METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 34A, OCTOBER 2003—2207

I. INTRODUCTION

THERE is the demand for high-temperature materialsthat can be used at around 2073 K in new and challengingapplication fields, such as a heat exchanger in a ram jetengine or a thruster in a satellite.[1] However, the suitablematerials are very limited. To develop a high-temperaturematerial that can be used at 2073 K, we noted an fcc andL12 coherent two-phase structure, because it is well knownthat the fcc and L12 coherent two-phase structure plays animportant role in strengthening Ni-based superalloys, whichare used in many application fields, such as the gas turbineof a generator in a power station or the jet engine in anairplane at high temperatures.[2] We have designed “refractorysuperalloys” based on Ir with an fcc and L12 coherent two-phase structure.[3] Ir is chosen as a base material because ofits high melting temperature (2720 K). Ir crucibles are usedto produce a single crystal of an electric memory or laserchip at 2223 K, and Ir is suitable as a structural material ataround 2073 K.

In previous studies, we investigated several binary alloysand found that Ir-Nb, Ir-Ta, Ir-Hf, and Ir-Zr alloys exhibitthe fcc and L12 two-phase coherent structure.[4] The meltingtemperature of these alloys is above 2273 K. It was foundthat the microstructure depends on the lattice misfit betweenthe fcc and L12 phases. For example, the cuboidal L12 pre-cipitates formed in the Ir-Nb and Ir-Ta alloys whose latticemisfit was 0.4 pct, while the maze structure formed in theIr-Zr alloy whose lattice misfit was 2.5 pct. We investigatedmainly Ir-Nb and Ir-Zr alloys. The compressive strength ofthe Ir-Zr alloys at 1473 K was higher than that of the Ir-Nballoys because of a fine maze structure, which prevented dis-

location movement.[5] However, the Ir-Nb alloy showedslower minimum creep rates than the Ir-Zr alloy at 1773 K.[6,7]

This is because the maze structure in the Ir-Zr alloys changedto a coarse lamellar structure through a discontinuous coars-ening process by a large lattice misfit during the creep testunder 137 MPa at 1773 K. Cuboidal L12 precipitates in theIr-Nb alloy were very stable in size and shape, with no changeoccurring after 300 hours of creep testing. We consider thatcontrol of the lattice misfit is important to develop an alloywith good properties in both high-temperature strength andcreep resistance. The lattice parameters of Ir3Nb and Ir3Zrare 0.38915 and 0.3941 nm, respectively, from the dataobtained from JCPDS. It is considered that the addition ofZr to Ir3Nb increases the lattice parameter of Ir3Nb andchanges the lattice misfit between the fcc and L12 phases.In order to evaluate changing the lattice parameter on theaddition of Zr to Ir-Nb alloys, the present investigation wascarried out.

An investigation of the phase diagram of Ir-Nb-Zr hasbeen performed. Gyurko and Sanchez suggested that thereis an Ir3Nb � Ir3Zr two-phase structure.[8] However, Millercalculated the Ir-Nb-Zr ternary-phase diagram by the first-principle method and showed complete solid solubility be-tween Ir3Nb and Ir3Zr in the temperature range between 1200and 1900 K.[9] Freitas and Sanchez characterized the phaseof Ir3(Nb0.95Zr0.05), Ir3(Nb0.5Zr0.5), and Ir3(Nb0.05Zr0.95)and found that all alloys showed a single L12 phase.[10] Wealso determined the Ir corner of the ternary Ir-Nb-Zr phasediagram by microstructural characterization and found thatNb and Zr were fully miscible in the L12 phase.[11] Thus,the wide fcc and L12 two-phase region existed from theIr-Nb to the Ir-Zr side. The determined phase diagram at1773 K is shown in Figure 1.[11] This suggests that it ispossible to prepare the alloys with different lattice misfits.However, the mechanical properties of the Ir-Nb-Zr alloyshave not yet been studied systematically. In this study, themicrostructure and high-temperature mechanical propertiesof the Ir-Nb-Zr alloys with the fcc and L12 two-phasestructure were investigated for first time. Then, the strengthand creep properties were discussed in terms of the latticemisfit, precipitate morphology, and deformation structure.

