cxpectccl to bc NiAl ancl NirAl.
..Ft!, 6, -^Alloy.\'o. I, as rast. Elertrolytic polish,At, 26.927a.
. -et_., Z ^Alloy No. 2, as cast. nfechan,ical polish,At, 16.81Ta.
Ftc. 33. Alloy 90XI158, oJler 680 hr. creep-rupl.ure lesling al 815' C. (1500" F.), Etchonl:alcoholio chrome regia, )(1500,
Can
. J. R
es. D
ownl
oade
d fr
om w
ww
.nrc
rese
arch
pres
s.co
m b
y D
AL
HO
USI
E U
NIV
ER
on
10/2
9/12
For
pers
onal
use
onl
y.
KINSEY AND S?EWART: A Ni-At-Mo CREEP RESISTANT ALLOY 85
A comparison of the microstructures of Allol's 2 and 3, as shor'vn in Figs. 7
and 8, will enable us to cstablish the identity of these two phases. Thcre isa dark-etching phase and a light-etching phase present. Allo-v 2, which has alower Ni : Al ratio than Al1o1' 3, shouid be expccted to have more of the NiAlphase than Alloy 3. Since there is more of the dark-etching phase in Allol' 2
than in Alloy 3, it is assumed that this dark-etching phase in these structuresis NiAl and the light-etching phase is NiaAl.
The Ni : Al ratio of Alloy 4 indicates that it u''ill consist structurall1, oftwo phases, alpha and NisAl. Its microstructure, as 3hown in Figs. 10 and 11,
indicates the presence of trvo phases. Most of the structure is characteristicof the structure obtained u'hen a secondary phase precipitates from a prirnarysolid solution orving to a shift in solubility u'ith falling temperature. Fromthe position of this alloy in the nickel-aluminum s-vstem, it r'vould be expectedtof.reeze out largely as alpha solid solution and to precipitate NirAl on coolingbelow 1200'C. It is apparent, then, that the precipitated phase in this alloyis Ni:Al and that the matrix is alpha solicl solution.
Attention is drarvn to the clear rvhite areas visible in Fig. 10. '\ccordingto Alexander and Vaughan (1), these are areas of NigAl that, on freezing, were
the alpha-NiaAl eutectic. They havc shown that this eutectic cannot existbelo.,v 1370o C., ou,ing to the incrcase of solubility of nickel in NirAl that occurs
between 1385' C. and 1370o C., rvhich results in the disappearance of u.hat littlealpha has separated in the eutectic. The fact that some eutcctic formed inAlloy 4 may be explained by the fact that sufficient segregation occurredduring freezing to permit this formation.
Nickel- Aluminum-Molybdenum Alloys
The filct that molybdenum is a vital constituent of man-v high temperatureallorrs prompted a decision to investigate its influence on nickel aluminumallol's. Tire program adopted involved preliminary tests at room temper-ature, to be follol'ved b). crcep and creep rupture tests on such alloys as piovedcastable and displal'ed promise from consicleration of tlieir room temperatureproperties.
For a preliminary stud-v of room temperature propcrties, these molybdenum-moclified allo-vs rvcre clivided into four groups, each group having a constantNi : Al ratio, so that, in effect, an attcmpt ll-as maclc to stucly the influcnceof various molrrbdcnum atlditions to the allo1-s listed in fable III. Alloys 1,
3, and 4 of Table III rvere the starting point for three of these groups. 'fhe
fourth group \vas basecl on a Ni : Al ratio of 19 :7, ivhich represents a single-phase alpha solid solution a11o1'.
These alloys are listed in Tablc V. Fig. 12 shorvs the position of these
alloys on the ternar-v nickel-aluminum-mo11'bdcnum triangle" This chartalso contains the nickel aluminum phase diagram (1) and the nichcl-molyb-denum phase diagram (2).
