International Journal of Refractory Metals & Hard Materials 24 (2006) 292–297www.elsevier.com/locate/ijrmhm

Creep and tensile tests on refractory metals at extremelyhigh temperatures

Bernd Fischer a, Stefan Vorberg a, Rainer Völkl b,Manuel Beschliesser c,¤, Andreas HoVmann c

a FH Jena—University of Applied Sciences, Jena, Germanyb University Bayreuth, Germany

c PLANSEE Aktiengesellschaft, Technology Centre, 6600 Reutte, Austria

Received 10 August 2005; accepted 29 October 2005

Abstract

The need for mechanical properties at elevated temperatures is high for Wnite element modelling, process optimization, research anddevelopment or quality assurance purposes. Obtaining of this data is diYcult, for refractory materials such as molybdenum or tungstenreliable data including precise strain measurement is required up to 2500 °C.

Over the last years a cooperation between PLANSEE and the University of Applied Sciences, Jena, Germany was built up. Within theinternal research program “Basisdaten” (basic materials data) creep data and mechanical properties of molybdenum, tungsten and theiralloys at very high temperatures could be achieved.

After a description of the unique test equipment at the University of Applied Sciences, Jena, the results of creep and tensile tests onmolybdenum and tungsten sheet material are presented.© 2005 Elsevier Ltd. All rights reserved.

Keywords: Mechanical testing; Non-contact strain measurement; Tensile strength; Creep strength

1. Introduction

In the Weld of high-temperature applications, an increas-ing demand can be observed for the use of refractory metalsas engineering materials. Therefore, a reliable determina-tion of the high-temperature mechanical properties is essen-tial. The high-temperature mechanical properties arerequired for the development of optimized alloys and pro-cesses, speciWcations in design engineering, modelling ofcomponent performance in industrial application and forquality assurance in production.

In many cases no data for very high temperatures isavailable and calculations and constructions are performedon the basis of experience and knowledge. With that no reli-

* Corresponding author. Tel.: +43 5672 600 2766; fax: +43 5672 600 536.E-mail address: manuel.beschliesser@plansee.com (M. Beschliesser).

0263-4368/$ - see front matter © 2005 Elsevier Ltd. All rights reserved.doi:10.1016/j.ijrmhm.2005.10.011

ability and safety calculations can be performed. Forresearch and development activities improvements in prop-erties cannot be captured on a quantitative basis.

Because no commercial measuring system exists forcreep and tensile tests conducted on metals at temperaturesup to 3000 °C, special devices were developed and con-structed at the University of Applied Sciences in Jena, Ger-many. The samples are heated directly by an electriccurrent. Temperature measurement and control areachieved contact-free by a pyrometer and a PID-controllerwith a temperature tolerance of §5 K. The strain measure-ment is also accomplished contact-free by a high resolutionCCD-camera. Obtained data then are analyzed by theimage processing program “SuperCreep” which has alsobeen developed at the University of Applied Sciences.Results gained from these devices have been transferred toindustrial partners over the last years which gave proof ofthe reliability of this testing equipment.

B. Fischer et al. / International Journal of Refractory Metals & Hard Materials 24 (2006) 292–297 293

In cooperation with PLANSEE uniaxial creep and ten-sile tests on sheet material of molybdenum, tungsten andtheir alloys which were taken from the commercial produc-tion line of PLANSEE have been performed.

The Wrst part of this paper covers tensile tests of tung-sten sheet material (thickness 1 mm). These tests shouldgive answer to the question whether the results obtained inJena Wt to the results obtained at PLANSEE. Therefore,tensile tests on a sheet of the same batch were performed inJena (testing temperature 1800–2500 °C) and in the TestingLaboratories in the Technology Centre of PLANSEE (test-ing temperature 1400–2100 °C).

The second part investigates the creep performance ofMo and ML (molybdenum based material, doped withLa2O3 particles) sheet material (thickness 2 mm) at 1400 °Cand 1600 °C up to 200 h.

2. Experimental

2.1. General design of test facilities at the University of Applied Sciences, Jena

The interest of the research group in Jena is focused onmetallic materials for ultra-high-temperature applications.Envisaged test temperatures range up to 3000 °C for Re/Walloys [1]. Commercial test facilities either do not have therequired speciWcations or are too expensive. This is the rea-son why these test facilities [2–6] were designed and built atthe University of Applied Sciences, Jena.