Compressive Strength and Creep Properties of Ir-Nb-ZrAlloys between 1473 and 2073 K

Y. YAMABE-MITARAI, Y. GU, and H. HARADA

The microstructure and strength at 1473 and 2073 K and creep properties at 2073 K were investigatedin three Ir-Nb-Zr alloys with the fcc and L12 two-phase structure. The microstructure and latticemisfit were investigated using scanning electron microscopy (SEM), transmission electron microscopy(TEM), and X-ray diffractometry (XRD). Compression and creep tests were performed, and theirdeformation structures were observed using SEM and TEM. At 1473 K, the strength of the Ir-Nb-Zralloys was higher than that of the binary Ir-Nb and Ir-Zr alloys, but they were almost equivalent at2073 K. However, the ternary alloys showed great improvement on creep at 2073 K. The time forthe 2 pct creep strain of the Ir-Nb-Zr alloy was about 100 hours, while it was 1 hour for the binaryalloys. The deformation mechanisms for compressive strength and creep resistance in these Ir-Nb-Zralloys are discussed in terms of the deformation structure.

Y. YAMABE-MITARAI, Senior Researcher, Y. GU, Project Researcher,and H. HARADA, Project Director, are with the National Institute for Mate-rials Science, Ibaraki 305-0047, Japan. Contact e-mail: mitarai.yoko.nims.go.jp

This article is based on a presentation made in the symposium entitled“Fundamentals of Structural Intermetallics,” presented at the 2002 TMSAnnual Meeting, February 21–27, 2002, in Seattle, Washington, under theauspices of the ASM and TMS Joint Committee on Mechanical Behaviorof Materials.

16_02-101A-Symp.qxd 11/9/03 4:14 PM Page 2207

2208—VOLUME 34A, OCTOBER 2003 METALLURGICAL AND MATERIALS TRANSACTIONS A

Fig. 2—Secondary electron images of (a) alloy A (Ir-12.5 at pct Nb-3 atpct Zr), (b) alloy B (Ir-8.5 at pct Nb-6 at pct Zr), and (c) alloy C (Ir-4.25Nb-9Zr) heated at 1773 K for 72 h.

Fig. 1—The experimentally determined phase diagram of the Ir corner ofthe Ir-Nb-Zr at 1773 K.[11]

*PHILIPS is a trademark of Philips Electronic Instruments Corp.,Mahwaha NJ.

II. EXPERIMENTAL PROCEDURES

Three alloys with nominal compositions of Ir-12.5Nb-3Zr,Ir-8.5Nb-6Zr, and Ir-4.25Nb-9Zr (at. pct), were prepared as15 g button ingots by the arc-melting method. These alloycompositions are located on the line between the Ir-17 at. pctNb and Ir-12 at. pct Zr alloys with 50 pct of the volumefraction of the L12 precipitates in each binary system, asshown in Figure 1. In Figure 1, the square symbols showthe nominal composition. The solid symbols represent theactual composition of the alloy, determined experimentallyby electron-probe microscopy analysis (EPMA) in a prev-ious study. The composition of alloy A deviates from thenominal composition, but the others almost have the targetedcompositions.

Cylindrical samples, 3 mm in diameter and 6 mm inheight, were cut from the button ingots and heat treated at1773 K for 72 hours in a vacuum furnace. The samples werethen cooled in the furnace. These heat-treated samples weresliced into 200-mm-thick disks for all samples. One piece ofthe sliced sample in each alloy was embedded in a resin,mechanically ground with sandpaper, and then electrolyticallyetched in an ethyl alcohol solution of 5 pct HCl to observethe microstructure using scanning electron microscopy (SEM,PHILIPS* XL30). They were also used for the investigation

of lattice parameters by X-ray diffractometry (XRD). Themeasured 2u was between 40 and 120 deg. Each peakfrom the fcc or L12 phases was clearly detected using XRD.Then, the lattice parameter was determined by least-squaresrefinements. The lattice misfit (d) was calculated by the followingequation:

where af and aL are the lattice parameters of the fcc and theL12 phases, respectively. Other sliced samples were polishedto obtain a thickness of 30 mm using a dimpling machine.

d � (af � aL)/af

They were then ion milled into thin foils to observe the pre-cipitate morphology using transmission electron microscopy(TEM, PHILIPS CM200).

