Constitutional and microstructural investigation of the pseudobinary NiAl–W system

lable at ScienceDirect

Intermetallics 19 (2011) 342e349

Contents lists avai

Intermetallics

journal homepage: www.elsevier .com/locate/ intermet

Constitutional and microstructural investigation of the pseudobinary NiAleWsystem

Srdjan Milenkovic a,b,*, André Schneider a,c, Georg Frommeyer a

aMax-Planck-Institut für Eisenforschung, Max-Planck-Str. 1, D-40237 Düsseldorf, GermanybMadrid Institute for Advanced Studies of Materials (IMDEA Materials Institute), C/Profesor Aranguren s/n, 28040 Madrid, SpaincBenteler Tube GmbH, Residenzstraße 1, D-33104 Paderborn, Germany

a r t i c l e i n f o

Article history:Received 16 August 2010Received in revised form18 October 2010Accepted 20 October 2010Available online 20 November 2010

Keywords:A. Nickel aluminides, based on NiAlB. CrystallographyB. Phase diagramsC. Crystal growthD. Microstructure

* Corresponding author. Present address: Madrid Inof Materials (IMDEA Materials Institute), C/Profesor ASpain.

E-mail address: [email protected] (S. Milenko

0966-9795/$ e see front matter � 2010 Elsevier Ltd.doi:10.1016/j.intermet.2010.10.019

a b s t r a c t

Pseudobinary NiAleW section in the range 0.7e1.8 at% was characterised regarding solidificationmicrostructure and constitution. The existence of eutectic reaction was confirmed and occurs attemperature of 1664 � 2 �C and the composition of 1.5 at% W. As-cast microstructures indicated thatthe NiAleW system is an anomalous eutectic with skewed coupled zone and that the NiAl phase isa preferential phase for eutectic nucleation. Directionally solidified alloy containing 1.5 at% W exhibitedentirely eutectic structure characterised by eutectic cells with the average value of 500 mm, interfibrespacing 3.5 mm, fibre diameter 300 nm, and volume fraction of theW phase 1.4%. The solubility of W in theintermetallic phase is <0.04 at% and the solubility of Ni and Al in the W phase is negligible. Crystallo-graphic orientation between the constituent phases was established to be cube on cube. Based on theresults obtained by DTA, microstructural and compositional analysis, a new isoplethal section NiAleWofthe ternary AleNieW system has been proposed.

� 2010 Elsevier Ltd. All rights reserved.

1. Introduction

Rapid advances in high temperature technology dictate ever-growing need for new materials with superior properties atelevated temperatures. Over the past decade ordered intermetallicshave received considerable attention as potential candidates forthe high temperature structural applications. Major efforts weredirected to the improvement of low temperature ductility and hightemperature strength. Among a large number of the intermetallicswith attractive properties, a particular interest has been devoted tothe B2-ordered intermetallic compound NiAl and its alloys. In termof thermophysical properties, B2 NiAl offers several advantagesover the currently used nickel-based superalloys. It has a highermelting point (1674 �C) [1], a significantly lower density(5.86 g/cm3), and distinctly higher thermal conductivity (76W/mK)at ambient temperature [2]. Furthermore, NiAl possesses excellentoxidation resistance at high temperatures, which can be furtherimproved by alloying with yttrium and other refractory elementslike zirconium and hafnium. However, the structural use of NiAl

stitute for Advanced Studiesranguren s/n, 28040 Madrid,

vic).

All rights reserved.

is hindered by few major drawbacks: poor ductility and fractureresistance at ambient temperatures and low strength and creepresistance at elevated temperatures. In order to achieve the desir-able balance of properties, several approaches have been consid-ered. A special emphasis has been devoted to the incorporation ofrefractory metal or intermetallic (Heusler or Laves) phases throughdirectional solidification (DS) of eutectic alloys. DS eutectics canbe classified as natural in-situ composites since their structureconsisting of two ormore separate solid phases is formed in a singlestep. The advantages DS eutectics have over single-phase inter-metallics include an improvement in both strength and toughnessby various intrinsic and extrinsic mechanisms, as well as naturalchemical and mechanical compatibility between the reinforcementand the matrix, which is extremely important feature for hightemperature applications in hostile environments for extendedperiods of time.