Ohmic heating was chosen for easy access to the sample,fast heating and cooling cycles, and simplicity in design andoperation. Low temperatures at the grips allow the use ofinexpensive copper without the need for an active coolingsystem. Usually specimens in the form of thin wires orstrips are tested. The test facilities permit tests either in airor under a protective gas atmosphere. All functions arecomputer controlled with the software LabView and Super-Creep [4–6], the later developed at the University ofApplied Sciences, Jena for strain measurements by meansof digital image analysis.

Both constant tensile load creep tests and high-tempera-ture tensile tests can be executed. For constant engineeringstress creep tests, the load is applied to the sample througha steel pull rod by means of calibrated weights. Alterna-tively, the specimen chamber can be mounted in a commer-cial test machine. The steel pull rod is then connected withthe load cell at the crosshead of the test machine. A sche-matic diagram of the test facilities is given in Fig. 1.

The temperature is measured with a pyrometer. Anadjustable response time down to 1 ms guarantees securetemperature control at high heating rates. A problem oftenencountered in pyrometry is that the spectral emissivity ofthe investigated material has to be known as a function oftemperature, time and wavelength. Pt/Rh alloys show veryslow oxidation at low temperatures. At temperatures aboveabout 1000 °C their oxide scales evaporate. Neuer et al. [7,8]therefore recommend Pt/Rh alloys as reference materials.

Hence simple calibration of the test system for materialswith unknown emissivity can be performed with a thin foilof a Pt/Rh alloy pasted on the specimen.

Strain is measured with a video extensometer controlledby the software SuperCreep. A variable exposure time ofthe CCD camera from 1 to 1000 ms allows images to begrabbed up to 2000 °C without introducing Wlters in theoptical path. Telecentric lenses are used to avoid perspec-tive distortions. SuperCreep continuously determines thedistances between corresponding markers in the centralzone of the specimen where the temperature is uniform.Suitable markers for high temperature tests can be made bylaser cutting samples with small shoulders from the sheetmaterial (Fig. 2a), or simply by winding thin wires aroundthe specimen (Fig. 2b).

2.2. Performance of test facilities

The temperature distribution was measured by a secondpyrometer on a 100 mm long strip of a Pt–10% Rh alloy [5].The maximum temperature was 1500 °C at the specimencentre. In a zone 30 mm around the specimen centre thetemperature was 1500§5 °C. In a zone 10 mm between themarkers the temperature was within 1500§ 2 °C. During afollowing creep test the temperature between the markerscould be held to 1500§ 3 °C until necking occurred. Due toohmic heating the temperature outside a necked region isalways lower, whereas in the necked region the temperatureis kept at the desired value.

The test facility allows fast heating and cooling cycles. Amaximum temperature of 1506 °C at the centre of the Pt–10% Rh DPH strip was reached 12 s after turning on power.After 15 s the maximum temperature was held at1500§1 °C. After switching oV the power, the specimentook about 20 s to cool down to 750 °C. Heating and cool-ing rates of +100 °C/s and ¡30 °C/s, respectively, can bereached [5].

Fig. 1. Test facility to measure tensile and creep properties of metallicmaterials at temperatures up to 3000 °C.

294 B. Fischer et al. / International Journal of Refractory Metals & Hard Materials 24 (2006) 292–297

With SuperCreep and the CCD camera the gauge lengthfrom the digital image can be determined to a subpixelaccuracy of 0.2 pixels. At a maximum measurable strain of�t 60%, a maximum error ��t§0.07% of the strain wasdetermined [6]. The repeatability therefore is less than0.25% of full-range output. The accuracy can be improvedby increasing the initial distance between the markers.However the maximum measurable strain then decreases.

To verify the performance of the video extensometer thethermal expansion of pure Pt was studied [6]. The tempera-ture was increased in Wve steps from 1000 °C up to 1250 °C.Two hundred to three hundred measurements were per-formed at each temperature. The standard deviations���D§(0.04–0.05)%, given in Fig. 3, are in good agreementwith the estimated error of ��t§0.07%. Underlying zerostrain at 1000 °C, Fig. 3 shows that the measured meanthermal strains lie always in between the values reported byBeck [9] and Arblaster [10].

