The evolution of grain-boundary cracking evaluated throughin situ tensile-creep testing of Udimet alloy 188

C.J. Boehlerta) and S.C. LonganbachDepartment of Chemical Engineering and Materials Science, Michigan State University,East Lansing, Michigan 48824

M. Nowell and S. WrightEDAX-TSL, Inc., Draper, Utah 84020

(Received 27 May 2007; accepted 6 November 2007)

In situ scanning electron microscopy was performed during elevated-temperature(�760 °C) tensile-creep deformation of a face-centered-cubic cobalt-based Udimet 188alloy to characterize the deformation evolution and, in particular, the grainboundary-cracking evolution. In situ electron backscatter diffraction observationscombined with in situ secondary electron imaging indicated that general high-anglegrain boundaries were more susceptible to cracking than low-angle grain boundariesand coincident site-lattice boundaries. The extent of general high-angle grain-boundarycracking increased with increasing creep time. Grain-boundary cracking was alsoobserved throughout subsurface locations as observed for postdeformed samples. Theelectron backscattered diffraction orientation mapping performed during in situtensile-creep deformation proved to be a powerful means for characterizing the surfacedeformation evolution and in particular for quantifying the types of grain boundariesthat preferentially cracked.

I. INTRODUCTION

Recent progress in the development of in situ scanningelectron microscopy (SEM) techniques for understandingthe deformation behavior of materials, combined with thehigh-beam current and stability of a hot Schottkey fieldemission gun (FEG) and the latest camera technology forrapidly obtaining high-quality electron backscattered dif-fraction Kikuchi patterns (EBSPs), now allows for aunique capability for imaging surface deformation. It ispossible to make unprecedented submicron resolvedmeasurements of the local two-dimensional crystal-structure distribution in materials during high-temperaturetensile-creep deformation. This article describes the tech-nique that has been developed to acquire secondary elec-tron (SE) detector images and electron backscattered dif-fraction (EBSD) orientation maps in situ during tensile-creep deformation at temperatures as high as 760 °C.

Udimet 188 (also known as Haynes Alloy No. 188) isa cobalt–nickel–chromium–tungsten commercially avail-able alloy with good creep strength and oxidation re-sistance up to 1093 °C. It has exhibited good fabrica-bility, tensile strength and elongation-to-failure, weld-

ability, low-cycle fatigue resistance, and corrosion resis-tance.1–6 It has useful applications in gas turbines, com-bustors, flame holders, liners, and transition ducts. Acomplete understanding of the physical mechanisms re-sponsible for the elevated-temperature creep behaviorand associated microstructure–property relationships ofUdimet 188 is lacking. Previous data have suggested thata high fraction of low-angle boundaries (LABs) and co-incident site lattice boundaries (CSLBs) significantly en-hances resistance to creep and grain-boundary sliding atelevated temperatures.7 In this work, an in situ combinedtechnique of surface imaging and EBSD orientation map-ping during tensile-creep testing was used to gain anunderstanding of the deformation evolution in Udimet188 and in particular the sequence of grain-boundarycracking.

II. EXPERIMENTAL

The as-received Udimet alloy 188 sheet material wassubjected to the following strain annealing sequence:cold rolled to 10% or 25% deformation, followed by asolution treatment at 1191 °C for 1 h, followed by aircooling. This sequence was repeated four times. Theoriginal thickness of the as-processed sheet was 8.4 mm,while the final thickness of the 10% cold-rolled sheet was5.5 mm and the final thickness of the 25% cold-rolledsheet was 2.8 mm. The chemical composition of the

a)Address all correspondence to this author.e-mail: boehlert@egr.msu.edu

DOI: 10.1557/JMR.2008.0058

J. Mater. Res., Vol. 23, No. 2, Feb 2008 © 2008 Materials Research Society500

as-processed material is listed in Table I. Spatially re-solved EBSD orientation maps were obtained from pol-ished sections to determine orientation relationships be-tween grains and also to determine the grain-boundarycharacter distribution (GBCD).