The cylindrical samples heat treated at 1773 K for 72 hourswere used for compression tests at 1473 and 2073 K and forcompression creep tests at 2073 K under 137 MPa. Thetest was carried out at a strain rate of 10�4/s in a vacuumatmosphere using an Instron 8560 or a Shimazu (Kyoto,Japan) AG-I testing machine. Before testing, the sample waskept at the testing temperature for 15 minutes. The result-ing deformation structures were characterized using SEMand TEM techniques.

III. RESULTS

A. Microstructure

The typical microstructures of alloys heat treated at 1773 Kfor 72 hours are shown in Figures 2(a) and (b). In alloy A,

16_02-101A-Symp.qxd 11/9/03 4:14 PM Page 2208

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 34A, OCTOBER 2003—2209

Fig. 3—Dark-field images of (a) and (b) alloy A (Ir-12.5Nb-3Zr), (c) through (e) alloy B (Ir-8.5Nb-6Zr), and (f) through (h) alloy C (Ir-4.25Nb-9Zr) heattreated at 1773 K for 72 h. These images were taken from the L12 superlattice spot.

the dendritic structure with the fcc structure was present(Figure 2(a)). A coarse fcc and L12 lamellar structure wasobserved around the dendrite arms. Fine precipitates wereobserved in the dendrite arms as well as in the interden-drite area. From the contrast of the image, we can identythe fine L12-phase precipitates in the fcc phase and the finefcc-phase precipitates in the L12 phase. In alloy B, the micro-structure consisted of dendrite arms and the coarse fccparticles with fine L12 precipitates (Figure 2(b)). Betweendendrite arms, the L12 phase with fine fcc precipitates wasobserved. In alloy C, irregularly shaped coarse fcc and L12

phases with fine precipitates were observed in both phases.The precipitate morphology of alloys A through C was

observed using TEM. Dark-field images from the L12 phaseare shown in Figures 3(a) through (f). As shown in theobservation by SEM, the microstructure was not homoge-neous. Thus, we found a different precipitate morphologyin each alloy. Broadly, the structures can be classified intothree groups. The first group (type A) includes fine cuboidalL12 precipitates in the fcc phase 100 nm in size, as shownin Figures 3(a), (c), and (f). In this case, the precipitatesare surrounded by {100} habit planes, and they tend toalign along the �001� orientation. Furthermore, we alsofound platelike fcc precipitates in the cuboidal L12 phase(Figures 3(a) and (f)). These fcc plates formed whenthe composition of precipitates with a nonequilibrium stateformed from the as-cast microstructure, and then the com-

position changed to the equilibrium state from the non-equilibrium state during the heat treatment.[12] The secondgroup (type B) is a platelike fcc phase in large L12 parti-cles, as shown in Figures 3(b), (d), and (g). The habit planesof these fcc plates were also the {100} planes. The groupA and B structures were coherent, but the third group (typeC) was an incoherent or semicoherent L12 phase with anirregular shape, as shown in Figures 3(e) and (h). The mazestructure observed in the Ir-Zr alloy was not found in thesealloys.

B. Lattice Misfit

The lattice parameters of the fcc and L12 phases of thealloys are shown as a function of the Zr composition inFigure 4. For reference, the lattice parameters of the Ir-basedbinary alloys represented by the Ir-15Nb and Ir-15Zr binaryalloys are also plotted.[13] Although we do not have thelattice-parameter data for the Ir-17Nb and Ir-12Zr alloys, itis considered that the lattice misfits of each binary alloyare equivalent in the two-phase region. The lattice param-eters of the fcc phase of the alloys did not depend on thealloy composition. However, the lattice parameter of the L12

phase increased with increasing Zr content, because thelattice parameter of Ir3Zr is larger than that of Ir3Nb. Thelattice misfit estimated from the data in Figure 4 is shownin Figure 5. The lattice misfits of ternary alloys were between

16_02-101A-Symp.qxd 11/9/03 4:14 PM Page 2209

2210—VOLUME 34A, OCTOBER 2003 METALLURGICAL AND MATERIALS TRANSACTIONS A

Fig. 4—The lattice parameters of the Ir-Nb-Zr alloys heat treated at1773 K for 72 h as a function of the Zr concentration.

