International Journal of Machine Tools & Manufacture 41 (2001) 16811685
The effects of machined workpiece surface integrity on thefatigue life of -titanium aluminide
A.R.C. Sharman a,*, D.K. Aspinwall b,c, R.C. Dewes b, D. Clifton d, P. Bowen a,ca School of Metallurgy and Materials, University of Birmingham, Edgbaston, Birmingham, B15 2TT, UK
b School of Manufacturing and Mechanical Engineering, University of Birmingham, Edgbaston, Birmingham, B152TT, UK
c IRC in Materials for High Performance Applications, University of Birmingham, Edgbaston, Birmingham, B152TT, UK
d Department of Mechanical Engineering, Edinburgh University, The Kings Building, Edinburgh, UKReceived 19 February 2001; accepted 27 February 2001
Tensiontension tests on turned, electro-chemical machined (ECM) and electro-discharge textured (EDT)specimens made from Ti45Al2Nb2Mn+0.8 vol% TiB2 alloy, showed the turned specimens to have ahigher fatigue strength 475 MPa. It is likely that this was due to the presence of highly compressivesurface residual stresses caused by the turning operation. 2001 Elsevier Science Ltd. All rights reserved.
Keywords: Titanium aluminide; Surface integrity; Fatigue life; Machining
The surface integrity produced by different machining procedures can significantly affectfatigue performance . Indeed according to Zlatin and Field , the major mechanical propertyaffected by the type of machining operation and its severity is high cycle fatigue strength.-Titanium aluminides (-TiAl) are attracting considerable interest as possible replacements for
nickel-based superalloys in the high pressure compressor and low pressure turbine stages of gasturbine engines, due to their low density (3.76 g/cm3) and relatively high operating temperatureof up to 800C . The literature on the fatigue properties of -TiAl with respect to microstructure,alloying, testing, etc. is extensive, however, very little of this data refers to the effects of machined
* Corresponding author. Tel.: +44-121-414-3541.E-mail address: firstname.lastname@example.org (A.R.C. Sharman).
0890-6955/01/$ - see front matter 2001 Elsevier Science Ltd. All rights reserved.PII: S0 890- 695 5(01 )000 34- 7
1682 A.R.C. Sharman et al. / International Journal of Machine Tools & Manufacture 41 (2001) 16811685
surface integrity. Here, the majority of data has been reported by Mantle and Aspinwall , Trailand Bowen , Bentley et al.  and Jones and Eylon . The major conclusions from this datawere that the fatigue life obtained was relatively unaffected by the specimen surface conditionand no evidence of preferential crack growth from surface defects was seen. Greater knowledgeof the effects of surface integrity on fatigue life is critical to the acceptance of -TiAl alloys inservice, especially in the aerospace industry.
2. Experimental work
The workpiece material used was a grain-refined -TiAl comprising Ti45Al2Mn2Nb+0.8 vol% TiB2 XD produced by Howmet, USA. This was HIPped at 1260C/170 MPa for4 h then heat treated at 1010C for 50 h to improve mechanical properties. The material had anear fully lamellar microstructure with a 50100 m colony size and a bulk hardness of 350400 HK0.025.
Fatigue tests (SN) were conducted using an Amsler Vibrophore electro-magnetic resonancetesting machine in tensiontension configuration. Tests were carried out at an R ratio of 0.1 (whereR is the ratio of maximum stress to minimum stress applied over the fatigue cycle) and wereperformed at room temperature with specimen run-out being taken as 1.2107 cycles.
Four surface conditions were produced by turning, electro-chemical machining (ECM) andelectro-discharge texturing (EDT). The turned samples were made using operating conditions thatwere known to cause minimal surface damage. The ECM samples were initially turned oversizeand subsequently machined to remove 250 m from the test surface. The EDT samples wereprocessed in a similar fashion. Two levels of EDT operating energy were selected, one usinghigh energy parameters (EDT-H) to give specimens with deep cracks and one using low energyparameters (EDT-L) to produce shallow cracks. Table 1 details the surface integrity of thefatigue specimens.
Sample machined surfaces were prepared for two-dimensional (2-D) topographical, microstruc-tural and surface analysis. Sections were hot mounted in bakelite, ground using SiC paper andpolished with SiO2 solution. After polishing they were etched for 7 s in Krolls reagent. Surfaceroughness measurements (Ra and Rt) were conducted on a surface profilometer with a cut-offand evaluation length of 0.8 and 4 mm, respectively. Each measurement was repeated five timesat different positions on the surface and an average taken. Surface residual stress values were
Table 1Surface integrity data for the fatigue specimens
Machining procedure Max. crack depth (m) Ra (m) Residual stress
Turning 5 0.84 Highly compressiveECM 0 1.43 NeutralEDT-L 10 1.32 TensileEDT-H 40 3.98 Tensile
1683A.R.C. Sharman et al. / International Journal of Machine Tools & Manufacture 41 (2001) 16811685
Fig. 1. SN curves for turned, ECM and EDT specimens.
measured using a purpose built X-ray diffraction (XRD) machine, which had been modified toovercome the line broadening problems encountered when using XRD on -TiAl alloys .