Can
. J. R
es. D
ownl
oade
d fr
om w
ww
.nrc
rese
arch
pres
s.co
m b
y D
AL
HO
USI
E U
NIV
ER
on
10/2
9/12
For
pers
onal
use
onl
y.
CANADIAN JOURNAL OF RESEARCH. VOL, 27, SEC. F.
TABLE V
Cnnurcer, ANALvsES oF NIcKEL-ALUMINUM-MoLyBDENUM ALLoYS oBTATNIID
Chemical analysis, /eAlloy sought Alloy obtained
Aluminum Nlolybdenum
30M4430M16230M25030M35030M45030M550
57M5057M10057M15057M250J / IVIJJU57M45057M550
90M5090N,I10090N{ 1 5090M25090M350
190M50190N{ 100190M150190M250
30M4730M14935M20230M24729M38628NI51 1
54NI3957M10655M15657M13159M32057M38456N,I.143
s0M5386M104
130M17794M24889M357
177M46165M105172M16r185M255
a1 nt63 .8061 .9056.6545. 5535.95
81 .0276.O171.3674.058.2JZ.+47 .3
23.732t.2417 .9018.7015.8312.95
15.0413.3613.0712.99.89.28.4
9.48o ?7
5.837 .206.5
4.955. 104.63.8
14.9120.224.6538.6251.10
3.9410.6015.5713. 1032.O38.4++.J
J.Zl10.4317.6724.8035 .7
85.2580.2,/o..)68.057.8
90. 584.479 .370.7
4.5510.5016.125 .5
Room Temperature Mechan'ical Properlies of Nickel-Alunainum-Molybd.enumAlloys
Three bars from each of the alloy heats listed in Table V were pulled in roomtemperature tensile tests. The results are listed in Table VI. Vickers hard-ness values are also recorded in this table.
The results of these tensile tests are summarized graphically in Fig. 13,
which shows the influence of molybdenum on the tensile strength and hardnessof alloys of several nickel aluminum ratios. Evidently mollbdenum ismost effective as a strengthener w-hen nickel and aluminum are in such aratio that a binary alloy of nickel and aluminum r'vould consist structurallyof NisAl f alpha solid solution, or alpha solid solution alone (see Fig. 1).
M etall o gr a 1> hy oJ N i, ck el- A lumin um- M ol,y b d enum A ll o y s
Figs. 14 to 21 illustrate the significant metallographic features of the alloyshaving the chemical and mcchanical properties listed in Tables V and VIrespectively.
These photomicrographs should be interpreted rvith reference.to the phase
zone boundary limits set out in Table II.
Can
. J. R
es. D
ownl
oade
d fr
om w
ww
.nrc
rese
arch
pres
s.co
m b
y D
AL
HO
USI
E U
NIV
ER
on
10/2
9/12
For
pers
onal
use
onl
y.
KINSE/ AND STEWART: A Ni-Al-Mo CREEP RESISTANT ALLOY
. Ftc. 21. , ,1llo2 1-85 '11255, us cost. frechonical polish; etchont, ferric chlori.de l>lus hydro-chLoric acid in alcohol. X200. Ni, 70.7/6: AI, S t7o; Mn, 2 j'.5To.
Can
. J. R
es. D
ownl
oade
d fr
om w
ww
.nrc
rese
arch
pres
s.co
m b
y D
AL
HO
USI
E U
NIV
ER
on
10/2
9/12
For
pers
onal
use
onl
y.
KIN'SEY AND STEWART: A Ni-At-Mo CREEP RESISTANT AI'LOY 89
When the Ni : Al ratio is 3 : 1, the alloy may be expected to be basicallyNiAl. The addition of molybdenum to such an alloy causes the formation
of a second phase. This is shown in Figs. 14 and 15, rvhich are the micro-
structures of Alloys 30M149 and 30N{247 respectively.
In as-cast binary nickel-aluminum alloys with a Ni : Al ratio of 5.7 : l,previous r,vork has shown that two phases, namely NiAl and NiBA], may be
present. These phases have been tentatrvely identified in Figs. 7 and 8.
