SCIENGE&ENGINEERINGAMaterials Science and Engineering A251 (1998) 255-261 256E. Y. Iuanou et al./Materials Science and Engineering A251 (1998)255-261 Synthesis of a nanocrystalline W-25 wt.% Re alloy by mechanicalalloying E.Y. Ivanov a, C. Suryanarayana b,*, B.D. Bryskinc Tosoh SMD, Inc., 3600 Gantz Road, Grove City, OH 43123-1895, USA Department of Metallurgical and Materials Engineering, Colorado School of Mines, Golden, CO 80401-1887, USA ° Rhenium Alloys, Inc., 1329 Taylor St., Elyria, OH 44036-0245, USA Received 26 January 1998 Abstract The formation of a nanocrystalline tungsten-rhenium solid solution by mechanical alloying in a SPEX mill using tungstencarbide vial and balls was investigated. The milling process was monitored by X-ray diffraction and electron microscopytechniques. It was shown that mechanical alloying of a W-25 wt.% Re powder mixture resulted in the formation of a W-Re solidsolution with a very small volume fraction of the phase. The grain size of the solid solution phase is in the nanometer range.Sintering of the mechanically alloyed powder resulted in a dense and high-purity product. O 1998 Elsevier Science S.A. All rightsreserved. Keywords: Mechanical alloying, W-Re alloy; Nanocrystals; X-ray diffraction; Electron microscopy 1.Introduction Nanocrystalline materials are polycrystalline materi-als with a grain size of usually<100 nm. They exhibitunusual physical, chemical, and mechanical properties[6], and can be synthesized by many methods, includingmechanical alloying [6,7]. Mechanical alloying (MA) isa powder metallurgy process involving repeated weld-ing, fracturing, and rewelding of powder particles in ahigh-energy ball mill. It has been shown that MA canresult in the formation of stable and metastable phasesincluding solid solutions, intermediate phases, and evenamorphous alloys [8,9]. However, it has been repeatedly ( * Corresponding author . Tel.: + 1 303 2733178 ; fa x: +1 303 2733795; e-mail: csuryana@mines.edu ) ( 0921-5093/98/$19.00 C 19 9 8 Elsevier Science S.A . All rig h ts res e rved.PII S0921-5093(98)00620-0 ) W-Re alloys are synthesized by blending the elementalpowders and subsequent sintering. Since Re is in closeAdditions of rhenium to tungsten result in increased contact with W, occasionally the Re concentration canlow-temperature ductility, improved high-temperature be high leading to the formation of the undesirableostrength and creep resistance as well as higher recrystal- phase. Attempts were made earlier to homogenize W-lization temperatures. Tungsten-rhenium alloys with Re alloys either by ball milling or by coating the W3-5 wt.% Re (all compositions in this paper are ex- particles with Re films. These met with only a limitedpressed in weight percentage unless specified otherwise) success. Bryskin and Carlen [4,5] have discussed theare the most commonly used. However, to increase the effect of the o phase on the formability and producthot workability, avoid the recrystallization embrittle- quality after high-temperature consolidation and ther-ment, and to control the microstructure these alloys momechanical treatment of powder-metallurgy pro-need to be doped with K, Al, or Si. The strength and cessed W-Re alloys. Hence, it would be desirable tohardness of W-Re alloys, beyond about 5-9% Re, obtain a homogeneous solid solution of Re in W toincrease with increasing Re content, and thus W-24- avoid the above-mentioned problems.27% Re alloys are the most desirable with respect tostrength and ductility. Further, these alloys do notrequire any doping. But, beyond the equilibrium maxi-mum solid solubility limit, which has been determinedto be anywhere from 24 to 37% Re depending on thesynthesis/processing method [1-3], precipitation of theo phase occurs during thermomechanical processingand this leads to cracking of the alloys. Generally, observed that MA in a steel vial and with steel grindingmedium introduces Fe contamination which can be asmuch as 20 at.% [10]. Use of a vial and grindingmedium made of the same material as the powder to bemilled can substantially minimize/avoid contaminationof the powder. Ivanov and Wickersham [11] used tung-sten carbide vial and balls and produced nanocrys-talline tungsten powder with a purity of 99.99%. Thepurpose of the present work is to synthesize high-purityW-Re solid solution alloy by MA and study the sinter-ing behavior of the powder so produced. A W-25% Repowder mixture without MA is also sintered (this willbe referred to as the ‘blended’alloy) to compare thesintering behavior and ascertain the advantages of MA. Mechanical alloying of W-Re powder mixtures wasconducted earlier [12]. Even though a single-phase solidsolution was obtained in a W-25% Re powder mixture,the powder was contaminated with about 3500 ppm ofiron. Further, during sintering of the milled powder, itwas observed that the samples presintered at 1700 Kwere only 57% dense and that during sintering at 2600K, the compact swelled revealing a porous structureinside. Extensive grain growth was also reported withthe grain size being much larger than in the ‘blended'alloy. 