- Email: [email protected]
Investigation on the surface and near-surface characteristics of Ti–2.5Cu after various mechanical surface treatments
Surface & Coatings Technology 205 (2011) 3644–3650
Contents lists available at ScienceDirect
Surface & Coatings Technology j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / s u r f c o a t
Investigation on the surface and near-surface characteristics of Ti–2.5Cu after various mechanical surface treatments E. Maawad a,⁎, H.-G. Brokmeier a,b, L. Wagner a, Y. Sano c, Ch. Genzel d a
Institute of Materials Science and Engineering, Clausthal University of Technology, Agricolastr. 6, D-38678 Clausthal-Zellerfeld, Germany Helmholtz-Zentrum Geesthacht Centre for Materials and Coastal Research, D-21502 Geesthacht, Germany Toshiba Corporation, 8 Shinsugita-cho Isogo-ku Yokohama 235-8523, Japan d Helmholtz-Zentrum Berlin (BESSY), Albert-Einstein-Str. 15, D-12489 Berlin, Germany b c
a r t i c l e
i n f o
Article history: Received 14 September 2010 Accepted in revised form 5 January 2011 Available online 12 January 2011 Keywords: Mechanical surface treatment Shot peening Ball-burnishing Laser shock peening Ultrasonic shot peening Residual stress
a b s t r a c t As a result of the variety of mechanical surface treatments, studies of surface and near-surface characteristics are becoming increasingly important in a variety of industrial fields. This is to achieve more gains by balancing between optimum conditions and costs. In the present study, shot peening (SP), ball-burnishing (BB), laser shock peening (LSP) and ultrasonic shot peening (USP) processes were performed on the α-titanium alloy Ti– 2.5Cu after two different heat treatments. The influence of such surface treatments on the surface and nearsurface characteristics, such as residual stress, work hardening or dislocation density and surface roughness, was studied. The depth profiles of residual stress and full width at half maximum (FWHM) were obtained by using energy-dispersive X-ray diffraction. Results show that the BB process produced the highest and deepest maximum residual stress and LSP produced the lowest work hardening close to the surface. © 2011 Elsevier B.V. All rights reserved.
1. Introduction Mechanical surface treatments are widely used to prolong fatigue life of many engineering components in the industry. This is attributed to the changes of surface and near-surface characteristics by producing plastic deformation and compressive residual stress [1]. Shot peening (SP) is most commonly used due to its low cost and can be used on small or large areas depending on requirements. However, there are several techniques which could be competitive such as ballburnishing (BB), laser shock peening (LSP) and ultrasonic shotpeening (USP). BB leads to a smoother surface and deeper compressive residual stress than SP process [2]. Disadvantages of this process are difficulty in tooling development and poor access to some locations, for example, on engine components. Compared to the SP process, LSP can produce deeper compressive residual stresses [3]. Therefore, it is used mainly for increasing fatigue strength [4]. Furthermore, the compressive residual stresses are produced with less work hardening than that with SP, allowing for less thermal relaxation of these stresses when the part is subjected to high temperatures [5]. In addition, the process, effect and application
Abbreviations: SP, shot peening; BB, ball-burnishing; LSP, laser shock peening; USP, ultrasonic shot peening. ⁎ Corresponding author. IWW, Agricolastr. 6, D-38678 Clausthal-Zellerfeld, Germany. Tel.: + 49 5323 72 2758; fax: + 49 5323 72 2766. E-mail address: [email protected] (E. Maawad). 0257-8972/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.surfcoat.2011.01.001
of LSP without coating were studied [6]. It was concluded that LSP without coating prevents stress corrosion cracking (SCC) and prolongs fatigue life. On the other hand, the major disadvantage of LSP is that it currently requires a laser system specially designed for production runs with high average power and large pulse energy. Recently, USP has been studied to improve the durability of structural parts [7]. In USP, unlike SP where the shot is propelled by compressed air, the shot is energized by a sonotrode vibrating at ultrasonic frequency. The main characteristic of such process could give a smooth surface compared to SP because of using polished bearing ball with a large diameter. Furthermore, it can give much energy in short time due to using high frequency [8]. There are several researches studying the influence of various surface treatments on the surface and near-surface characteristics in different alloys. Turski et al. [9] investigated on the surface roughness, microstructure, level of plastic work and through thickness residual stress distribution in austenitic steel AISI 304 after different surface treatments. It was concluded that the samples show compressive stresses to a significantly greater depth for the LSP, ultrasonic impact treatment and water jet cavitation peening samples compared to the more conventional SP treated sample. The influences of deep rolling (DR) and LSP on the fatigue behavior of Ti–6Al–4V and austenitic steel AISI 304 were investigated by R.K. Nalla et al. [5] and I. Nikitin et al. [10], respectively. It was found that DR induced larger near-surface compressive residual stresses and degree of work hardening compared to LSP. Recently, a new family of surface severe plastic deformation processes (SPD) using high energy balls, widely known as the surface