Based on this identification, it may be assumed that the dark-etching phase
in the nickel-aluminum-molybdenurn allo)'s having a Ni : Al ratio of 5. 7 : 1
will be NiAl. The structurcs of t1'pical alloys of this group, Alloys 57M106
and 55M156, are shown in Figs. 16 and 17. r
The eutectic appearing in Alloys 57N{106 and 55M156 is characteristic of
the presence of molybdenum. 'fhis eutectic can be caused to disappear from
Alloy 57M106 by hcating to 1000o c. (1832'F.). It is therefore assumed thatthe presence of this eutectic in Alloy 57M106 in the as-cast condition is due tosegregation on freezing, and that undcr conditions approaching equilibriumit $'i11 not occur in this alloy.
In alloys of the 9 : 1 Ni : Al ratio group, molybdenum again causes the
occurrence of a characteristic phase. This is illustrated by Fig. 18, which is
the microstructure of Alloy 94\4248. (The NiAl phase is absent from thisgroup of alloys.) As the molybclenum contcnt increases, this phase becomes
more massive, as is shotvn by F'ig. 19, 'which is the microstructure of Alloy89]l 35 7.
Alloys having a Ni : Al ratio greater than 15 : 1 are structurally single-
phase alpha solid solution up to at least l6TaMo; Fig. 20 is characteristic of
this microstructure. At 257a N,Io a second phase appears; this may be seen
in Fig. 21.
An attempt has been made tentativel-v to establish the idcntity of the phase
that is associated with the presence of molybdenum' This identilication is
based on the rvork of Ellingcr (2).
It wouid appear that this phase, which occurs as a eutectic in Allo1'5 57M 106'
55M156, 94M248, and 89M357 (Figs. 16, 17,18, and 19), and in a primarvform in Alloys 30N'I149, 30M247 and 89NI357 (Figs' 14, 15, and 20), could be
the delta (NiNio) phase of the nickel-mo11'b6"tt,t- system.
On this basis, Alloys 30M149, 30M247, and 89M357 are all hypereutectic
tvith respect to molybdenum, since primary NiMo is present. The absence
of primary NiMo in Alloys 57M106, 55N,I156, and 94M248 r'vould indicate
that these alloys are hypoeutectic with respect to molybdenum'
Summary of Metallogra phic Feol ures
- 1. The addition of molybdenum to alloys 61 nickel and aluminum causes
the appearance of a characteristic phase tentatively identified as NiNIo. This
phase will occur in the primary form r,vhen the alloy is hypereutectic withrespect to molybdenum and it luill occur in the eutectic form when the alloy ishypoeutectic with respect to molybdenum.
Can
. J. R
es. D
ownl
oade
d fr
om w
ww
.nrc
rese
arch
pres
s.co
m b
y D
AL
HO
USI
E U
NIV
ER
on
10/2
9/12
For
pers
onal
use
onl
y.
90 cA\ADrAn JIDRNAL oF REIEARCTI. voL. 27. "sEC. F.
2. The limiting mol.vbdenum content at which the phase NiMo will occurvaries u'ith the nickel : aluminum ratio.
3. The NiAl phase pcrsists in thc ternary alloys of nickel, aluminum, andmoly[dsrum at nickel : aluminum ratios gencralry similar to those in thebinary nickel-aluminum alloys.
4. The presence of the NiAl phase has a marked weakening effect on thealloy and renders it less rcsponsive to strengthening by additions of molyb-denum.
Creep-rupture Properties of Some Nickel-Aluminum-Molybdenum Alloys at 815. C. (1500. F.)
The alloys selected for creep-rupture tests are risted in Tabre vII. Thischoice was dictated by the r,r'ish to inclucle a represcntative alloy of each ofthe groups sholvn in Table V. The molr.bdenum contents r,rrere chosen fromconsideration of the roon temperature properties presentccl in 'fable vLThe cast test bars u'ere of thc design show-n in Fig. 4.