2. Experimental procedure Tungsten powder (99.999% pure and 10 pm averageparticle size) from H.S. Starck and Re particles (99.99%pure agglomerates with the average size of 30 um) fromRhenium Alloys were used as the starting materials.The blended powder with a nominal composition ofW-25% Re was mechanically alloyed in a standardSPEX 8000 shaker mill using tungsten carbide vial andtungsten carbide balls. A ball-to-powder weight ratio(BPR) of 5:1 was used. To minimize contamination, theloading and unloading of the powders into and fromthe containers was always done inside an argon-filledglove box. Mechanical alloying was conducted for differenttimes and a small quantity of the powder was taken outat regular intervals for microstructural and crystalstructure characterization. Microstructural characteri-zation was done in a scanning electron microscope(SEM) equipped with an energy dispersive X-ray spec-trometer (EDS) and aa Philips CM200 transmissionelectron microscope (TEM) operating at 200 kV. Thecrystal structure of the resulting phases was determinedby X-ray diffraction (XRD) in a Rigaku D2100H X-raypowder diffractometer using Cu K, radiation. Chemicalanalysis of the powders was performed by an atomicabsorption spectrometer for trace elementsand byLECO for carbon, nitrogen, and oxygen. The milled powders were sintered following the tech-nique described in Ref. [13]. In brief, the milled powderwas cold pressed to 12.5 mm dia pellets, presintered at1773 K for 15 min and then sintered at 2623 K in ahydrogen atmosphere for 3 h. The microstructural fea-tures of the sintered and ‘blended’specimens werecharacterized using an SEM and the microhardness wasmeasured with a 100 g load using a Wilson-Tukonhardness tester. 3. Results and discussion 3.1. Microstructure of the mechanically alloyed powder Fig. 1(a) and (b) show the scanning electron micro-graphs of the as-received W and Re powders, respec-tively. The as-received W powder is characterized bywell-formed crystals with an average size of 10 um. Theas-received Re powder exhibits 30 um clusters of fusedparticles with individual sizes of 2-3 um. The MAprocess has significantly reduced the particle size ofboth the W and Re powder particles and also changedtheir shapes. It can be clearly seen that typical rounded (a) 10um (b) 2.5um Fig. 1. Scanning electron micrographs of as-received (a) tungsten and(b) rhenium powder particles. lum Fig. 2. Scanning electron micrograph of the W-25% Re powdermixture milled for 14 h. powder particles with a size of about 1 um are presentin of the W-25% Re powder mixture milled for 14 h(Fig.2). EDS analysis has shown that complete alloyinghas occurred and that pure W or Re particles are nolonger present. The particle size distribution in theas-received W and Re powders as well as the milledpowder mixture, as determined using a Microtrac, isshown in Fig. 3. It can be seen that while the as-re-ceived W and Re particles are about 10 and 30 um insize, respectively, the average size of the W-25% Repowder milled for 14 h is about 1.3 um. Fig. 4(a) showsa TEM micrograph of the powder milled for 14 hshowing that the material is now in a nanocrystallinestate with a grain size of about 10-15 nm. The corre-sponding electron diffraction pattern is shown in Fig.4(b) which confirms that a W(Re) solid solution withb.c.c. structure is obtained. Fig. 5 shows a high-resolu-tion transmission electron micrograph of the W-25%Re powder mixture milled for 14 h, showing well-re-solved lattice fringes of the W(Re) solid solution phase. (a) (b) Fig. 4.(a) Transmission electron micrograph of the W-25% Repowder mixture mechanically alloyed for 14 h showing the formationof a nanocrystalline structure. (b) Electron diffraction pattern fromthe above showing the presence of a W(Re) solid solution with thebcc structure. 