Fig. 5—The lattice misfit of the Ir-Nb-Zr alloys as a function of the Zrcomposition.

Fig. 6—The compressive strength of the Ir-Nb-Zr alloys at 1473 and 2073 K.

Fig. 7—The creep curves of the Ir-Nb-Zr alloys at 2073 K under 137 MPa.

1.3 and 1.8 pct and were located between those of the Ir-Nband Ir-Zr alloys.

C. Mechanical Properties

The compressive strengths at 1473 and 2073 K are shownas a function of the Zr concentration in Figure 6. For com-parison, the strengths of the binary alloys are also presented.[5]

At 1473 K, the strengths of the ternary alloys were higherthan those of the binary alloys. On the other hand, at 2073 K,only the ternary alloy C (I-4 25Nb-9Zr) showed a higherstrength than the binary alloys, but the other two alloysshowed almost the same strength levels as those of the binaryalloys.

To investigate its potential as a high-temperature material,a creep test was performed for alloy A (Ir-12.5Nb-3Zr)at 2073 K under 137 MPa (Figure 7). For comparison, thecreep curves of the binary alloys under same conditionare also presented.[7] The binary alloys showed tertiarycreep during a few hours. On the other hand, the secondarycreep continued for up to 90 hours, and the creep life forthe 2 pct creep strain was about 100 hours in the Ir-Nb-Zralloy.

D. Deformation Structure

Figure 8 shows the deformation structure observed usingSEM of alloy A (Ir-12.5Nb-3Zr), tested at 1473 and 2073 K.The dendritic structure did not change during the test at bothtesting temperatures (Figures 8(a) and (d)). It is seen fromthe high-magnification images (Figures 8(b) and (e)) thatcoarsening did not occur, and the cuboidal precipitatesretained their size and shape. However, cracks at the inter-dendrite region were observed in Figures 8(c) and (e). Thedeformation structure was almost equivalent at 1473 and2073 K. On the other hand, although the samples retainedtheir dendritic structure after the creep test at 2073 K, theirmicrostructure clearly coarsened, as shown in Figures 9(a)and (b). The fine precipitates in the fcc phase changed tocoarse particles (shown in Figure 9(a)) or a rod or plateshape (shown in Figure 9(b)). The coarse L12 phase wasalso observed in the interdendrite region. Cracks were alsoobserved between fcc particles or the fcc/L12 interface.

The deformation structure in the transverse cross sectionof alloy A (Ir-12.5Nb-3Zr) after the compression test wasobserved in TEM (Figures 10(a) through (d)). After the com-pression test with the strain of 4.8 pct at 1473 K, a largenumber of dislocations bowed out and made dislocationloops in the fcc-phase channels in the type-A structure(the fcc phase with cuboidal L12 precipitates) (Figure 10(a)).

16_02-101A-Symp.qxd 11/9/03 4:14 PM Page 2210

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 34A, OCTOBER 2003—2211

Fig. 8—The deformation structure of alloy A (Ir-12.5Nb-3Zr) compression tested at (a) through (c) 1473 K and (d ) and (e) 2073 K.

A similar structure was observed in the Ir-Nb alloy tested at1473 K.[14] Although the Burgers vector was not determinedin this observation, they are a/2�110�-type dislocations, basedon the analysis of dislocations observed in the Ir-Nb alloy.This structure is similar to that of Ni-based superalloys duringthe incubation period of creep.[15,16] Using the stereo imagingmethod, it has been established that these dislocations inNi-based superalloys are located in the fcc matrix channel.[15]

Thus, the dislocation loops observed in the Ir-Nb-Zr alloyare in the fcc matrix channel. We believe that we observedthe dislocation loops, which were located in the horizontalchannel to the applied stress. In the type-B structure (L12

phase with platelike fcc precipitates), a few dislocation loopswere observed (Figure 10(b)). These dislocation loopsexpanded from the platelike fcc phase. We believe thatwe observed the dislocation loops, which spread in theinvisible fcc channel normal to the visible fcc channel inthe micrograph. In Figures 10(c) and (d), the deformationstructure of the sample tested with the strain of 0.86 pctat 2073 K is shown. In the type-A structure, dense dislocationswere observed in the whole area studied (Figure 10(c)).