It is known that in all NieAleX phase diagrams, where X isa refractory metal, there is a large two-phase field between B2 NiAlphase and either a terminal solid solution of refractory phase (Cr,Mo, Re, V, W) or intermetallic phase of the Heusler or Laves type(Hf, Nb, Ti, Ta). Various studies confirmed that pseudobinaryNiAleX sections are all of eutectic type. The most extensivelystudied systems are NiAleCr, NiAleMo and NiAleCr(Mo) [3e17].While the systems with Ta, Nb, Re and V were also investigated[18e21], although to a lesser extent, there is very limited

S. Milenkovic et al. / Intermetallics 19 (2011) 342e349 343

information regarding the NiAleWsystem. In the first investigationof the Ni-rich corner of the AleNieW system employing thermaland microstructural analysis, Budberg [22] was first to derive theNiAleW isopleth and three isothermal sections at 800, 1000and 1200 �C. It was reported that eutectic reaction occurs at thecomposition of 13.4 at% W and temperature of 1600 �C. The solu-bility of tungsten in NiAl was determined to be 10 at% at theeutectic temperature and 6 at% at 1500 �C. In contrast, Kaufman[23] and Nash [24,25] obtained through thermodynamic calcula-tions and EPMA measurements, respectively, the solubility of w0.2at% W at 1250 �C. Extrapolation of Budberg’s values to 1250 �Cyields w2 at% W, which is ten times higher than the latter. Thedifficulties involved in achieving equilibrium in this system suggestthat more reliance should be placed on the results of Kaufmanand Nash, particularly in view of the limited experimental tech-niques used by Budberg. Thus, Alekseeva [26] corrected the partialpseudobinary NiAleW section to account for the results of Kauf-man and Nash, obtaining the new composition of the eutecticreaction ofw1.4 at%W. In addition, Schäffer [27] studied NiAl alloysprocessed in a Tamman furnace, with 0.25e3.0 at% W and set theeutectic point at 1640 �C and 0.9 at% W. Furthermore, Tiwari et al.examined NiAl alloys containing 1 at%W in hot extruded and zone-directionally solidified alloys [28,29]. They observed a two-phasealigned microstructure, which consisted of primary NiAl grainssurrounded by the NiAleW eutectic. More recently, Popovic et al.[30] obtained new experimental data on the phase equilibria inNi-rich alloys at 1100, 1000, and 900 �C and computed threeisothermal sections at these temperatures. Combining their newresults with literature data, they optimized the thermodynamicinteraction parameters using the CALPHAD approach.

Considering the discrepancies in the literature, this work wasundertaken in order to investigate constitution and microstructureof NiAleW pseudobinary section. The main objective is to deter-mine with accuracy eutectic composition and temperature, as wellas to examine the effect of the composition and the growthconditions on the microstructure and morphology of the NiAleWeutectic alloys.

2. Experiments

As a startingmaterial, nickel (99.97wt.%), electrolytic aluminium(99.9999wt.%) and tungstenpellets (99.9wt.%)wereused topreparethe alloys. The compositions of all the alloys were close to thenominal values for themajor elementswith the deviations less than0.5 at% for Ni, 0.8 at% for Al, and 0.1 at% for W, indicating negligiblelosses during processing. The impurity content was 550 ppm Fe,100 ppm O, <100 ppm Si, 65 ppm C, <10 ppm N and <2 ppm S onaverage. Pre-alloys were prepared by induction melting under inertatmosphere and drop casting into a cylindrical copper mould.Subsequently, the as-cast ingots were machined to fit aluminacrucibles and directionally solidified in a Bridgman type crystalgrowing facility. The experiments were conducted at the tempera-ture of 1700�10 �C, thermal gradient of approximately 40K/cmandgrowth rate of 30 mm/h.