2.3. Experimental details Jena

The measurements on refractory metals were executed inwater-cooled chambers under reducing atmosphere (Ar–5%H2). Samples for creep and tensile testing were produced bylaser cutting and grinding. Specimen geometry was 120£4£ 1 mm (length£width£ thickness). The gauge lengthwas 10 mm. For tensile tests the testing speed was 10 mm/

Fig. 2. Images of self-radiating specimens with markers for the videoextensometer. The initial gauge length deWned by two correspondingmarkers is approximately 10 mm. (a) Pt–10% Rh specimen at 1500 °C withfour small shoulders laser-cut out of sheet. (b) Pt-wires wound round aspecimen as markers.

min which corresponds to an initial strain rate of 2£10¡2 s¡1. The strain was measured in a direct mode with aCCD camera and the Software SuperCreep.

2.4. Experimental details PLANSEE

Tensile tests have been performed in a high-temperaturetesting apparatus in vacuum in the Testing Laboratories inthe Technology Centre of PLANSEE. Samples for tensiletesting were produced by spark erosion and grinding. Spec-imen geometry was 80£ 16£1 mm (length£width£thickness). The gauge length was 20 mm, testing speed wasset to 20 mm/min which corresponds to an initial strain rateof 2£ 10¡2 s¡1. The strain was recorded in a non-directmode via cross-head movement.

2.5. Material

Tungsten (W) has the highest melting point (TmD3420 °C) and lowest vapour pressure (pD 1£ 10¡5 Pa at2500 K) of all metals [11]. Therefore, tungsten is commonlyused in high-vacuum industry and applications at high tem-peratures. Tungsten sheet material for this investigation hasbeen produced using powder metallurgy route (O < 5 �g/g).Final thickness of 1 mm has been achieved by rolling. Ten-sile tests have been conducted on stress-relieved W sheetmaterial.

Pure molybdenum (Mo) exhibits high melting point(TmD2620 °C), low vapour pressure (pD 7£ 10¡2 Pa at2500 K) and high corrosion resistance to molten glass andmetals [11]. For example, it is used in lightning technology,electronics, automotive industry and high-temperature fur-nace industry.

Additions of lanthanum oxide, La2O3 particles (0.7 wt.%)to pure molybdenum (PLANSEE trade name: ML) resultin formation of a layered, Wbrous structure upon recrystalli-

Fig. 3. Measured thermal strain of pure Pt compared to literature dataaccording to Beck [9] and Arblaster [10].

B. Fischer et al. / International Journal of Refractory Metals & Hard Materials 24 (2006) 292–297 295

zation. ML sheet material has excellent creep properties.Examples of application are furnace parts and wires forheating elements. Creep tests have been conducted onstress-relieved Mo and ML sheets which have been rolled to2 mm thickness and further ground to 1 mm (Mo:OD 24�g/g, CD8 �g/g, ML: OD 1035 �g/g, CD6 �g/g).Testing temperature was 1400 °C and 1600 °C which corre-sponds to homologous temperatures T/Tm of 0.58 and 0.65respectively. It can be assumed that steady-state creep pre-dominates at these temperatures.

3. Results and discussion

3.1. Tensile tests

Table 1 lists the number of tested tensile-specimens atthe given testing temperature. Note the overlap of testsbetween 1800 °C and 2100 °C to achieve quantitative judge-ment if a transfer of data recorded on diVerent testingequipments can be done.

Results of the tensile tests conducted on the stress-relieved tungsten sheet (thickness 1 mm, initial strain rate2£10¡2 s¡1) are presented in Tables 2 and 3. Mean valuesare shown in Fig. 4. The elongation to fracture is not listeddue to the fact that the methods of strain-measurement(Jena: direct via SuperCreep, PLANSEE: non-direct viacross-head movement) lead to diVerent i.e. non-comparableresults.