The dislocations were spread more widely than in the sampletested at 1473 K, and a “zigzag” morphology was observedin many dislocations. The “zigzag” morphology consistedof dislocations along the �110� and �100� orientations, asshown by the arrows marked “s” and “m,” respectively. Thissuggests that mobile dislocations with different Burgersvectors meet and react and, as a result, dislocations alongthe �110� slip direction with a curve along the mismatch�100� orientation form, as reported by Field et al. in creptNi-based superalloys.[17] In the type-B structure as well asin the sample tested at 1473 K, a few dislocation loopswere observed (Figure 10(d)). We did not find a deformationstructure in the type-C structure in both samples tested at1473 and 2073 K.

Three kinds of images of alloy A (Ir-12.5Nb-3Zr), takenunder different diffraction conditions using TEM after thecreep test at 2073 K under 137 MPa, are shown in Figures11(a) through (c). In most of the area, the precipitateschanged to an irregular, long shape. Thus, it is consideredthat all three kinds of structures (types A through C) changedto the same irregular structure. Compared with the secondary

16_02-101A-Symp.qxd 11/9/03 4:14 PM Page 2211

2212—VOLUME 34A, OCTOBER 2003 METALLURGICAL AND MATERIALS TRANSACTIONS A

Fig. 9—The deformation structure of alloy A (Ir-12.5Nb-3Zr) crept at2073 K in compression.

electron image (SEI) in Figures 9, it is considered that thecoarse rod or plate precipitates corresponded to the irregularlyshaped precipitates observed using TEM. The area A withfine precipitates, shown in Figure 9(b), appeared whenirregular rod-shaped precipitates were observed from the tipdirection of the rod. The dislocation network was clearlyobserved at the interface in Figures 11. The dislocations withan a/2[01 �-type Burgers vector were invisible in the beamdirection of [ 11] (Figure 11(a)). When they were observedusing [ 1] and [0 0], dislocations with a/2�1 0] and a/2[ 01]-type Burgers vectors were invisible (Figures 11(b) and (c)).Our observation showed that this network consisted of threekinds of Burgers vectors, i.e., a/2[01 �, a/2 � 1 0], and2/a[ 01], and would, thus, show a hexagonal configurationon the {111} plane, as suggested by Lasalmonie andStrudel.[18] In addition to the dislocation network, dislocationsin the L12 phase suggest that creep deformation occurs byshearing of L12 phase in the Ir-Nb-Zr alloy.

IV. DISCUSSION

A. Precipitate Morphology and the Lattice Misfit

It was difficult to get a homogeneous microstructure,because the solid-solution treatment cannot be performedon the tested alloys. However, after heat treatment, fine L12

and fcc precipitates were formed in the primary fcc and L12

phase, respectively. The lattice misfits of ternary alloys werebetween positive 1.3 and 1.8 pct. They are larger than thoseof conventional Ni-based superalloys with cuboidal precip-itates, which is below 1 pct in absolute value. In a previousstudy, we reported that the precipitate shape depends onthe lattice misfit.[4] When the lattice misfit was small (around

111

112111

1

0.3 pct), cuboidal precipitates were formed, and, for a largemisfit of around 2.5 pct, platelike precipitates were formed.The present study showed that the coherent cuboidal L12

phase existed in all tested alloys (type A in Figures 3(a),(c), and (f)), even though their lattice misfits were quite large,in the range from 1.3 to 1.8 pct. We also found a coherentstructure consisting of the platelike fcc phase in the L12

phase (type B in Figures 3(b), (d), and (g)) in the samesamples. This suggests that, in the Ir-Nb-Zr alloys, theprecipitate morphology did not depend on the lattice misfitalone. Instead, it was determined by which phases werematrix and precipitate, i.e., when the fcc phase precipitatedin the L12 phase, the precipitate shape became platelike. Thismay be due to the difference in the elastic constants in eachphase. A maze structure could be formed if the density ofthe fcc phase was large in the type-B structure. Thus, weconsider that the maze structure in the Ir-15Zr alloy alsoformed by precipitation of the platelike fcc phase in the L12

phase. This is possible for the Ir-15 at. pct Zr alloy, becausethe primary phase of the ascast sample was the L12 single phase.