Directionally solidified ingots were prepared by grinding,mechanical polishing and chemical etching with a solution con-sisting of 68 vol% glycerine, 16 vol% HNO3 and 16 vol% HF for furthermetallographic analysis. The microstructural analysis includedcharacterisation using optical and scanning electron microscopy(SEM). The scanning electron microscopy was performed ona Hitachi S-530 SEM system with integrated EDAX DX-4 Detectorfor quantitative (EDX) energy diffraction x-ray analysis on 1 mmdiamond polished samples. In addition, the Automatic CrystalOrientation Mapping e ACOM technique (also known as orienta-tion imaging microscopy, OIM) was used [31,32]. It is based on the

automatic evaluation of electron backscatter diffraction (EBSD)patterns, which are collected by stepwise scanning of the electronbeam on a regular grid over the sample in an SEM. The analysis ofan individual EBSD pattern yields first of all the crystallographicorientation of the crystal at the position of the electron beam. In thepresent study ACOM has been applied in a high-resolution, high-intensity SEMwith field emission gun (JEOL JSM 6500 F). For ACOMa high-speed CCD camera (DigiView) for pattern recording andthe TSL OIM analysis software was used.

Electronprobemicroanalysis (EPMA)wasperformedwithCamecaSX 50 and SX 100 instruments. Pure Ni, Al and W were used asstandards, and the error of the resulting compositions is�1% relative.The previous EPMAvalues were finally calibrated by a comparison ofvalues obtained for single-phase alloys with respective valuesresulting from ICP-OES. For the determination of the phase compo-sitions, at least 15 points were measured per phase. Very fine-scaledeutectic regions were analysed with a defocused electron beamcovering areas of up to 10 mm in diameter with one analysis.

DTAmeasurements were performed on a Setaram Setoys 18 DTAsystem with high temperature measuring head. Directionallysolidified samples of 3 mm diameter and 2 mm length were heatedunder argon in alumina crucibles up to the melting. The heatingand cooling rates were 10 K/min. The accuracy of the temperaturedeterminationwas DT¼ 1 �C. All temperatures were determined bylinear regression of the slope during the heat flow to the trans-formation free curve. For more detailed experimental informationsee [33].

3. Results and discussion

3.1. As-cast microstructures

Initially, as-cast ingots were analysed in order to study solidifi-cation behaviour and growth morphology of the eutectic andoff-eutectic alloys in the NiAleW eutectic system. Since literaturesurvey indicated that the eutectic reaction occurs at the composi-tion of 1.2 � 0.3 at% W samples within the range from 0.7 to 1.8 at% W were prepared and analysed.

In all the samples a coarse grain two-phase microstructurewas observed. Depending on the composition, the microstructureconsisted either of NiAl dendrites and eutectic, as in this case ofalloys with 0.7 and 0.9 at% W (Fig. 1a), or of W dendrites (lightcircles), NiAl enriched layers, so called halos, around them (greyareas) and eutectic (Fig. 1b and c), as in the case of the alloys with1.2e1.8 at% W. With increasing content of tungsten, the fraction ofits dendrites increased. However, the fraction of eutectic structuredid not change significantly with the composition. The formationof both primary phases and eutectic structures is characteristic fornon-equilibrium solidification of the so called anomalous eutectics,in which the coupled zone, a range on the phase diagram definingcompositions and temperatures at which the two eutectic phasescan grow with similar velocities, is skewed to one side of thephase diagram. The consequence is that at high undercoolings themicrostructure of an alloy with eutectic composition may not befully eutectic, which unable the estimation of the eutectic compo-sition from the as-cast microstructures. On the other hand, theskewed coupled zone enables production of fully eutectic struc-tures from the off-eutectic alloys at certain conditions, allowingthus to change the volume fraction of the constituent phases, whichis otherwise fixed by the phase diagram. SinceWhas amuch highermelting point than NiAl and the eutectic composition lies muchcloser to NiAl than toW, the assumption about the shape of coupledzone is in agreement with the observed microstructures.

The formation of a halo of one phase around a primary dendriteof another phase is a common occurrence during the solidification

Fig. 1. As-cast microstructures: a) NiAle0.7W, b) NiAle1.2W, c) NiAle1.8W. Right pictures are the higher magnification of the pictures on the left.

S. Milenkovic et al. / Intermetallics 19 (2011) 342e349344

of off-eutectic alloys, but it may also occur in sufficiently under-cooled eutectic alloys. Halo formation has been the subject ofa number of investigations [34e38] and the conditions requiredfor its formation are reasonably well understood [37]. Whenprimary a dendrites nucleate and start to grow at some under-cooling below the equilibrium liquidus temperature, the nucleationof the eutectic phases can be either hindered or promoteddepending on the capacity of primary phase to serve as a nucleatingagent. The ability of one solid to heterogeneously nucleate anothersolid has been extensively studied, particularly in eutectic systems.