E. Maawad et al. / Surface & Coatings Technology 205 (2011) 3644–3650
mechanical attrition treatment (SMAT), has been developed to induce in workpieces the desired nanocrystalline surface layer [11]. SMAT produces a thicker nanocrystalline surface layer, a deeper plastically deformed region with greater abundance of twins, faults and dislocations, and a thicker surface region with larger residual compressive stresses than SP [12,13]. These differences make SMAT more effective in improving the fatigue properties than SP. Ti–2.5Cu has many applications including sheets, forgings and extrusions for fabricating components such as bypass ducts of gas-turbine engines as well as airframes in aeronautical industry. As a result of the variety of material responses to various surface treatments, the main aim of the present work is to study the influence of SP, BB, USP or LSP on the surface and near-surface characteristics in Ti–2.5Cu, such as residual stress, surface roughness and work hardening. 2. Material and methods 2.1. Material The Ti–2.5Cu alloy was received as a 10 mm thick rolled plate. Four specimens were cut from this plate with a dimension of 20 × 20 × 5 mm3 perpendicular to the rolling direction followed by two different heat treatments. Solution heat treatment (SHT) was done at 805 °C for 1 h followed by water-quenching. The other was SHT followed by double aging (SHT + A) by annealing the material at 400 °C for 8 h and then annealing at 475 °C for 8 h. The microstructures of Ti–2.5Cu after the two conditions are illustrated in Fig. 1.
3645
Table 1 Tensile properties and hardness of Ti–2.5Cu.
E (GPa) σy (MPa) UTS (MPa) El (%) εf = ln (Ao/Af)
SHT
SHT+A
105 520 610 28 0.62
110 685 770 16.4 0.57
E = modulus of elasticity, σy = yield stress, UTS = ultimate tensile stress, EL= elongation, and εf = true fracture strain.
It was observed that both microstructures consist of α grains and stringers of the eutectoid component α + Ti2Cu. Threaded cylindrical tensile specimens were machined having a gage length and a diameter of 20 and 4 mm, respectively. Tensile properties and hardness of Ti–2.5Cu with different heat treatments are listed in Table 1. Ti–2.5Cu (SHT + A) shows a pronounced increase in modulus of elasticity, yield strength, ultimate tensile stress and hardness due to the precipitation hardening influence of Ti2Cu. In contrast, the ductility decreased after aging. 2.2. Mechanical surface treatments The specimens were mechanically surface treated on both surfaces with 20 × 20 mm2 by applying SP, BB, LSP or USP. SP was performed at TU Clausthal in Germany using cast steel (S330) having an average shot diameter of 0.8 mm and a hardness of 460 HV. While USP was applied at MTU Aero Engines in Germany using 100Cr6 bearing steel balls with a diameter of 1.5 mm and a hardness of 700–800 HV. The main reason for using 0.8 mm shots in SP and 1.5 mm balls in USP is their higher mass and therefore the higher compressive residual stresses that can be achieved. Peening was performed to roughly 100% coverage in SP and USP using the same Almen intensity of 0.20 mmA. The influence of the different peening media parameters with the same Almen intensity used in SP and USP, such as speed, diameter and hardness, on the surface and near-surface characteristics was considered when discussing the experimental results. Furthermore, Ti–2.5Cu specimens were ball-burnished at TU Clausthal using a conventional lathe and a hydrostatic tool by which a hard metal ball (Ø6 mm) is pressed with a pressure of 300 bar onto the specimen surface. LSP was carried out at Toshiba Corp. in Japan using a compact Q-switched and frequency-doubled Nd:YAG laser with a power density of 5 GW cm− 2, a pulse duration of 8 ns and a spot size of 0.8 mm in diameter. Water was used as a tamping material limiting the thermal expansion of plasma gas. LSP was done without coating having some characteristics which were mentioned elsewhere [6]. 2.3. Residual stress and full width at half maximum
Fig. 1. Microstructure of (a) Ti–2.5Cu (SHT) and (b) Ti–2.5Cu (SHT + A), average grain size ≈ 20 μm (SHT = solid solution heat treatment, A = aging).