B. Compressive Strength

In the type-A structure tested at 1473 K, dislocationsspread in the narrow fcc channel (Figure 10(a)). Whencuboidal precipitates align along the cuboidal orientation,it is difficult to form Orowan loops around individual pre-cipitates and, thus, a mobile a/2�110�-type screw dislocationslips on one {111} plane, leaving segments of mixed char-acter. As a result, dislocation loops form.[19] This is also abypass mechanism, but not an Orowan mechanism. In ourobservation, cuboidal precipitates aligned along the cuboidalorientation (Figure 3). The dislocation loops in the fcc phasein Figure 10(a) suggested that deformation of the Ir-Nb-Zralloy was not governed by an Orowan mechanism, but byanother type of bypass mechanism similar to the one oper-ating in Ni-based superalloys. The local deformation by dis-location loops is often observed during the incubation periodof creep in the Ni-based superalloys.[15] Note that the localdeformation by dislocation loops was observed in thecompression-tested sample in the Ir-Nb-Zr alloy. There isnot enough time to spread dislocations for a long distanceduring the compression test. Thus, only small dislocationloops of 100 to 200 nm in size were observed, and defor-mation occurred locally. Another interesting point is thatdeformation also occurs in the fcc plate in the L12 phase,but not in the large L12 area in the type-B structure (Figure10(b)). We observed only a few dislocation loops in thetype-B structure. The fcc plates formed independently inthe type-B structure and, as a result, the formation of thefcc channel was incomplete, compared to those of the longcontinuous fcc channel in the type-A structure. Thus, it wasdifficult to spread dislocation loops for long distances in thetype-B structure. This result indicates that the deformationresistance is higher in the type-B structure than in the type-Astructure.

The strength of the ternary alloys was higher than that ofthe binary alloys at 1473 K. It is interesting to note that thedeformation structure of type A was similar to that of binaryalloys such as the Ir-Nb alloy.[14] However, the dislocationdensity was higher in the Ir-Nb alloy than in the Ir-Nb-Zralloy. It is believed that the solid-solution hardening effect

16_02-101A-Symp.qxd 11/9/03 4:14 PM Page 2212

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 34A, OCTOBER 2003—2213

Fig. 10—The dislocation structure of alloy A (Ir-12.5Nb-3Zr) tested at (a) and (b) 1473 K and (c) and (d) 2073 K. The zone axis and diffraction directionare (a) [101] and [020], (b) [101] and [1 ], (c) [001] and [ 00], and (d ) [101] and [20 ].2211

in the fcc phase determines the deformation resistance, becausethe fcc phase is mainly responsible for deformation. Asa consequence, the resistance to dislocation motion in thefcc phase in the Ir-Nb-Zr alloy was higher than that in theIr-Nb alloy. In the Ir-Nb-Zr alloys, the maze structure waschanged to the semicoherent structure with piled-up dislo-cations at the interface during the test.[5] On the other hand,in the Ir-Nb-Zr alloy, both the type-A and type-B structureswere very stable during the test. This suggests that themicrostructural stability of the Ir-Nb-Zr alloy causes higherdeformation resistance than that in the Ir-Zr alloy. It can besaid that the existence of the type-B structure increasesdeformation resistance in the tested alloys, because it wasnot deformed.

At 2073 K, dense dislocations spread in the fcc phase inthe type-A structure (Figure 10(c)), while in binary alloys,shearing into precipitates by dislocations was observed at2073 K.[5,14] In spite of the difference of the deformationstructure, ternary alloys showed strengths that were almostequivalent to those of the binary alloys. This suggests thatboth bypass and shearing mechanisms are not effective inimproving strength at 2073 K. However, it is interesting tonote that we observed only a few dislocation loops in the fccphase in the type-B structure. We consider that the type-B

structure is still stable and can prevent the movement ofdislocations at 2073 K.