3. Results and discussion
Fig. 1 shows the SN curves for all the specimens, which are relatively flat with large variationin the number of cycles to failure at similar stress levels, as seen in previous -TiAl results . Turned specimens produced the highest run-out strength of 475 MPa. ECM specimens had aslightly lower run-out strength of 440 MPa, whilst the EDT-L and EDT-H specimens had signifi-cantly reduced values of 357 and 225 MPa, respectively.
The low fatigue strength exhibited by the EDT specimens was to be expected given the presenceof cracks running directly into the bulk of the material as shown in Fig. 2. Due to the extremelyhigh temperature generated during the EDT process (12,000C), a molten pool of metal is formedat the workpiece surface. This cools rapidly giving rise to tensile residual stresses and microcracks.
Fig. 2. Typical crack morphology of the EDT specimens.
1684 A.R.C. Sharman et al. / International Journal of Machine Tools & Manufacture 41 (2001) 16811685
Fig. 3. Selective etching of the workpiece produced by ECM.
The cracks were not restricted to the recast white layer but also extended into the material belowas found in previous work on electro-discharge machining (EDM) of -TiAl . These crackstended to follow lamellae interfaces and colony boundaries both of which have been shown tooffer very little resistance to fatigue crack propagation [10,11]. In the current work, fractographyhas shown that crack propagation to failure occurred from cracks induced by the EDT process.
The relatively rough surface (1.43 m Ra) of the ECM specimens was produced by a selectiveetching of individual lamellae within each colony as shown in Fig. 3. Although ECM surfacescontained no cracks, the selectively etched lamellae would be expected to act as points of stressconcentration and the poor adherence of individual lamellae would offer very little resistance tocrack initiation.
In contrast, although the turned specimens contained small shallow cracks (5 m deep) asshown in Fig. 4, these cracks did not propagate during fatigue cycling. Fractography has shownthat for the turned specimens, crack initiation occurred due to interlamellar plate failure withinlamellae colonies which were oriented at a favourable angle to the applied load. This occurredat both near surface and internal sites. The higher life exhibited by the turned specimens wouldbe produced by the increased number of cycles required to propagate a crack through the com-pressive residual stress layer.
Fig. 4. Typical crack morphology of the turned specimens.
1685A.R.C. Sharman et al. / International Journal of Machine Tools & Manufacture 41 (2001) 16811685
The presence of cracks penetrating into the bulk of the fatigue specimens, in combination withthe presence of tensile residual stresses, resulted in the substantially reduced fatigue life of theEDT specimens compared to those which had been turned and ECMed. The lower life of EDT-H compared to EDT-L specimens reflects the presence of deeper cracks and higher tensile residualstresses. It is likely that the higher fatigue life of the turned specimens was due to the highlycompressive residual stresses present in the machined surface.
The authors would like to thank the UK Engineering and Physical Sciences Research Council(EPSRC), Rolls-Royce Plc (Wayne Voice and Colin Sage), De Beers Industrial Diamonds Ltd(Matthew Cook), and Sandvik Coromant UK (Andy Smith) for funding and technical support.
 G.E. Dieter, Mechanical Metallurgy, McGraw-Hill, New York, 1988, ISBN 0-07-100406-8. N. Zaltin, M. Field, Procedures and precautions in machining titanium alloys, Titanium Sci. Technol. 1 (1973)
489504. W. Voice, Future use of gamma titanium aluminides by Rolls-Royce, Aircraft Engng Aero Technol. 71 (4) (1999)
337340. A.L. Mantle, D.K. Aspinwall, Surface integrity and fatigue life of turned gamma titanium aluminide, J. Mater.
Process. Technol. 72 (1997) 413420. S.J. Trail, P. Bowen, Effects of stress concentrations on the fatigue life of a gamma based titanium aluminide,
Mater. Sci. Engng A 192/193 (1995) 427434. S.A. Bentley, A.L. Mantle, D.K. Aspinwall, The effect of machining on the fatigue strength of a gamma titanium
aluminide intermetallic alloy, Intermetallics 7 (1999) 967969. P.E. Jones, D. Eylon, Effects of conventional machining on high cycle fatigue behavior of the intermetallic alloy
Ti47Al2Nb2Cr, Mater. Sci. Engng A 263 (1999) 296304. EPSRC IMI grant Ref. GR/L33993, 1996. A.L. Mantle, E.Abboud, D.K. Aspinwall, Productivity and workpiece surface integrity effects when electrical
discharge wire machining a gamma titanium aluminide, in: Proceedings of the 14th IMC, Dublin, 1997, pp. 443450, ISBN 897606-16-8.
 P. Bowen, N.J. Rodgers, A.W. James, Fracture and fatigue of cast gamma TiAl based aluminides, in: Y.-W. Kim,R. Wagner, M. Yamaguchi (Eds.), Gamma Titanium Aluminides, Proceedings of the TMS95, 1995, pp. 849865, ISBN 0-87339-304-X.
 J.P. Campbell, K.T.V. Rao, R.O. Ritchie, On the role of microstructure in fatigue crack growth of gamma basedtitanium aluminides, Mater. Sci. Engng A 239/240 (1997) 722728.