The GBCD describes the proportions of the differenttypes of grain boundaries as described by the coincidentsite lattice (CSL) model.8,9 In the CSL model, the crystalorientation of either side of a grain boundary is deter-mined and the associated misorientation across the grainboundary is then measured. If the lattices of two adjacentgrains were allowed to interpenetrate, certain combina-tions of orientation relationship between the two latticeswould result in a periodic array of coinciding sites. Aboundary with a high density of coincident points impliesa good fit between the adjacent grains.8 The reciprocaldensity of the CSL points is denoted by �. A twin bound-ary is represented as �3, indicating that every third site atthe boundary is coincident. LABs are defined as thoseboundaries containing misorientations between 2° and15°. In this work, the boundaries with low � values (� �29), and the LABs were together termed “special”boundaries. The general high-angle boundaries (GHABs)were defined as boundaries containing misorientations>15° that exhibited � values >29. Lehockey and Pa-lumbo7 have shown that a high fraction of special bound-aries significantly enhances the resistance to creep ofpure nickel at elevated temperatures. Other studies10–19

have demonstrated significant improvements in a numberof properties, including stress-corrosion cracking, fa-tigue, weldability, and creep, with an increase in the spe-cial boundary fraction. In terms of stress-corrosion crack-ing, Ni, Ni–16Cr–9Fe–xC, and Alloy 600 have shownsignificant improvements with increased fractions of spe-cial boundaries,11,13,16,19 and these findings have beenattributed to both the intrinsic corrosion resistance, andthe resistance to solute segregation and precipitation ex-hibited by special boundaries. In terms of creep resis-tance, it has also been shown that the presence of a highfraction of special boundaries decreases the creep rate ofNi, Ni–16Cr–Fe, Alloy 625, and Alloy 738.7,10,15,17,18 Infact, pure Ni microstructures containing special grain-boundary fractions in excess of 50% exhibit improve-ments of 15-fold and fivefold, respectively, in steady-state creep rate and primary creep strain.7,15

The EBSD hardware and software used in this workwere manufactured by EDAX-TSL, Inc. (Mahwah, NJ).The specimens were ground mechanically using 15-,6-, and 1-�m diamond suspension for 600 s and then

were polished using 0.06-�m colloidal silica for 3600 s.The step size used to obtain the EBSD orientationmaps was 0.5 �m. The reported fractions of GHABs,LABs, CSLBs, and twins (�3) were the averaged val-ues taken from several orientation maps, performedon the cross sections, rolling faces, and longitudinalsections of the thermomechanically processed sheetmaterial.

Flat dog-bone-shaped samples, with gauge section di-mensions of 3 mm wide by 2.5 mm thick by 10 mm long,were cut via electrodischarge machining (EDM). TheEDM recast layers were removed through grinding. Im-ages of the sample geometry and tensile stage are avail-able in Ref. 20, and details of this apparatus and thetensile-creep-testing technique can be found else-where.21,22 The specimens were glued to a metallicplaten and polished to a metallographic finish (a 60-nmparticle size was used for the final polish) using an au-tomatic polishing machine. The tensile-creep experi-ments were performed using a screw-driven tensile stagebuilt by Ernest F. Fullam, Inc. (Clifton Park, NY) placedinside the chamber of a FEI (Hillsboro, OR) XL-30 FEGSEM. A 4.4-kN load cell was attached to the end of thedisplacement assembly to monitor the load. The experi-ments were performed in load-control mode, and thesamples were loaded to the creep stress at a rate of 3.7N/s (50 lbs/min). Thereafter, the load was maintained at±8 N of the desired creep load. The samples were radi-antly heated at a rate of approximately 1 °C/s, using a6-mm-diameter tungsten-based heating element placedjust below the gauge section of the sample. The heatingelement was surrounded by an insulating ceramic sheath,and only one face of the heating element was exposed. Adirect current power supply operated at a constant volt-age was used to produce the current needed to resistivelyheat the heating element. Cooling was provided using anopen-bath chiller located outside the SEM chamber,which continuously circulated distilled water at 15–20 °Cthrough the copper and flexible tubing to the heater as-sembly. The temperature was monitored by a thermo-couple spot-welded directly to the sample’s gauge sec-tion. An additional thermocouple was placed below theheating element and was also monitored throughout theexperiments. The samples were held at temperature for30 min prior to the addition of a load to minimize thermalstresses. The load displacement–time relationship wasobtained live during the experiments using MTESTW,version F 8.8e, data acquisition and control software(Admet, Inc., Norwood, MA). The displacement data ac-quired during the experiments comprised that of thesample as well as the gripping fixtures; therefore, thedisplacement values reported in this study do not repre-sent the sole displacement of the sample’s gauge section.At selected intervals during the experiments, SE imageswere obtained from identical areas of the sample without