TABLE VIIAr-r-oys sELDCTED FoR cRE-brr,-RUpruRE sTLTDIES
Allol'Chemical analysis, /6
Aluminum Molybdenun'r
30M25056NI 10090M15090M200
190M250
56/()/otz71
2510t.)2025
19t,)98+
The liquidus and solidus temperatures of thcse alloys werc first measuredas an aid in determining the proper casting temperature. Both the freezingcurve technique, employing an immersion platinum-rhodium thermocouple,and the metallographic n-rethod were employed. Table vIII lists the resultsof these measurements.
TABLtr VIIILlqurous AND SoLIDUS T.rJMpltl{ATURES
Chemical analysis, Freezing curvedata
Metallographicdata
Solidus,'c.
Not determined
1325 - 13+01285 - 1300
Not determined
AlloyNi NIo
35M24676M15195M138
106M217213M257
58.6/J.O79.O72.972.4
21.6IJ. 1
13.821 .725.7
16.616.98.36.9J-+
Can
. J. R
es. D
ownl
oade
d fr
om w
ww
.nrc
rese
arch
pres
s.co
m b
y D
AL
HO
USI
E U
NIV
ER
on
10/2
9/12
For
pers
onal
use
onl
y.
' KlliSEy AND STEWART: A Ni-Al-Mo CRDEP RESTSTANT ALLOY 9l
On the basis of the data listed in Table VIII, a casting temperature of1650'C. (3002'I.-.) was chosen for all alloys but 30M250, for which a castingtemperature of 1750" C. (3182" F.) r,vas usecl. The temperature of the test-bar moldslvas held lrithin the range of 649" to 677o C. (1200" to 1250'F.).
All test bars cast \vere subjected to X-raf inspection as a final check onsoundness. It tvas not possible to producc good test bars from Alloy 30VI250.Trro heats r'r'ere attcmpted. All bars of Atr1oy 30N'I250 produced containedshrinks and hot tears. This alloy was thcrcfore eliminated at this point.
The chemical analysis of the alioy heats that produced good test bars arelisterl in Table tX.
'IABLE IXCnnltrcer- ANALysrs oF ALLoys pRoDLTcING RADTocRApHIcALLy souND T]tsr BARS
\lloy
53X,I10.190N{ 1 5895Id255
2.58N{252
All but tr,vo of the crccp-rupture tests werc carried out at 815" C. (1500'F.).Thc results of these tests are summarized in 'fable X. The elongation vs.tinre curves for r\11o)'s 90\,I 158 and 95111255 are presented in Figs. 22 and23.
TABLE XSr,lrlrany oF cREEp-RUpruRE TESls er 815o C. (1500'F.)
I
-.1Siress, I
p.s.r. i
29,9+820,030
53 hI 104
90IlI 158
95M255
258M252
35 ,00029,94527 ,4+O
40, 00035 ,00035 ,00030,00027,600
25 ,00022,SOO13 ,500
Test temperature,9820 C. (18000 F.). Adapter bar broke in seven hours. Elong. 1|/p*
. _+N^o1-q: AJter-bei.ngstres-sedtot5,000p.s.i.atg82'C.Jorsewnhours,th,isbarwasused,foratest at 815" C. with a stress oJ 35,000 p.s.i. -
Between these tbsts the bar was unstressed. and, coi!.ed,lo room temperature.
Chemical analysis, /6
100
352236556913
0.033
o.o12o.oo27o.oo22
Extensometers not functioning
Extensometers not functioningSee note below*
Extensometers not f Inctioning
Can
. J. R
es. D
ownl
oade
d fr
om w
ww
.nrc
rese
arch
pres
s.co
m b
y D
AL
HO
USI
E U
NIV
ER
on
10/2
9/12
For
pers
onal
use
onl
y.
CANADIAN JOURNAL OF RESEARCTI. VOL. 27, SEC. F.