3.2. Supersaturated solid solutions The XRD pattern of the blended powder mixtureshowed the presence of reflections from both W (b.c.c.,a=0.3187 nm) and Re (h.c.p., a=0.2761 nm andc=0.4457 nm) [14]. After MA, the intensities of the Wand Re reflections decreased and their peak widthsincreased due to both a reduction in crystallite size andincrease in lattice strain. Both W and Re continued tobe present after milling for 30 min. However, withcontinued MA, the peak heights continued to decreaseand the peak widths increased. Additionally, the peakpositions changed. For example, the Wpeak positionsshifted to higher 20 values and the Re peak positionsshifted to lower 20 values, suggesting that interdiffusionof these elements into each other has occurred leadingto the formation of W(Re) and Re(W) solid solutions.The XRD pattern of the powder milled for 5 h stillcontained a significant amount of Re(W) solid solution. Fig. 5. High-resolution transmission electron micrograph of the W-25% Re powder mixture mechanically alloyed for 14 h. Fig. 6 shows the XRD pattern of the W-25% Repowder mixture milled for 14 h which contains mostlythe W reflections, suggesting the formation of ananocrystalline W(Re) solid solution with a b.c.c. struc-ture and a=0.3151 nm. Note that some minor peaks,for example at 20=44.3 and 64.8° are also present.While the peak at 20=44.3° is due to Re, the peak at20=64.8° is due to the formation of the o phase.Hence, it can be concluded that at this stage the alloypowder consists of W(Re) and very small amounts (lessthan about 5%) of Re(W) solid solution and the ophase. Further milling up to 24 h did not change theconstitution of the powder (except that the Re contentin the W(Re) solid solution is slightly higher than at 14h) and all the three phases continued to be present.However, with increasing milling time, the milled pow-der was picking up more oxygen and so the milling wasstopped at 14 h. Thus, all the characterization work wasdone on the powder milled for 14 h. Table 1 lists the interplanar spacings and latticeparameters of the W(Re) solid solution phase. Thelattice parameter of W decreased with increasing millingtime due to incorporation of Re into the W lattice. Fig. 6. X-ray diffraction pattern of the W-25% Re powder mixturemechanically alloyed for 14 h. Table 1 Interplanar spacings (nm) and lattice parameter of the W(Re) solidsolution obtained by milling the W-25% Re powder for differenttimes Reflection 14 h 24 h 110 0.22315 0.22245 200 0.15767 0.15701 211 0.12862 0.12814 220 0.11143 0.11105 Lattice parameter (nm) 0.3151 0.3139 Average crystallite size (nm) 5-7 5-6 Lattice strain (%) 0.6 The crystallite size of the solid solution phase wasobtained by measuring the full width at half the maxi-mum intensity of the W peaks as a function of thereciprocal space variable (s)=4 sin 0/a where 20 is thediffraction angle and A is the wavelength of the X-radi-ation used. It has been shown that the peak broadeningdue to crystallite size is independent of s while thebroadening due to strain is proportional to s [15]. Thevolume averaged crystallite size and rms strain arecalculated using the standard equations and it wasfound that the crystallite size is only about 5-7 nmindicating that a nanocrystalline supersaturated solidsolution is obtained by MA. The lattice strain in thepowder milled for 14 h is about 0.6%. There is no information available in the literature onthe variation of the lattice parameter of W with Recontent and that of Re with W. But, considering thatthe atomic radius,r of Re is smaller than that of W(rw=0.1408 nm and rRe=0.1375 nm), it is expectedthat the lattice parameter of W decreases with Recontent and that the lattice parameters of Re increasewith the incorporation of W into the lattice. This is whathas been actually observed in the present investigation.The lattice parameter of pure W (a=0.3187 nm) de-creased to 0.3151 nm after milling for 14 h and to 0.3139nm after milling for 24 h for the W(Re) solid solution,suggesting that a greater amount of Re is present in theW(Re) solid solution at longer times of milling. It wasreported earlier [12] that a W(Re) solid solution with alattice parameter of 0.3121 nm was obtained by millingthe W-25% Re powder mixture in a SPEX mill for 10h. There are some significant differences between theresults of our investigation and those reported in ref.[12]. While we used a BPR of 5:1 the earlier investiga-tors used a BPR of 10:1. The higher BPR of 10:1 resultsin higher energy input into the powder and this mighthave led to the formation of a W(Re) solid solutioncontaining 25% Re. Since a higher amount of Re ispresent in the solid solution, the lattice parameter of theb.c.c. solid solution was smaller in their case. The Fig. 7. X-ray diffraction pattern of the W-25% Re powder (a)mechanically alloyed for 14 h and (b) mechanically alloyed for 14 hand then annealed at 1773 K, indicating the presence of a secondphase. longer time required to achieve the formation of a solidsolution in our case is due to the lower BPR of 5:1 usedin our experiments. It should also be mentioned thatour milled powder contained, in addition to the W(Re)solid solution, trace amounts of Re(W) solid solutionand the o phase. This result suggests that the choice ofa correct BPR is critical in achieving the desired consti-tution of the powder. But, the advantage of using alower BPR is that contamination of the powder wouldbe less. The W-Re nanocrystalline solid