TABLE I. Measured composition of the Udimet 188 alloy in wt%.

Co Ni Cr W Fe Mn Si C Al Mo

36.8 24.2 22.1 12.0 2.5 1.0 0.4 0.1 0.1 0.4

C.J. Boehlert et al.: The evolution of grain-boundary cracking evaluated through in situ tensile-creep testing of Udimet alloy 188

J. Mater. Res., Vol. 23, No. 2, Feb 2008 501

interrupting the creep experiment. The pressure withinthe SEM chamber never exceeded 3 × 10−6 Torr through-out the duration of the experiments, which lasted up to70 h. Two of the experiments were performed at tem-perature (T) � 650 °C and � � 300 MPa. The other fiveexperiments were performed such that the applied stresswas 220 MPa and the temperature was 760 °C to mimicconventional creep experiments performed in air on thissame material.23 For these experiments, EBSD orienta-tion maps were acquired both prior to and after defor-mation for identical areas of the sample. For one of thesesamples, EBSD orientation maps were acquired in situ atseveral intervals during the creep deformation for iden-tical areas of the sample without interrupting the creepexperiment. This was accomplished by tilting the SEMstage to approximately 66° with respect to the incidentbeam and inserting the EBSD detector to a distance ofapproximately 21 mm from the center of the electronbeam. The working distance for this configuration was29 mm, and the tilt correction feature was used to adjustfor any foreshortening in the SEM images. The in situEBSD orientation maps were acquired over an area ofapproximately 200 �m × 200 �m using a step size

of 0.5 �m in 1200 s (approximately 140,000 points wereacquired with a speed of approximately 150 points persecond). The amount of displacement measured by thetesting assembly during this 20-min period was typically7 �m, and this small drift did not significantly disturb theorientation-mapping accuracy. The confidence of cor-rectly indexing the EBSPs was high initially as >95% ofthe EBSPs were correctly indexed, though with increasedtime and displacement, the EBSP quality decreased. Thiswas expected to be a result of several factors, includingsome surface topography due to the deformation/grain-boundary cracking and the small amount of surface oxi-dation.

III. RESULTS AND DISCUSSION

The grain boundaries were intact and uncracked afterthe thermomechanical processing. Figure 1 illustrates theEBSD data obtained at room temperature prior to heat-ing. The orientation maps [Figs. 1(a) and 1(c)] show thatthe alloy exhibited an equiaxed microstructure and theorientations were distributed fairly evenly. Pole figureanalysis indicated that the microstructure was not

FIG. 1. (a) EBSD orientation map and (b) grain-boundary misorientation angle chart for a 25% cold-rolled microstructure. (c) EBSD orientationmap and (d) grain-boundary misorientation angle chart for a 10% cold-rolled microstructure.