Frc.22. Creep-rupture curztes, Alloy No. 90M158 at 815" C. (1500' F.)'
I a 'n'
,rnKUT IUf,L 1 WNVLJ
FOR ALLOY gOM I58TEI.4PERATURE IsOCPF (BIs"C)srREss - A-27440 PS.r. I, B- 29.945 PS r. I
136 HRS.
l
CREEP RUPTURE CURVES
- FOR ALLOY 95M255 -TE|\,4PERATURE --- l500oF (8l5qc\
STRESSES A-275CO PS,I,
tlllllc-4o,oooPsl.I | | D-35.OOOPSI.
BRONroo IRS.
BrcKE AT5S HRS
o 1 I I
BROKE A1 .9l3tHRS.a-I / I
9 tr 2t
r P .f/ 2 .8
I
I / l> \-o f
con a-1
u-
9r P<TIME rcs
Frc. 23. Creep-rupture curtes, AIIoy No. 95M255 at 815 o C. Q5A0" F')'
Can
. J. R
es. D
ownl
oade
d fr
om w
ww
.nrc
rese
arch
pres
s.co
m b
y D
AL
HO
USI
E U
NIV
ER
on
10/2
9/12
For
pers
onal
use
onl
y.
RINSEY AND STEWART: A Ni-AT.MO CREEP RESISTANT ALLOY 93
In certain tests, as noted in Table X, the extensometers were not functioningproperly, and thcrcfore no creep data are available from these tests.
It rvill be noted that one test is reported on Alloy 95M255 at 982' C.
(1800'F.), using a stress of 15,000 p.s.i. This test lasted for seven hoursbcfore the adapter bar broke, at u'hich time the test bar had elongated t!/6.This test bar was later put into test at 815" C. (1500" F.) under a stress of35,000 p.s.i. It lasted for 236 hr. as compared to the 352 hr. of a duplicatebar that had been tested in the as-cast condition at 815o C. (1500" F.) usingastress of 35,000 p.s.i. All creep data for 35,000 p.s.i. at 815'C. (1500'F.)for Alloy 95NI255 u'ere obtained from the bar that was first tested at 982" C.
(1800'F.) and subsequentl)' tested at 815o C. (1500'F.).
The graphs shown in Figs. 24 and 25 summarize the results of the creep-
rupture tests on Alloys 90\"I 158 and 95NI255 and compare these alloys withsome of the existing standard and improved high temperature allo-vs. Thechemical analy'5gs of thc alloys uscd for comparison purposes are listed inTable XL These alloys arc: a standard vitaliium alloy, 30V-2(4); two modi-ficd vitallium alloys, 73Ie), and X-63(3);and alloys S-816 and Inconel "X".
Ftc. 24. Graph,ical summary of creep-rupture tests ot 8 I 5o C. (1 500" F.) on alloys 90 M 1 5 8ond, 9 5 M2 5 5 , and comparison with other h'igh temperature alloys.
Fig. 24 is a graph of log stress vs. log rupture time. Fig'. 25 contains twographs, one showing log stress vs. log time to 0.5/o total creep strain, and theother showing log stress vs. log rninimum crcep rate.
Can
. J. R
es. D
ownl
oade
d fr
om w
ww
.nrc
rese
arch
pres
s.co
m b
y D
AL
HO
USI
E U
NIV
ER
on
10/2
9/12
For
pers
onal
use
onl
y.
91 CANADIAN JOURNAL OF RESEARCH. VOL. 27, SEC. F.
Frc. 25. Creep-rupture results and comparisons at B15o C. (1500" F.).