Most studies have indicated that eutectic alloys exhibit non-reciprocal nucleating characteristics, meaning that one primaryphase acts as an effective heterogeneous nucleation site for theother phase, but not vice versa [34e38]. For some eutectic alloysa primary phase is so poor nucleant, that the surrounding meltbecomes severely undercooled and solute enriched leading todendritic growth of undercooled phase around the primary phaseprior to eutectic growth [37]. Thus, the above mentioned micro-structural feature indicates that in this system the NiAl phase isa preferential phase for eutectic nucleation.

S. Milenkovic et al. / Intermetallics 19 (2011) 342e349 345

Themicrostructural investigation of the as-cast samples enabledthe analysis of the growth characteristics of the constituent phases.According to the theory of crystal growth [39], the growth mode ofthe constituent phases determines directly the resultant micro-structure morphology. In an early work on single-phase materials,Jackson [40] proposed an interface-roughness parameter, a, todistinguish between ‘non-faceted’ phases that solidify withcomplete crystallographic isotropy via an atomically rough solid/liquid interface, and ‘faceted’ phases that have preferred crystallo-graphic growth directions associated with atomically smooth solid/liquid surface facets. The interface-roughnessparameterwas relatedto the entropy of fusion (DSf) by the expression ay ADSf/,R where Ris the gas constant and A is a crystallographic parameter (z1).According to this criterion, if a < 2, the solid/liquid interface isatomically rough and non-faceted interface results. On the otherhand, in the case of a > 2, the resulting solid/liquid interface issmooth and a faceted growth occurs.

The observation of the microstructures, where the primaryphases grow concurrently with the eutectic microstructure, allowsone to evaluate the solidification behaviour and the crystallo-graphic nature of the primary phases. Fig. 2a and b show thesolidification microstructure with the formation of primary phases,encompassed by the eutectic microstructure. It can be observedthat the primary dendrites of both phases grow totally isotropically,and that does not exist the preferential growth direction. This

Fig. 2. Growth morphology of primary dendrites: a) NiAl, b) W.

indicates that both aW and bNiAl phases have tendency to non-faceted growth. Table 1 shows thermodynamic parameters usedto calculate parameter a. Data for NiAl were taken from ref. [41].

It can be seen that microstructural observations are consistentregarding the aW phase, but it is rather intriguing that an inter-metallic phase with the ordered structure and high entropy offusion exhibit non-faceted nature. One plausible explanation couldbe that in the mentioned criterion only entropies of the purecomponents were used, without taking into account the solidsolubilities of the eutectic phases. However, this phenomenonhas already been observed during solidification of some eutecticscontaining intermetallic phases and was discussed by Elliot [42].

3.2. Directionally solidified microstructures

Since as-cast microstructures did not allow estimating eutecticcomposition, microstructural analysis was performed on samplesprocessed by directional solidification, which yielded more ther-modynamically stable microstructures due to much lower under-cooling and unidirectional heat extraction. It showed someinteresting results regardingmicrostructure evolution,morphology,uniformity and interphase spacing.

Transverse sections cut from the bottom, middle and top of eachingot revealed variations in the microstructure depending on thecomposition (Fig. 3). Alloys with 0.7 and 0.9 at% W presentedprimary dendrites of the bNiAl along the whole ingot (Fig. 3a),indicating that the composition was off-eutectic and/or that thegrowth occurred out of the coupled zone. Interdendritic regionexhibited eutectic structure with fibrous morphology (Fig. 3a). Onthe other hand, alloy with 1.2 at% W exhibited primary phasedendrites only in the bottom section, while at the central andtop sections fully eutectic structure was observed (Fig. 3b). Unlikeother alloys, only the alloy containing 1.5 at% W exhibited entirelyeutectic structure without any traces of primary phase dendritesalong the whole ingot, as depicted in Fig. 3c. Finally, the micro-structure of the alloy with 1.8 at%W consisted of primary dendritesof the aW phase and eutectic at the bottom section, whereasmiddle and top sections only of eutectic structure (Fig. 3d). At thelongitudinal section, the microstructure was quite similar, charac-terised by the primary phase dendrites and/or eutectic micro-structure with fibrous morphology. The cells were severalmillimetres in length along the growth direction. In the interiorof the cells the orientation of the fibres was not always parallel tothe growth direction and they were often inclined or perpendicularto the growth axis of the adjacent cells. According to solidificationtheory, when an off-eutectic alloy reaches a liquidus, primarydendrites nucleate and start to grow, while the composition ofliquid phase changes along the liquidus line until it reaches eutecticcomposition. Then the solidification process is completed andresulting microstructure consists of primary dendrites and eutecticmicrostructure. Applying this theory to above microstructuralobservations, it follows that the alloys with 0.7, 0.9 and 1.2 at% Ware hypoeutectic, alloy with 1.8 at% W hypereutectic, whereas alloywith 1.5 at% W fully eutectic, indicating that this is the eutecticcomposition.