solution obtainedby MA is metastable and decomposed on annealing.On annealing the milled powder at 1773 K, its XRDpattern showed an increase in peak intensity and adecrease in peak width reflecting the growth in crystal-lite size and decrease in lattice strain. Further, thepowder now consisted of a mixture of W(Re) solidsolution (with a Re content less than in the as-milledpowder), and a second phase (Fig. 7). In view of thevery low intensities of the peaks from the second phase(and consequently very low peak-to-background ra-tios), it has not been possible to unambiguously identifythe nature (crystal structure and lattice parameters) ofthe second phase. But, it is suspected that it could bethe a phase since a clear peak is observed at about20=64.8°. The ratio between the peak intensities of W and thesecond phase in the annealed sample was used todetermine the amount of the precipitated phase. Usingthis procedure on samples sintered at 2623 K, it wasdetermined that the amount of the second phase wasabout 1% in both the “blended’and ‘milled’alloys. 3.3. Microstructure of the sintered samples The microstructure of the W-25% Re samples (bothwith and without MA) sintered at 2623 K were studiedusing an SEM. Some of the milled powders were alsopreannealed at different temperatures before sintering.The density and hardness values of these samples aresummarized in Table 2. Fig. 8 shows the SEM micro-graphs of the W-25% Re samples sintered at 2623 K.One can clearly see that the ‘blended’alloy (withoutMA) has a large grain size of about 40-45 um while themilled powder compact has a grain size of about 25 um.This indicates that significant grain growth has oc-curred during sintering of the milled nanocrystallinepowder. Additionally, significant porosity is present inboth the samples and the porosity is randomly dis-tributed (Fig. 9). It should also be noted that theporosity isISnot exclusively related to thegrainboundaries and that it can be found inside the grainsalso. An important difference between these two alloysis that the pore size is larger in the ‘blended’ alloy,probably related to the larger grain size of the material. Fig. 10 shows the back-scattered SEM micrographsshowing the general microstructural features of thesintered samples. From the contrast in the microstruc-ture it is noted that the ‘blended'alloy exhibits largenon-homogeneity in composition. On the other hand,the milled sample appears much more homogeneous.The compositional uniformity of the sample was evalu-ated using the EDS technique. Since W and Re happento be neighboring elements in the periodic table, theirelectron energies are very close to each other andtherefore there is a strong interference between the Wand Re peaks. Consequently, an accurate quantitativedetermination of the amounts of W and Re has notbeen possible. In the sintered ‘blended'alloy (obtainedfrom the powder without MA), some pure W grainswere detected (labeled A in Fig. 10(a)). Further, thevariation in Re concentration was as much as 40% from Table 2Percentage of theoretical density and hardness of sintered W-25% Re samples under different conditions Condition Cold-pressed Presintered at 1773 K/15 min Sintered at 2623 K Hardness H、 ‘Blended'alloy 50.2 53.5 93.1 376 Mechanically alloyed for 14 h (MA) 57.4 62.9 92.6 432 MA+Annealed at 1273 K 54.5 57.5 89.6 385 MA+Annealed at 1473 K 57.5 64.4 90.4 383 (a) 100pm (b) Fig. 8. Scanning electron micrographs of the W-25% Re samplessintered at 2623 K. (a) ‘blended'(without MA) and (b) milled for 14h. the light to gray areas. On the other hand, in thesintered sample obtained from the milled powder, nopure W or Re grains were detected and the variation inthe Re concentration from region to region was <10%.This analysis clearly indicates that MA has greatlyaided in improving the compositional homogeneity andgrain size refinement of the sample in comparison tothe ‘blended’ alloy. 3.4. Contamination Contamination of the milled powder with gaseousinterstitial elements like oxygen and nitrogen or substi-tutional elements like Fe, Ni, and Cr from the steelmilling container and grinding medium is a frequentlyencountered problem during MA. Chemical analysis ofthe milled powder revealed that the iron content is only55 ppm, the same as in the starting powder material.This should be compared with the value of 3500 ppmreported earlier [12] when the powder was milled in astainless container and with 52100 steel balls. Contami-nation of up to 3000 ppm of Ni, Cr, and Ti was alsoreported in an attritor-milled W-Re powder [16]. Even though milling and handling of the powder wascompletely done under purified argon atmosphere, thesamples milled for 24 h in the present investigationcontained about 5000 ppm of oxygen. But, the oxygencontent in the sintered samples was