As a further basis of comparison, it is interesting to consider specific gravityvalues of existing alloys and of Alloys 90M158 and 95M255. These data are
included in Table XL
TABLE XI
Crmr.lrc.q.r ANALysES AND sPEcrFrc GRAVTTY oF CoMPARATTvE ALLoYS
S-816 llnconel "X"
CobaltNickelChromiumNlolybdenumTantalumColumbiumTt r n qsfenTitaniumAluminumIronManganeseSiliconCarbon
Specific gravity
+.)2020
4
3.54.5
4 max.I.J0.50.3s
60o
ZJ5z
57.51025
o
69
z-)6
I
2.50.770.50.40.04
75.6
15.8
8.40.4
0.24
66.5
25.5
7.00.8
0. 160.50.50.5 0.3
8.30 7.6 8.0
* Not aztail.able,
LpG srR[ss-vl -+LqG ;frMF +p,l5_%_cREEP_ 1STRA N - AT- l5OO: F.
)R ALLOYS 9OM.t58 8 95N4 255 |
aLLOYSL90r/ r58 I 95M255
CREEP RATE PER CEN
Can
. J. R
es. D
ownl
oade
d fr
om w
ww
.nrc
rese
arch
pres
s.co
m b
y D
AL
HO
USI
E U
NIV
ER
on
10/2
9/12
For
pers
onal
use
onl
y.
KINSEY AND STEWART: A Ni-Al'Mo CREEP RESISTANT ALLOV 95
The data presented in Table XII indicate for the various allol's the stress
to cause both rupture and 0. 5/o total creep strain in 100, 1000, and 5000 hr.
In Table XIII the data are arranged on a basis of minimum creep rate.
TABLE XIT
cnEBp-nuprunE DATA ar 815'C. (1500'F.), suurtenrzED ON A TIME B-dsIS
100 Hr 5000 Hr.
27 ,OOO26 ,000
TABLE XIII
cnnEp-nupruRE DATA .q.r 815'C. (1500'F.), su'uuanrzED ON A MINIMUM CREEP RATE BASIS
Stress,p.s.i.
35 ,00040 ,00036 , O00
Alloy 90M158
RuptureO.5/6 Creep
strarnO.5/6 Creep
strain0.57o Creer)
stfarn
Min.
rate,7o Pet
_1_0.035o. 0330. 0025
Stress,p.s.l.
Mtn.
rate,o/o Der
hr.
Stress,p. s.i.
0.o.0.
Stress,p.s.i.
)La1
20,
1Vlln.
creeprate,
7o |erhr.
Min.creeprate,
Ta Perhr.
73J(3)95M25590M158
able.o017.oo22
Not avail2e,sooJ o32,000i 0
28 ,000 003 1
001 1
Minimumcreeprate,
Alloy 73J (3) Alloy 95M255
Stress,p.s.l.
37,00029,50026,A0A22,O0020 ,000
Hours torupture
Hours torupture
Hours torupture
0.050.010.0050.0020.001
56570
2,00010,000
,t2 ,00035 ,00032 ,00026,00020,000
70300400
1,2005 ,000
32,00022,000
2003 ,000
Metallography oJ Creep-Rupture Test Bars'fhe microstructures of the test bars produced from the alloys listed in
Table IX are presented in Figs. 26 to 39-
To reveal the complete microstructure of Alloy 53M104, as shown in Figs.
26 and,27, a double etching technique was employed. An alcohol solution of
ferric chloride and hydrochloric acid lvas used to bring out the light gfay areas
but this reagent alone would not clearly reveal the eutectic network. Inorclcr to bring out this eutectic network an electrolytic etch was employed,
using an oqu"o.r. solution of sodium clanide as the electrolyte. With this
etch it was necessary to use care to avoid over-etching' About 5 to 15 secs'
at 1 to 3 v. rvas adequate. Neither the electroll'tic etch nof the acid etch
alone revealed the complete microstructure.
Can
. J. R
es. D
ownl
oade
d fr
om w
ww
.nrc
rese
arch
pres
s.co
m b
y D
AL
HO
USI
E U
NIV
ER
on
10/2
9/12
For
pers
onal
use
onl
y.
96 :ANADTAN JIIRNAL oF RESEARCH. vol. zz, sEC. F.