C.J. Boehlert et al.: The evolution of grain-boundary cracking evaluated through in situ tensile-creep testing of Udimet alloy 188

J. Mater. Res., Vol. 23, No. 2, Feb 2008502

strongly textured. A large volume fraction of twins wasobserved in this microstructure [Figs. 1(b) and 1(d)]. TheGHAB, LAB, CSLB, and twin boundary fractions were0.30, 0.50, and 0.39, respectively, for the 10% cold-rolled material. The GHAB, LAB, CSLB, and twinboundary fractions were 0.35, 0.12, and 0.45, respec-tively, for the 25% cold-rolled material. If we considerthat the combination of the LAB and CSLB fractionsconstitute the overall fraction of special boundaries,then the average special boundary fractions were 0.7 and0.65, respectively, for the 10% and 25% cold-rolled ma-terials. Thus, there was not a large difference in theGBCD between the two thermomechanically processedmicrostructures, both of which exhibited a majority ofspecial boundaries. Each microstructure exhibited asmall amount (<5 vol%) of fine spherical precipitates,which, based on previous literature,6 were believed to beMC6 or M23C6 carbide precipitates that were homoge-neously distributed throughout the equiaxed face-centered cubic (fcc) grains. The average fcc equiaxedgrain size (including twins) was approximately 30 �mand 39 �m for the 10% and 25% cold-rolled microstruc-tures, respectively.

Typical displacement-versus-time curves from the insitu tensile-creep experiments are shown in Fig. 2. For allof the tensile-creep experiments performed, the 10%cold-rolled material tended to exhibit slightly lower dis-placement values for a given time than the 25% cold-rolled material, which is consistent with previous obser-vations made for conventional creep experiments.23 Theslightly greater volume fraction of special boundariespresent in the 10% cold-rolled material may have ac-counted for the slightly lower displacement values ob-served as it has been observed that a high fraction of

special boundaries leads to higher creep-rupture lives fornickel.7 This result was suggested to be due to the in-ability of special boundaries to absorb extrinsic latticedislocations that facilitate sliding and cavitation. In ad-dition, Watanabe24 showed that special boundaries slidemuch less than random boundaries during creep. It isnoted that ductile failure was always evident in bothmicrostructures. Although the samples were not etched,the grain boundaries were observed using SE imagingdue to the thermal effects at temperatures above 600 °C.Figures 3(a)–3(j) illustrate SE images taken at sequentialdisplacement levels, indicated at the lower right-handcorner of the images, during tensile-creep deformationfor a 10% cold-rolled sample tested at T � 760 °C and� � 220 MPa. The evolution of grain-boundary crackingis revealed by these images. It is noted that the estimateddisplacement that the heated gauge section underwentwas less than this value as the displacement measurementincluded the entire sample and grip assembly. The mi-crocracks grew into larger cracks, and were only ob-served at the grain boundaries and did not proceed intothe grain interior. The grain-boundary cracks grew andbecame more extensive as the creep time and displace-ment increased. The cracks propagated both perpendicu-lar and parallel to the loading direction, though the larg-est number of cracks was for boundaries closer to per-pendicular to the loading axis rather than closer toparallel to the loading axis. Wider cracks were evidentfor boundaries perpendicular to the loading direction.The total number of boundaries characterized in the arearepresented by Fig. 3 was 92, in which 41 were GHABsand 51 were either CSLBs or LABs. Figures 4(a)–4(j)are EBSD image-quality maps for the same area. Theblack areas at the grain boundaries highlight regionswhere indexable EBSPs were not possible. Figure 5 in-dicates the boundaries that cracked during the experi-ment (black line overlaid) for the same area representedin Figs. 3 and 4. It is clear that the cracks tended to formon the GHABs. Combining the observations for all of thesamples examined (including well over 500 boundariesanalyzed with approximately 200 cracked boundaries),over 85% of the boundaries that exhibited cracking wereGHABs. This implies that CSLBs and LABs are lesssusceptible to grain-boundary cracking than GHABsduring the creep deformation of this alloy. Evaluationof polished gauge sections of the deformed creep speci-mens confirmed that grain-boundary cracking occurredwithin the subsurface bulk of the material as well. Due tothe amount of deformation and the inability to obtainEBSD data from subsurface locations prior to testing, itcould not be deciphered whether the majority of thesubsurface grain-boundary cracks originated at GHABs.Future work is intended to understand and model whycertain GHABs are more susceptible to cracking thanothers.