The main feature of eutectic microstructure was the formationof eutectic cells with fine fibrous morphology inside the colonyand the coarse regions at the colony boundaries, as depicted in

Table 1Thermodynamic parameters used to calculate parameter a.

Phase DHf (KJ/mol) T (K) DSf (J/mol K) a

aW 35 3695 9.47 1.14bNiAl 58.8 1923 30.6 3.67

Fig. 3. DS microstructures from the bottom (left), centre (middle) and top (right) of the ingot with: a) 0.7 and 0.9, b) 1.2, c) 1.5W and d) 1.8 at% W.

S. Milenkovic et al. / Intermetallics 19 (2011) 342e349346

S. Milenkovic et al. / Intermetallics 19 (2011) 342e349 347

Fig. 4a. The average cell size was dependent on the compositionand for the off-eutectic alloys (NiAle1.2W and NiAle1.8W) wasw200 mm, and for the eutectic alloy (NiAle1.5W) w500 mm.Comparison of the values measured at the top, middle and bottomof ingots showed that the variation along the length of each ingotwas negligible and within the observed experimental scatter.This observation is significant since it suggests that good micro-structural control should be possible during a typical DS productionrun of this alloy in the manufacture of an airfoil.

The formation of the eutectic colony structure has been asso-ciated to the combined effects of the impurities rejected fromthe solidifying eutectic, the growth rate, and the imposed thermalgradient at the solid/liquid interface, which produce a zone ofconstitutionally undercooled liquid ahead of the growing interface.Under these conditions of growth, the planar solid/liquid interfacebecomes unstable and transforms into cellular interface. Thegrowth characteristics of the colony structure have been investi-gated by several researchers [43,44] and it has been shown thatthe colonies are formed by the motion of a cellular solid/liquidinterface during growth. Further, it was observed that fibres do notgrow parallel to each other within the cells, but diverge towardscolony boundaries and tend to increase in thickness as theyapproach it. The direction inwhich the fibres grow in the cells givesa direct indication of the shape of the growing cellular interface.The formation of cell structure indicated that the applied growthconditions were below critical ones for achieving the planar solid/liquid interface.

Fig. 4. Microstructure of the eutectic cells at a) border and b) interior of the cell.

Interior of the cells consisted of regular fibre arrays with meandiameter of 300 nm and interphase spacing of 3.5 mm as shown inFig. 4b. The volume fraction of the W phase was measured to be1.4%.

3.3. Compositional analysis

The composition of the alloy and the constituent phases hasbeen determined using different methods. Initially, the conven-tional electron microprobe analysis (EMPA) was applied to deter-mine the overall composition of the eutectic aswell as compositionsof the constituent phases. It was found that the composition ofthe eutectic regions is 1.5% indicating that this is the exact eutecticcomposition. The NiAl matrix contained only traces of W (0.02%)and had the near-stoichiometric composition. The size of W fibres(300 nm) was quite below the resolution of this technique andthe composition of that phase could not be measured. Therefore,another approach was used. Recently, an electrochemical approachfor the selective phase separation of the NiAl-based eutectics hasbeen established [45e48]. It has been demonstrated that employingthe proper conditions it is possible to selectively dissolve either thematrix or the fibres without affecting the other phase. Hence, theW fibres were extracted using procedure described in [45] andanalysed by ICP-OES. After disintegration of the wires taken fromthe filter paper, chemical analysis was performed showing residualcontents of 0.17% and 0.38% for Ni and Al, respectively. These valuesare in agreement with an additional TEM study performed on theextracted wires [48]. The EDX-spectra showed no nickel oraluminium from the matrix phase present in the wires, indicatingthat the solubility of the Ni and Al in W is below the detection limitof the EDX which is 1 at%.