C. Creep Behavior

Although the strength at 2073 K did not improve in theternary alloys, the creep life improved drastically. Duringthe creep test, the precipitate shape changed to the irregu-lar shape with the dislocation network at the interface. Thedislocation network is often observed in Ni-based super-alloys during creep tests.[16] It is suggested that the formationof dislocation networks occurs by local reactions between linesegments of dislocations with a slip orientation of a/2�110�.[19]

Thus, the dislocation network has a sessile structure andobstructs dislocation motion. When deformation is governedby a shearing mechanism, a mobile dislocation will trapat the interface with the dislocation network by reactionbetween a mobile dislocation and dislocation network. Thus,we consider that the formation of the dislocation networkcauses the creep resistance to be high. The dislocation networkwas also observed in Ir-Nb-Ni alloys,[20] which showed asuperior creep property similar to that of the Ir-Nb-Zr alloyat 2073 K. We do not confirm the deformation structureyet in binary alloys crept at 2073 K, because the sample

16_02-101A-Symp.qxd 11/9/03 4:14 PM Page 2213

2214—VOLUME 34A, OCTOBER 2003 METALLURGICAL AND MATERIALS TRANSACTIONS A

Fig. 11—The dislocation structure of alloy A (Ir-12.5Nb-3Zr) after creeptesting for 300 h, at 2073 K. The zone axis is [101], and the diffractiondirections are (a) [01 ], (b) [ 1], and (c) [0 0].2111

quickly deformed and failed. However, shearing of precipitatesby dislocations was observed in the Ir-Nb alloy at 1923 K.[21]

From that observation, it can be said that the dislocationnetwork prevented shearing of precipitates and resulted inthe improvement of the creep property. Another importantpoint is that discontinuous coarsening, which was observedin the Ir-Zr alloy with the lattice misfit of 2.5 pct, did notoccur in the Ir-Nb-Zr alloy due to the lattice misfit below2 pct. Microstructure stability seems to play important rolein improving the creep properties.

D. Design of High-Temperature Materials

We have tried to add third elements such as Ni, Mo, Ta,C, B, and Pt to increase the strength at high temperatures.[22]

Although Ni-, C-, and Pt-added alloys improved the strengthat 1473 K, we have not succeeded in improving the strengthabove 1773 K. The addition of other elements caused achange in the constitutent phase or a loss of interface co-herence, resulting in a decrease of strength. In this study,Ir3Nb and Ir3Zr were fully miscible, and only the fcc andL12 two-phase region was, thus, formed without any otherphase. Furthermore, the coherent structure was also formed.From the viewpoint of the phase constitution andmicrostructure, the Ir-Nb-Zr is a very promising system.However, the strength at 2073 K was not improved. Theresults suggest that it is very difficult to improve the strengthat 2073 K by the simple addition of a third element. Analternative idea is microstructure control. As shown in Figures10, the platelike fcc precipitate in the L12 phase (type-B struc-ture) seems to have a high resistance against deformationbecause of a lack of dislocations in the deformed structure.If we can obtain a large amount of the type-B structure andan L12 matrix strengthened by the solid-solution hardeningeffect, it may be possible to improve the strength at 2073 K.On the other hand, the addition of a third element is veryeffective in improving the creep-life properties at 2073 K,as shown in the case of Ir-Nb-Zr and Ir-Nb-Ni alloys. Therod-type structure with a well-developed dislocation networkformed because the lattice misfit of the Ir-Nb-Zr alloy wasfound to be in the range of 1.3 to 1.8 pct, which is higherthan that of the Ir-Nb alloy but lower than that of the Ir-Zralloy. The present study showed that the dominant changeof lattice misfit improved the creep resistance. However, amore detailed investigation is necessary.

V. CONCLUSIONS

The microstructure and strength at 1473 and 2073 K andcreep behavior at 2073 K were investigated in three Ir-Nb-Zralloys (Ir-12.5Nb-3Zr, Ir-8.5Nb-6Zr, and Ir-4.25Nb-9Zrat. pct). The lattice misfits of these alloys were between1.3 and 1.8 pct. Two kinds of coherent structures coexistedin the alloy, i.e., cuboidal L12 precipitates in the fcc phaseand platelike fcc precipitates in the L12 phase. The strengthof the ternary alloys at 1473 K was higher than that of thebinary Ir-Nb and Ir-Zr alloys. This is because the structurewith platelike fcc precipitates in the L12 phase preventsthe movement of dislocations. However, the strength of theternary alloys at 2073 K was almost equivalent to that ofthe binary alloys. This is because the coherent structurecannot prevent the dislocation motion in ternary alloys aswell as those in binary alloys. On the other hand, the creepproperty at 2073 K was drastically improved in the ternaryalloy. The creep life for the 2 pct creep strain was about100 hours, while the binary alloys deformed more than 5 pctduring the first 2 hours. This enhancement has been attributedto the formation of a dislocation network at the interface,which prevents shearing of the L12 phase.