Residual stress measurements were performed by hard X-ray diffraction using synchrotron radiation at BESSY-II in Berlin. The characteristic of the used beamline EDDI offers a white X-ray beam with an energy range of 10–80 keV. The primary beam cross-section was set to 0.5 × 0.5 mm2. The angular divergence in the diffracted beam was restricted by a double slit system with apertures of 0.03 × 5 mm2 to Δθ ≤ 0.005°. The scattering angle was chosen to 8° considering the energy of X-ray. Energy-dispersive X-ray diffraction (ED) gives complete diffraction spectra for a fixed detector position. Any Bragg reflection was obtained by a different X-ray energy (wavelength) which means the signal of any reflection belongs to a different depth in the sample as schematically shown in Fig. 2. Due to the limited usable energy range provided by the 7T multipole wiggler which extends from about 10 keV to 80 keV, the maximum information depth for titanium accessible in reflection mode experiments is about 100 μm. In order to get a stress distribution in deeper region,
3646
E. Maawad et al. / Surface & Coatings Technology 205 (2011) 3644–3650
Fig. 2. Scheme showing a relation between energy (E) and penetration depth (τ).
three specimens were prepared. For the two of them, layer removal in a step of 100–150 μm was applied by electropolishing. The correction on the measured residual stress was made for the electropolished specimens using the equation described elsewhere [14]. Residual stresses were evaluated by means of the sin2ψ method, where ψ is the tilting angle, in steps of Δψ = 4° up to 80°. A modified multiwavelength approach [15] for any energy line E(hkl) gives an average penetration depth τ(hkl) (Eq. (1)). τðhklÞ ¼ ðτðhklÞmin þ τðhklÞmax Þ=2
ð1Þ
where τ(hkl)min and τ(hkl)max are the minimum and the maximum penetration depths corresponding to the maximum (ψmin) and minimum (ψmax) tilting angles, respectively. The diffraction elastic constants of alpha-reflections were calculated by the Kröner-model [16]. Residual stresses at the surface were determined by using laboratory X-ray diffraction (D-5000 at TU Clausthal) using Cu-Kαradiation with a wavelength of 1.54 Å. The {21.3}-Bragg peak was used. The moduli of elasticity of 105 GPa for Ti–2.5Cu (SHT) and 110 GPa for Ti–2.5Cu (SHT + A) with Poisson's ratio of 0.33 were used to calculate the diffraction elastic constants (S1 and 1/2 S2). The sin2ψ method was also used to evaluate the surface residual stresses with tilting angles range of −60° to +60° in steps of Δsin2ψ = 0.125. Work hardening or dislocation density distribution can be evaluated by means of X-ray diffraction peak broadening using characteristic parameters of individual peak profiles, especially the full width at half maximum (FWHM) and the integral breadths. It is well known that besides the instrumental contribution, there are two main types of broadening: the size and the strain components. The former one depends on the finite size of the coherent diffraction domains and the latter is caused by any lattice imperfection (point, line or plane defects). The strain field of linear defects, such as dislocations, is of long-range character, therefore their diffraction effects cluster around the fundamental Bragg reflections [17,18]. In the present work, FWHM was also determined by using ED. The (10.3) reflection was used due to its high multiplicity factor (n = 12) and corresponding reliability of the results as well as good statistics. The FWHM of (10.3) reflection was calculated by using the Pseudo-Voigt function [19] as shown in Fig. 3.
Fig. 3. (10.3) peak fitted by Pseudo-Voigt function to define FWHM.