By using the same approach, analytical chemistry experimentswere applied to the electrolyte used for selective dissolution. Thecontent of W in the NiAl matrix was 0.04%. A piece of theuntreated alloy was also completely disintegrated in aqua regia todetermine the overall composition of the alloy. The ICP-OESmeasurement showed that the overall composition of the alloy is49.5 at% Ni, 49 at% Al and 1.5 at% W. In summary, the resultsobtained by both techniques matched very well and showed thatthe eutectic composition is 1.5 at% W, the solubility of W in theintermetallic phase is <0.04 at% and the solubility of Ni and Al inthe W phase is negligible. The results of composition analysis areshown in Table 2.

3.4. DTA analysis

In order to accurately determine eutectic temperature, allthe samples were submitted to thermal analysis. The results aresummarized in Table 3. All investigated NiAleWalloys showed onlyone transformation temperature at 1664 �C. Generally, off-eutecticalloys show two peaks, associated with eutectic reaction and liq-uidus temperature, respectively. However, this could not beeobserved in any of the investigated alloys, neither on heating norcooling. The only difference was that the temperatures determinedduring the cooling cycle were lower for about 5 �C, due toundercooling.

Table 2Nominal (bold) and analysed composition of the alloy with 1.5 at% W.

Eutectic NiAl matrix W fibres

Ni Al W Ni Al W Ni Al W

Nominal 49.25 49.25 1.5 50 50 0 0 0 100EPMA 49.55 48.98 1.47 50 49.98 0.02 e e e

ICP-OES 49.51 49.00 1.49 50.2 49.76 0.04 0.17 0.38 99.45

Table 3Solidus temperatures determined by DTA.

Alloy Transformation temperature

Heating Cooling

NiAle0.7W 1662 1658NiAle0.9W 1662 1658NiAle1.2W 1663 1658NiAle1.5W 1664 1659NiAle1.8W 1666 1659

S. Milenkovic et al. / Intermetallics 19 (2011) 342e349348

In the case of hypoeutectic alloys, liquidus and solidus could notbe distinguished because of the overlapping of the peaks, due tovery close transformation temperatures. Considering that themelting point of pure NiAl is 1674 �C and of eutectic 1664 � 2 �C,and that liquidus of hypoeutectic alloys should be in the samerange, it is clear why the peaks could not be distinguished. On theother hand, the liquidus temperature of hypereutectic NiAle1.8Walloy could not be observed, possibly because it has not be reachedduring the experiments, due to very steep slope of the liquidus line.Summarizing, the onsets of the heat flow curves have been inter-preted as the eutectic melting temperatures and the eutectictemperature has been established to be 1664 � 2 �C. The observedtemperature is somewhat higher than reported in the literature,and the discrepancy is most probably related to different solidifi-cation conditions and experimental details. Based on the resultsobtained by DTA, microstructural and compositional analysis,a new isoplethal section has been proposed in Fig. 5.

3.5. Crystallography

An interesting aspect of directionally solidified eutectics isa preferred crystallographic orientation relationship between thephases. The existence of such specific crystallographic relationshipsis ascribed tominimization of interface energy between the phases.The high degree of thermal stability of many eutectics has also beenassociated to interfaces of low boundary energies. ConsideringNiAl-based eutectics, only few were characterised regarding crys-tallographic relationships [6,49e52]. Eutectic alloys that consistof phases with similar structures, such as NiAleCr, Mo, V, oftenadopt identical growth textures and form a unique orientationrelationship between the phases, any lattice parameter mismatchbeing accommodated by structural dislocations at the interfaces.However, no data were reported for the NiAleW system.