Tensile strength of tungsten sheet material linearlydecreases with increasing temperature in the investigated

Table 2Mechanical properties (tensile strength, Rm and yield strength, Rp0.2) ofstress-relieved tungsten sheet material generated in tensile tests in facilitiesof University of Applied Sciences, Jena

Temperature [°C] Rm [MPa] Rp0.2 [MPa]

1800 103.5 61.51800 105 62.51800 105.8 62.52000 76.5 512000 78 512000 78.2 522100 66 46.52100 66 46.32100 63.5 47.52200 53 38.52200 54.5 412200 54.3 41.82500 25.3 17.52500 24.2 202500 20.5 14.32500 19 14

temperature range. The data recorded in Jena Wt into thedata which have been recorded at PLANSEE. However, theevolution of the yield strength is diVerent although theabsolute deviation is only about 10 MPa. In the overlap-region the data recorded in Jena appears to be higher thanthe one determined at PLANSEE. Since determination of0.2% plastic yield strength involves strain measurement theorigin of error could be found in diVerent types of strainmeasurements at Jena and PLANSEE. It is unlikely thatthe eVect is caused by a combination of deviations in strainmeasurements and the known strain rate sensitivity of theyield strength of body-centred cubic metals since this eVectappears at lower temperatures (approx. <1000 °C for tung-sten) [12,13].

Table 3Mechanical properties (tensile strength, Rm and yield strength, Rp0.2) ofstress-relieved tungsten sheet material generated in tensile tests in facilitiesof Testing Laboratories in the Technology Centre of PLANSEE

a Indicates an experiment, where the test start was delayed although thesample had reached the testing temperature.

Temperature [°C] Proben Rm [MPa] Rp0.2 [MPa] A [%]

1400 3 183 82 41.101600 3 138 57 45.101600 4 141 59 52.801800 3 113 53 54.701800 4 114 51 54.702000 3 80 40 52.302000 4 33 45 62.402100 3 63 32 38.102100 4 42a 30 29.00

Fig. 4. Mean values of mechanical properties of stress-relieved tungstensheet material (thickness 1 mm) between 1400 °C and 2500 °C determinedin uniaxial tensile tests with an initial strain rate of 2 £ 10¡2 s¡1.

Table 1Number of tests performed on stress-relieved tungsten sheet material (thickness 1 mm) in Jena and PLANSEE in the temperature range of 1400–2500 °C

Tester Testing speed [mm/min] Strain rate [s¡1] 1400 °C 1600 °C 1800 °C 2000 °C 2100 °C 2200 °C 2500 °C

PLANSEE 20 2 £ 10¡2 1 2 2 2 2Jena 10 2 £ 10¡2 3 3 3 3 4

296 B. Fischer et al. / International Journal of Refractory Metals & Hard Materials 24 (2006) 292–297

3.2. Creep tests

Output data of creep tests are simple strain–time curveswhich can be further transferred into stress–rupture curves,deformation–time curves and creep rate–stress curves (i.e.Norton plots).

The results of the experiments on pure molybdenum andML sheet material are shown in Figs. 5 and 6.

The results show that the addition of obstructions to dis-location movement as realized by adding La2O3 particles tomolybdenum (ML) has a beneWcial eVect on both the creeprupture behavior and the minimum creep rate. For exam-ple, creep stress for life to rupture of 10 h (see Fig. 5)increases from about 32 to 89 MPa at 1400 °C and from 15to 43 MPa at 1600 °C (corresponds to an increase by a fac-tor of 2.8 in both cases). However, it has to be stated that aquantitative judgement and description of the creep curves

Fig. 5. Creep rupture plot of Mo and ML (stress-relieved) sheet material(thickness 2 mm) tested under constant load conditions in ArH2 atmo-sphere at 1400 °C and 1600 °C.

Fig. 6. Norton plot of Mo and ML (stress-relieved) sheet material (thick-ness 2 mm) tested under constant load conditions in ArH2 atmosphere at1400 °C and 1600 °C.

of Mo and ML is diYcult since parameters such as (sub-)grain size and dislocation density in the starting conditionof Mo and ML are diVerent. The Mo sheet material under-goes recrystallization during the creep experiment, more-over, grain growth is likely to occur since no second phaseis present. For ML sheet material, recrystallization isdelayed due the presence of La2O3 particles.