FIG. 2. Displacement-versus-time curves for creep experiments per-formed at T � 650 °C and � � 300 MPa.

C.J. Boehlert et al.: The evolution of grain-boundary cracking evaluated through in situ tensile-creep testing of Udimet alloy 188

J. Mater. Res., Vol. 23, No. 2, Feb 2008 503

FIG. 3. SE SEM images taken of the same area during a tensile-creep experiment at T � 760 °C and � � 220 MPa where the creep-displacementlevels are indicated in the lower right-hand corner of the images. The loading axis was horizontal.

C.J. Boehlert et al.: The evolution of grain-boundary cracking evaluated through in situ tensile-creep testing of Udimet alloy 188

J. Mater. Res., Vol. 23, No. 2, Feb 2008504

FIG. 4. EBSD image-quality maps taken during the tensile-creep experiment at T � 760 °C and � � 220 MPa for the same area indicated inFig. 3, where the creep-displacement levels are indicated in the lower right-hand corner of the images. The loading axis was horizontal.

C.J. Boehlert et al.: The evolution of grain-boundary cracking evaluated through in situ tensile-creep testing of Udimet alloy 188

J. Mater. Res., Vol. 23, No. 2, Feb 2008 505

IV. SUMMARY

This work describes the technique that has been de-veloped to acquire SE detector images and EBSD orien-tation maps during tensile-creep deformation at tempera-tures as high as 760 °C. EBSD observations during de-formation proved to be useful for quantifying the types ofgrain boundaries that have preferentially cracked. Tosummarize, the tensile-creep deformation evolution of acobalt-based superalloy, Udimet 188, processed using agrain-boundary-engineering approach, was evaluated us-ing surface observations during in situ tensile-creep test-ing in a SEM chamber. This alloy exhibited surface crack-ing that initiated and propagated preferentially at GHABs.

ACKNOWLEDGMENTS

This work was supported by the National ScienceFoundation through Grant DMR-0533954. The author isgrateful to Mr. Nathan Eisinger (Special Metals Corpora-tion, Huntington, WV) for overseeing the alloy pro-cessing, and to Mr. John Carpenter of EDAX-TSL, Inc. forhis technical assistance.

REFERENCES1. C.J. Lissenden, J.F. Colaiuta, and B.A. Lerch: Hardening behavior

of three metallic alloys under combined stress at elevated tem-perature. Acta Mech. 169, 53 (2004).

2. D. Zhu, D.S. Fox, and R.A. Miller: Oxidation and creep-enhancedfatigue of Haynes 188 alloy-oxide scale system under simulatedpulse detonation engine conditions. Ceram. Eng. Sci. Proc. 23.4,547 (2002).

3. J. Chen, P.K. Liaw, Y.H. He, M.L. Benson, J.W. Blust,P.F. Browning, R.R. Seeley, and D.L. Karstom: Tensile hold low-cycle fatigue behavior of cobalt-based HAYNES� 188 superalloy.Scripta Mater. 44, 859 (2001).

4. J.D. Whittenberger: Mechanical properties of Haynes alloy 188after exposure to LiF-22CaF2, air, and vacuum at 1093K for pe-riods up to 10,000 hours. J. Mater. Eng. Perf. 1, 469 (1992).

5. J.D. Whittenberger: Mechanical properties of Haynes alloy 188after 22,500 hours of exposure to LiF-22CaF2 and vacuum at 1093K. J. Mater. Eng. Perf. 3, 754 (1994).

6. R.B. Herchenroeder, S.J. Matthews, J.W. Tackett, and S.T. Wlodek:Haynes alloy No. 188. Cobalt 54, 3 (1972).

7. E.M. Lehockey and G. Palumbo: On the creep behavior of grainboundary engineered nickel. Mater. Sci. Eng., A 237, 168 (1997).

8. V. Randle: The Role of Coincident Site Lattice in Grain BoundaryEngineering (The Institute of Materials, London, 1996), pp.1–104.