Alloys 90M158 and 95NtI255 etched satisfactoril)' u.ith the electrolytictechnique alone (see Figs. 28 to 31 and 34to 37). However, as a further aidin studying the phases present in these alloys and also in studying structuralchanges that took place during creep-rupture testing, Figs. 32 and 33 areincluded. Thesc show the microstructure of Alloy 90NI 158 at 1500 diametersas revealed by a modified chrome-regia etch.
It was found difficult to reveal the microstructure of Alloy 258\[2s2 ina satisfactory manner. The reagent selected, nitric acid plus acetic acid inacetone, seemed to be the most effective.
D'iscuss'ion of Metallography as Related, to Creep-Rupture Properti,esAt this stage it is not possible positivell- to identifl' all phases revealecl in
the microstructure of the alloys studied.f'he electrolytic etch has shown that a phase that occurs in a eutectic manner
is common to Alloys 53NI104,90M158, and 95N{255. This phase is attackedand darkened by this electrolytic etching. Figs. 26 to 31 and 34 to 37 rvillillustrate this point. The identity of this phasc has previously been tenta-tively established as NiMo.
It rvill be noted that thc matrix of Alloy 53x4104 (see Figs. 26 and 2I)consists of a gray etching phase and a clear u'hite phase. This gray phase isetched by the ferric chloride - hydrochloric acid etch and has been tentativclyidentified as NiAl. It is significant that this alloy, 53M104, rvhich is the onlyalloy tested that contains this NiAl phase. yielded the poorest creep-ruptureproperties. A similar etching technique applied to Alloys 90N,{158, 95NlI255,and 258N4252 failed to reveal a similar phase.
'fhe tr,vo alloys that gave the best creep-rupturc performance at 1500'F.possessed sin-rilar microsrructures, as may be seen from Figs. 28 to 31 and 34to 37. 'fhis microstructure consists of :
(a) The previously ''rentioned NiMo eutectic phase that is darkened by
the electrolytic etch.
(b) clearl'hitc areas usually associated u,-ith Nix,{o eutcctic. 81' referenceto Fig. 10, these arcas are tentativei-v identified as NieAl and rvere originallythe alpha-NieAl eutectic (1).
(c) A mottled background that at a high magnification (see Figs.29,32,and 35) is seen to be a dark-etching, verl' finc netw-ork in a white background.The network is tentatively i6"t tttt.d as Ni;Al and the background as alpha(see Figs 10 and 11).
The clear lvhite areas of Ni:Al are more numerous in Alloy 90N,I 1s8 and theNiNIo eutectic is more prevalent and continuous in Alloy 95NI255 (see Figs.28 and 34). It is considered significant that these two microstructuralfcatures are always associated w-ith each otl-rer; this helps to confirm thepossibility that the clear NiaAl arcas are eutectic in oriqin.
Frc.36. Altoy (.)5J[255, oJter clcep-ruptu,re tesling ot 3'15" C' (1500" F') for 93a hr'Electrolytic etch: 1)rs sotliunt cyan,ide ot 3 zt. X200-
Frr;" 37. So,mc as F'ig. 36. X1500.
Can
. J. R
es. D
ownl
oade
d fr
om w
ww
.nrc
rese
arch
pres
s.co
m b
y D
AL
HO
USI
E U
NIV
ER
on
10/2
9/12
For
pers
onal
use
onl
y.
['r,a,rn \/ll
,_l'tg, 39,_ Alloy 258)t252, as cost. Etch: nitric acid plus oce.tic ocitl in acetone. X200.Ni, 72.101/s; Al, 2 .8ft; Mo, 25 .2ak.
_ ftc. 3.0.. Al.loy -258,11252, ulter creep-ruptlLre testittg rLt. 515" C. (1500" F.) for 460 hr.Ll.h: nttrt. acid plus uretic acid in ucetone. X200. Ni, Z2.lci; Al, Z.8o/o; M0,25.Zak.