The creep stress sensitivity of the minimum creep rate(Norton exponent n, see Fig. 6) is of the same order for Moand MLR at 1400 °C and 1600 °C respectively. A Nortonexponent between 3 and 8 describes that the creep mecha-nisms is mainly controlled by dislocation movement(power-law creep). This means that the deformation mech-anism is of the same nature in the investigated materials.However, the stress to activate the creep deformation ismuch higher for ML sheet material. For example, stress toobtain a minimum creep rate of 5£ 10¡8 s¡1 is 10 MPa forMo and 40 MPa for ML sheet material.

4. Conclusions

The investigation revealed that the test equipment at theUniversity of Applied Sciences, Jena is highly capable ofgenerating creep and tensile testing data of molybdenumand tungsten sheet material up to high temperatures.

The results of uniaxial tensile tests on tungsten sheetmaterial (thickness 1 mm) are in good agreement with datarecorded in the Testing Laboratories in the TechnologyCentre of PLANSEE. With the aid of the high-temperaturetesting equipment mechanical data of tungsten sheet mate-rial could be generated up to 2500 °C.

The results of creep experiments on molybdenum andLa2O3 particles doped molybdenum sheet material (thick-ness 2 mm) at 1400 °C and 1600 °C showed the beneWcialeVect of particles on both creep rupture strength and mini-mum creep rate.

Acknowledgment

Financial support came from PLANSEE’s research pro-ject “Basisdaten” which is installed to gain materials dataof PLANSEE products in diVerent product forms.

References

[1] Fischer B, Freund D, Lupton DF. Stress–rupture strength of rheniumat very high temperatures. In: Proceedings of the international sym-posium on rhenium and rhenium alloys, TMS annual meeting,Orlando, 1997. p. 311–20.

[2] Fischer B, Helmich R, Töpfer H. Anordnung zur Warmfestigkeitsprü-fung hochschmelzender, elektrisch leitender WerkstoVe. DD Patent245576 A3, May 1982.

[3] Fischer B, Töpfer H, Helmich R. Gerät für Warmfestigkeitsmessun-gen an hochschmelzenden Metallen nach Einwirkung silikatischerSchmelzen. Silikattechnik 1984;35(11):329–31.

[4] Völkl R, Freund D, Fischer B, Gohlke D. BerührungsloseDehnungsaufnahme an widerstandsbeheizten Metallzugproben mitHilfe digitaler Bildverarbeitung bei Prüftemperaturen bis 3000 °C. In:Proceedings of the conference WerkstoVprüfung 1998, Bad Nauheim,

B. Fischer et al. / International Journal of Refractory Metals & Hard Materials 24 (2006) 292–297 297

Germany, Deutscher Verband für Materialforschung und -prüfunge.V., Berlin, Germany, WerkstoVprüfung Berichtsbände, ISSN 0941-5300, No. 638, 1998. p. 211–8.

[5] Völkl R, Freund D, Fischer B. Economic creep testing of ultra-hightemperature alloys. J Testing Eval 2003;31(1):35–43.

[6] Völkl R, Fischer B. Mechanical testing of ultra-high temperaturealloys. Exp Mech 2004;44(2):121–8.

[7] Neuer G, Pohlmann P, Schreiber E. Gesamtemissionsgrad und spe-ktraler Emissionsgrad von Hochtemperaturmaterialien. TechnicalReport IKE 5-249, Stuttgart, 1998. ISSN 0173-6892.

[8] Neuer G, Jaroma-Weiland G. Spectral and total emissivity of hightemperature materials. Int J Thermophys 1998;19(3):917–29.

[9] Beck G. In: Degussa AG, editor. Edelmetall-Taschenbuch. Hüthig,Heidelberg, Frankfurt Main, 1995. ISBN 3-7785-2448-8.

[10] Arblaster JW. Crystallographic properties of platinum. PlatinumMetals Rev 1997;41(1):12–21.

[11] Martienssen W, Warlimont H. Springer handbook of condensed mat-ter and materials data. Berlin, Heidelberg: Springer-Verlag; 2005.

[12] Stickler Ch. Mikroplastizität und zyklisches Spannungs-Dehnungs-verhalten von Ta und Mo bei Temperaturen unter 0,2 Tm, Wien,1998.

[13] L. Hollang, Fließspannung und Verfestigung hochreiner Molybdän-Einkristalle. Stuttgart, 1996.