ACKNOWLEDGMENTS

We appreciate the technical support received from Messrs.S. Nishikawa and T. Maruko, Furuya Metal Co., Ltd.

16_02-101A-Symp.qxd 11/9/03 4:14 PM Page 2214

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 34A, OCTOBER 2003—2215

REFERENCES1. N. Tanatsugu, T. Sato, Y. Naruo, T. Kashiwagi, T. Mizutani,

T. Monji, and K. Hamabe: Acta Astronautica, 1997, vol. 40,pp. 165-70.

2. N.S. Stoloff: in Superalloys II, C.T. Sims, N.S. Stoloff, and W.C. Hagel,eds., John Willey & Sons, New York, NY, 1987.

3. Y. Yamabe, Y. Koizumi, H. Murakami, Y. Ro, T. Maruko, andH. Harada: Scripta Mater., 1996, vol. 35(2), pp. 211-15.

4. Y. Yamabe-Mitarai, Y. Ro, T. Maruko, and H. Harada: Metall. Mater.Trans. A, 1998, vol. 29A, pp. 537-49.

5. Y. Yamabe-Mitarai, Y. Ro, S. Nakazawa, T. Maruko, and H. Harada:Def. Diffus. Forum, 2001, vols. 188–190, pp. 171-84.

6. Y. Yamabe-Mitarai, S. Nakazawa, and H. Harada: Scripta Mater., 2000,vol. 43, pp. 1059-64.

7. Y. Yamabe-Mitarai, S. Nakazawa, and H. Harada: JSME Int. J. Ser.A, 2002, vol. 45(1), pp. 2-7.

8. A.M. Gyurko and J.M. Sanchez: Mater. Sci. Eng. A, 1993, vol. 170,pp. 169-75.

9. G.A. Miller: Master’s Thesis, University of Texas at Austin, Austin,TX, 1994.

10. C.A. Freitas and J.M. Sanchez: Proc. THERMEC ’97, TMS, Warrendale,PA, 1997, pp. 1567-73.

11. Y. Yamabe-Mitarai and H. Harada: J. Alloy and Comp., in press.12. Y. Yamabe-Mitarai and H. Harada: Phil. Mag. Lett., 2002, vol. 82(3),

pp. 109-18.13. Y. Yamabe-Mitarai and H. Harada: Mater. Sci. Eng. A, in press.14. Y. Yamabe-Mitarai and H. Harada: unpublished research, 2001.15. T.M. Pollock and A.S. Argon: Acta Metall. Mater., 1992, vol. 40(1), pp. 1-30.16. S. Tian, J. Zhang, X. Wu, H. Yang, Y. Xu, and Z. Hu: Metall. Mater.

Trans. A, 2001, vol. 32A, pp. 2947-57.17. R.D. Field, T.M. Pollock, and W.H. Murphy: Superalloys 1992, 1992,

pp. 57-66.18. A. Lasalmonie and J.L. Strudel: Phil. Mag., 1976, vol. 32(5), pp. 937-49.19. M.V. Natahl, R.A. Mackay, and R.V. Miner: Metall. Trans. A, 1989,

vol. 20A, pp. 133-41.20. Y.F. Gu, Y. Yamabe-Mitarai, S. Nakazawa, and H. Harada: Scripta

Mater., 2002, vol. 46, pp. 137-42.21. Y. Yamabe-Mitarai, S. Nakazawa, and H. Harada: Material Science

Forum, in press.22. Y.F. Gu, Y. Yamabe-Mitarai, Y. Ro, and H. Harada: Key Eng. Mater.,

2000, vols. 171–174, pp. 669-76.

16_02-101A-Symp.qxd 11/9/03 4:14 PM Page 2215