9. V. Randle: Refined approaches to the use of the coincidence sitelattice. J. Met. 50(2), 56 (1998).

10. E.M. Lehockey, G. Palumbo, and P. Lin: Improving the weldabil-ity and service performance of nickel- and iron-based suparalloysby grain boundary engineering. Metall. Mater. Trans. A 29, 3069(1998).

11. B. Alexandreanu, B.M. Capell, and G. Was: Combined effect ofspecial grain boundaries and grain boundary carbides on IGSCC.Mater. Sci. Eng., A 300, 94 (2001).

12. E.M. Lehockey, G. Palumbo, P. Lin, and A.M. Brennenstuhl: Onthe relationship between grain boundary character distribution andintergranular corrosion. Scripta Mater. 36(10), 1211 (1997).

13. C. Cheung, U. Erb, and G. Palumbo: Applications of grain bound-ary engineering concepts to alleviate intergranular cracking inalloys 600 and 690. Mater. Sci. Eng., A 185, 39 (1994).

14. G. Palumbo and K.T. Aust: Special properties of � grain bound-aries, in Materials Interfaces: Atomic Level Structure and Prop-erties, edited by D. Wolf and S. Yip (Chapman and Hall, NewYork, 1989), pp. 190–211.

15. G. Palumbo, E.M. Lehockey, and P. Lin: Applications for grainboundary engineered materials. J. Met. 50(2), 40 (1998).

16. P. Lin, G. Palumbo, U. Erb, and K.T. Aust: Influence of grainboundary character distribution on sensitization and intergranularcorrosion of alloy 600. Scripta Mater. 33(9), 1387 (1995).

17. G.S. Was, V. Thaveeprungsriporn, and D.C. Crawford: Grainboundary misorientation effects on creep and cracking in Ni-basedalloys. J. Met. 50(2), 44 (1998).

18. V. Thaveeprungsriporn and G. Was: The role of coincidence-site-lattice boundaries in creep of Ni-16Cr-9Fe at 36 °C. Metall.Mater. Trans. A 28, 2101 (1997).

19. G. Palumbo and K.T. Aust: Structure-dependence of intergranularcorrosion in high purity nickel. Acta Metall. Mater. 38(11), 2343(1990).

20. C.J. Boehlert, C.J. Cowen, S. Tamirisakandala, D.J. McEldowney,and D.B. Miracle: In situ scanning electron microscopy observa-tions of tensile deformation in a boron-modified Ti-6Al-4V alloy.Scripta Metall. 55, 465 (2006).

21. J.P. Quast and C.J. Boehlert: Comparison of the microstructure,tensile, and creep behavior for Ti-24Al-17Nb-0.66Mo (at.%) andTi-24Al-17Nb-2.3Mo (at.%) alloys. Metall. Mater. Trans. A 38,529 (2007).

22. C.J. Cowen and C.J. Boehlert: Comparison of the microstructure,tensile, and creep behavior for Ti-22Al-26Nb (at.%) and Ti-22Al-26Nb-5B (at.%). Metall. Mater. Trans. A 38, 26 (2007).

23. C.J. Boehlert: The effect of cold rolling on the creep behavior ofUdimet alloy 188. World Sci. Eng. Acad. Soc. Trans. Heat MassTransfer 1, 156 (2006).

24. T. Watanabe: An approach to grain-boundary design for strongand ductile polycrystals. Res. Mech. 11(1), 47 (1984).

FIG. 5. EBSD inverse pole figure orientation map indicating theboundaries that cracked during the experiment (black line overlaid) forthe same area indicated in Figs. 3 and 4. Several of the grain bound-aries are labeled (R � GHAB, �3 � twin, L � LAB, �”X” �CSLB), and it is clear that the cracks tended to form on the GHABs.The loading axis was horizontal.

C.J. Boehlert et al.: The evolution of grain-boundary cracking evaluated through in situ tensile-creep testing of Udimet alloy 188

J. Mater. Res., Vol. 23, No. 2, Feb 2008506