In this study orientation imaging microscopy was used todetermine the crystallographic orientations of the two eutectic

Fig. 5. NiAleW isoplethal section of the ternary AleNieW system.

phases (NiAl andW)and their interfacial planes. EBSDpatternswererecorded fromthe individual phases at different locationswithin theas-grown composite. Fig. 6 shows the EBSD patterns from the NiAlmatrix and the W fibres. The growth direction was found to beparallel to<100> directions in both NiAl andW. In addition, in bothphasesplanesparallel to interfaceboundarieswere (110) planes. The(110) habit planes are the close packed planes in the bcc and B2crystal structures of W and NiAl, respectively. As such, they arefrequently occurring as interphase boundaries due to aminimum ininterface energy. The lattice parameters of W and NiAl are 3.16 and2.89 Å, respectively, which gives the lattice mismatch of around 9%.When two crystals have small lattice mismatch, it is possible tomaintain interface coherency by straining one or both lattices. Asthese coherency strains increase the interface energy, for misfitslarger than 5% it becomes energetically more favorable to form

Fig. 6. EBSD patterns obtained from (a) NiAl matrix and (b). W fibres.

S. Milenkovic et al. / Intermetallics 19 (2011) 342e349 349

a semicoherent interface in which the mismatch is periodicallytaken up by misfit dislocations. The distance between the misfitdislocations decreases with the increase in elastic strain. The latticemismatch of around 9% in this case suggests that the interfacebetween the NiAl matrix and W fibres is semicoherent.

4. Conclusions

Solidification microstructure and constitution of the pseudobi-nary NiAleW eutectic system were investigated and from theobtained results following conclusions were drawn:

i. The pseudobinary eutectic NiAleW system exhibits eutectictemperature of 1664 � 2 �C and the eutectic composition of1.5 at% W. Based on the results obtained by DTA, micro-structural and compositional analysis, a new isoplethalsection has been proposed.

ii. As-cast microstructures indicated that the NiAleW system isan anomalous eutectic with skewed coupled zone and thatthe NiAl phase is a preferential phase for eutectic nucleation.Regarding the crystallographic nature of the primary phases,both eutectic phases show tendency to non-faceted growth.

iii. Directionally solidified alloy containing 1.5 at% W exhibitedentirely eutectic structure characterised by eutectic cells withfine fibrous morphology inside the cells and the coarseregions at the cell boundaries.

iv. The average value of the cell size was 500 mm, interfibrespacing 3.5 mm, fibre diameter 300 nm, and volume fractionof the W phase 1.4%. The composition of the matrix is nearstoichiometric with only traces of W (<0.04%). The solubilityof both Ni and Al in W is negligible.

v For the both phases the growth direction is parallel to<100>and the habit plane is (110).

Acknowledgments

Authors wish to thank Ms. Schaff for performing directionalsolidification experiments, Mr. Bialkowski for the microprobeanalysis and Ms. Nelesen for the EBSDmeasurements. The financialsupport of the Max-Planck-Society for the advancement of Scienceis gratefully acknowledged.

References

[1] Rablbauer R, Frommeyer G, Stein F. Mater Sci Eng A 2003;343:301.[2] Darolia R. J Metals 1991;43:44.[3] Walter JL, Cline HE. Metall Trans 1973;4:33.[4] Johnson DR, Joslin SM, Oliver BF, Noebe RD, Whittenberger JD. Conference

Proceedings, Mater. In: Miracle DB, Graves J, Anton DL, editors. Inter-

metallicemetallic polyphase in-situ composites, vol. 273. Pittsburgh, PA:Research Society Symposium Proceedings; 1992. p. 87e92.

[5] Walter JL, Cline HE. Metall Trans 1970;1:1221.[6] Johnson DR, Chen XF, Oliver BF, Noebe RD, Whittenberger JD. Intermetallics

1995;3:99.[7] Jiang DT, Guo JT. Mater Sci Eng A 1998;225:154.[8] Guo JT, Cui CY, Chen YX, Li DX, Ye HQ. Intermetallics 2001;9:287.[9] Heredia FE, He MY, Lucas GE, Evans AG, Dève HE. Acta Metall Mater

1993;41:505.[10] Chen XF, Johnson DR, Noebe RD, Oliver BF. J Mater Res 1995;10:1159.[11] Subramanian PR, Mendiratta MG, Miracle DB. Metall Trans A 1994;25:2769.[12] Misra A, Wu ZL, Kush MT, Gibala R. Mater Sci Eng A 1997;239e240:75.[13] Misra A, Wu ZL, Kush MT, Gibala R. Philos Mag A 1998;78:533.[14] Raj SV, Locci IE. Intermetallics 2001;9:217.[15] Misra A, Gibala R, Noebe RD. Intermetallics 2001;9:971.[16] Cui CY, Guo JT, Ye HQ. Mater Sci Eng A 2004;385:359.[17] Bei H, George EP. Acta Mater 2005;53:69.[18] Zeumert B, Sauthoff G. Intermetallics 1997;5:563.[19] Ferrandini PL, Araujo FLGU, Batista WW, Caram R. J Cryst Growth