Can
. J. R
es. D
ownl
oade
d fr
om w
ww
.nrc
rese
arch
pres
s.co
m b
y D
AL
HO
USI
E U
NIV
ER
on
10/2
9/12
For
pers
onal
use
onl
y.
KINSE\ AND STEWART: A Ni'At-Mo CREEP RESISTANT' AI'I'OY 97
After exposure to the temperature of testing at 815o C. (1500'F.) there is
evidence that the NieAl nctu,ork has coalesced. A comparison of Fig. 29
with Fig.31, and of Fig.35 ivith Fig.37, u'-ill illustrate this'
Further to studlr the phases present in Allo-vs 90M158 and 95I11255' various
other etching reagents \\rcre emplo),cd. A modified alcoholic solution of
chrome-regia scemed to be the most useful. The microstructures of Alloy
90M158 (as developcd by this reagent), both before and after creep-rupture
testing, are shown at a magnification of X 1500 in Figs. 32 and 33. The NiNIo
eutectic phase that w-as darkencd by the electrolytic etch was only outlined by
the chrome-regia etch. The network structure show-n in Fig. 29 has been
brought out more clearll, by this chrome-regia etch in Fig. 32. The effect
on this network of prolonged exposure to a temperature of 815" C. (1500'F')is well illustrated by Fig. 31. Here the coalescence of this network phase is
unmistakable.'lhe fourth alloy tested, 248M252, is not of much interest since its properties
$,ere quite inferior. Its microstructure, as shor'vn in Fig. 38, is largely single-
phase, probably alpha solid solution, rvith small patches of a second phase
rvhich can be resolvcd into a eutectic-like appearance at high magnification.
The microstructure is radically altered during creep-rupture testing, as may
be seen in Fig. 39. From a study of Ellingcr's work (2), this new phase in
Fig. 39 would appear to be the nickel-molybdenum phase, gamma (Ni3Mo)'
General Conclusions
It is evident that the r,vork reported in this paper has provided the startingpoint for the development of an entirely ner'v series of high temperature alloy5
basecl on nickel, aluminum, and one or more other elcments, such as mol1'b-
denum.
There is evidence that the microstructure of these alloys may be used as a
criterion of their usefulness, since any alloy containing an apprcciable propor-
tion of the phase NiAl woulcl not be expected to be suitable.
While most of the clata presented are onll' the results of single melts and
single tests, the indications afe nevcrthcless quite promising. One of the
allo-vs procluced (95\'1255) compares quite favorably u,'ith the best of other
knorvn high tempcrature alloys. Alloys of this nerv series are readily melted
by high frequency induction and present no unusual casting difficulties.
Acknowledgments
This work iras been made possible through the co-operation of the Bureau
of N{ines and the National Research Council's Associate Cornmittec on High
Temperature N{etals. Gratcful acknor,r'lcdgment is made for assistance and
aclvice received, during the preparation of this paper, from H, H. Bleakney,
Can
. J. R
es. D
ownl
oade
d fr
om w
ww
.nrc
rese
arch
pres
s.co
m b
y D
AL
HO
USI
E U
NIV
ER
on
10/2
9/12
For
pers
onal
use
onl
y.
98 :ANADIAN Jot)-RNAr, oF RESEARCH. voL. zz, sEC. F.
Associate Research officer of the National Research council and secretaryof the Associate Committee on High remperature N{etals. It is also desiredto acknorvledge the work of R. F. cole, presently of the Ford Motor companyof Canada, Limited, who assisted in the earlier stages of this project.
Referencesl. ArBxeNnBn, \\r. O. and \,'eucneN, N. B. J. Inst. Metals, 61 :247_260. Ig37.2. Enrwcor, F. H. Trans. Am. Soc. Metals, 30 : 607_63g. lg+2.3. Ernnnr,r.w, E. Trans. Am. Soc. Metals, 39 :261-2g0. lg+7.4. GneNr, N. J. Trans. Am. Soc. Metals, 40 : 5g5-616. 194g.