2005;275:147.[20] Webber JG, Van Aken DC. Scipta Metall 1989;23:193.[21] Milenkovic S, Coelho AA, Caram R. J Cryst Growth 2000;211:485.[22] Budberg PB. Zh Neorg Khim 1958;3:694.[23] Kaufman L, Nesor H. Can Metall Q 1975;14:221.[24] Nash P, Fielding S, West DRF. Met Sci 1983;17:192.[25] Nash P. Conference Proceedings. In: Koch CC, Liu CT, Stoloff NS, editors. High-

temperature ordered intermetallic alloys I, vol. 39. Pittsburgh, PA: Mater.Research Society Symposium Proceedings; 1985. p. 411e21.

[26] Aleekseva ZM. In: Petzow G, Effenberg G, editors. Ternary alloys: a compre-hensive compendium of evaluated constitutional data and phase diagrams,vol. 8, AleNieTb to AleZneZr. Wiley-VCH; December, 1992. p. 52.

[27] H. J. Schäffer, 1997. Entwicklung und Eigenschaftscharakterisierung hoch-warmfester Werkstoffe mit intermetallischer NiAl-Matrix, Diss., RWTHAachen, VDI Verlag, Düsseldorf.

[28] Tiwari R, Tewari SN, Asthana R, Garg A. J Mater Sci 1995;30:4861.[29] Asthana R, Tiwari R, Tewari SN. Mater Sci Eng A 2002;336:99.[30] Popovic J, Broz P, Bursik J. Intermetallics 2008;16:884.[31] Adams BL, Wright SI, Kunze K. Metall Trans A 1993;24:819.[32] Schwartz AJ, Kumar M, Adams BL, editors. Electron backscatter diffraction in

material science. New York (: Kluwer Acad/Plenum Publishers; 2000.[33] Stein F, Palm M. In: Morris DG, Naka S, Caron P, editors. Intermetallics and

superalloys euromat 1999, Vol. 10. Weinheim: Wiley-VCH; 2000. p. 336e44.[34] Kofler A. J Aust Inst Met 1965;10:132.[35] Hogan LM. J Aust Inst Met 1962;7:188.[36] Sundquist BE, Bruscato R, Mondolfo LF. J Inst Met 1963;91:204.[37] Gigliotti Jr MFX, Colligan GA, Powell GLF. Metall Trans 1970;1:891.[38] Bluni ST, Notis MR, Marder AR. Acta Metall Mater 1995;43:1775.[39] Jackson KA, Hunt JD. Trans Met Soc AIME 1966;236:p.843.[40] Jackson KA. Mechanism of growth of liquid metals and solidification. Cleve-

land, Ohio: American Society for Metals; 1958. p.174.[41] Miracle DB. Acta Metall Mater 1993;41:p.649.[42] Elliott R. Intern Metall Rev 1977;219:p.161.[43] Kraft RW, Albright DL. Trans Met Soc AIME 1961;221:95.[44] Weart HW, Mack DJ. Trans Met Soc AIME 1958;212:664.[45] Hassel AW, Smith AJ, Milenkovic S. Electrochim Acta 2006;52:1799.[46] Cimalla V, Röhlig CC, Pezoldt J, Niebelschütz M, Ambacher O, Brückner K, et al.

J Nanomater; 2008:638947.[47] Milenkovic S, Hassel AW. Phys Status Solidi A 2009;206:495.[48] Milenkovic S, Smith AJ, Hassel AW. Phys Status Solidi A 2010;207:858.[49] Cline HE, Walter JL. Metall Trans 1970;1:2907.[50] Cline HE, Walter JL, Lifshin E, Russell RR. Metall Trans 1971;2:189.[51] Rablbauer R, Fischer R, Schneider A, Frommeyer G. In: Euromat 2001; 2001.

p. 48. Rimini.[52] Chen YX, Cui CY, Guo JT, Li DX. Mater Sci Eng A 2004;373:279.