Vickers tester, a nominal force of 100 gf (HV0.1) and a loading time of 10 s. The hardness testers in the Duramin series conform to the standard (DIN EN ISO 6507). The average of three measurements was taken at each depth to construct the hardness–depth profiles. 3. Results and discussions 3.1. Surface roughness The surface roughness (Rz) after the various surface treatments was determined and compared to the electropolished (EP) sample as a reference. As seen in Fig. 4, the surface roughness of shot peened Ti– 2.5Cu is much higher than that of the electropolished reference, while a remarkable improvement on the surface roughness was observed after BB (Fig. 5a) compared to that after SP (Fig. 5c). Furthermore, the surface roughness after USP became smoother than that after SP. As seen in Fig. 5b, wider dimples with smaller amplitude were produced after USP than that after SP (Fig. 5c). The reason of this result was considered that the balls in the case of USP are set into a random motion inside a component-specific peening chamber to act on the component, while the shot stream is nearly perpendicular to the surface in the case of SP. Therefore, the influence of the shot velocity on the indent depth or roughness is more significant in SP compared to that in USP. Furthermore, the bearing balls used for USP have higher spherical accuracy and smoother surface than the shots used in SP process. Contrary to what is expected for LSP with coating, a rough surface with linear furrows appeared after LSP without coating (Fig. 5d). When no protective laser-absorbent coating was used on the specimen, severe surface melting and vaporization would be induced
2.4. Surface roughness The surface roughness of the various conditions was determined by means of an electronic contact (stylus) profilometer instrument (Perpethometer). The average absolute value of the five highest peaks and the five lowest valleys over the evaluation length (Rz) was reported (DIN 4768). The average of three roughness measurements was taken. The parameter (Rz) was used rather than the average roughness comparing all the peaks and valleys to the mean line (Ra). Quite different surfaces could have the same Ra and consequently perform in a different manner [20]. 2.5. Microhardness Microhardness was determined by means of a Struers Duramin tester using a square base pyramid shaped indenter for testing in a
Fig. 4. Surface roughness after various surface treatments (SP = shot peening, BB = ball-burnishing, LSP = laser shock peening, and USP = ultrasonic shot peening).
E. Maawad et al. / Surface & Coatings Technology 205 (2011) 3644–3650
3647
Fig. 5. Images of the mechanically treated surface by (a) BB, (b) USP, (c) SP and (d) LSP (SP = shot peening, BB = ball-burnishing, LSP = laser shock peening, and USP = ultrasonic shot peening).
by LSP with high-power and long-duration laser pulses [21,22]. This can result in resolidified droplets and craters leading to very rough surfaces [23]. To prevent the surface from such damage, laser pulses with much lower energy and shorter duration were used in this study. 3.2. Work hardening/dislocation density The microhardness–depth distributions of Ti–2.5Cu after various mechanical surface treatments are shown in Fig. 6. It was found that the bulk hardness in Ti–2.5Cu (SHT) is 250 HV0.1 approximately (Fig. 6a), while the bulk hardness in Ti–2.5Cu (SHT + A) is about 320 HV0.1 (Fig. 6a). This increase of the hardness is explained by the precipitation hardening influence. The maximum work hardening induced by SP, BB, LSP and USP was observed at the surface which gradually decreased in the near-surface regions as shown in Fig. 6a and b. In both conditions of the heat treatments, it was found that the repeated dimpling at the surface by SP to achieve uniform surface coverage resulted in the highest cold worked layer or dislocation density at the surface. In contrast, LSP produced the lowest work hardening at the surface, although each point on the surface was hit 18 times (laser spot area × irradiation density) by the laser pulse. This is a result of the stress created by shock wave propagation rather than cold work as in SP. This lower magnitude of work hardening after LSP supports the previous work hardening results of laser peened hypoeutectoid steel [3]. In this study, laser beam with a small footprint (Ø0.8 mm) and a lower power density (5 GW cm− 2)
generated a less planar pressure wave to propagate shallower into the specimen producing plastic deformation with a depth of 0.45 and 0.6 mm from the treated surface in the case of SHT and SHT + A, respectively. In contrast, it was found in Ref. [9] that a deeper plastic deformation (2 mm from the treated surface) was produced after LSP with a larger footprint (3 × 3 mm) and a higher power density (10 GW cm− 2) resulting in a higher pressure in 304 austenitic stainless steel. It should be also taken into account that the tensile properties of Ti–2.5Cu and 304 austenitic stainless steel are different when comparing the influence of LSP parameters. This feature of LSP leads to a thermal stability of residual stress close to the surface. It was found that the work hardening induced by BB slightly decreased close to the surface (up to 100 μm in depth) compared to SP, while it increased in deeper region. This is due to less shearing between the ball and the specimen surface while using lubricant as well as due to larger ball size of 6 mm in diameter. The increase of work hardening or dislocation density in the near-surface region retards dramatically fatigue crack nucleation [24]. This is a result of a decrease of dislocation mobility which results in a limited number of slip planes [25] and corresponding slip band intrusions and extrusions. On the other hand, it accelerates fatigue micro-cracks propagation. Furthermore, it was observed that USP produced relatively higher work hardening than LSP and lower than SP and BB. This is a result of fewer impacts by large balls required to achieve the surface coverage. The work hardening behavior close to the surface after these surface treatments was supported by determining FWHM as shown in
3648
E. Maawad et al. / Surface & Coatings Technology 205 (2011) 3644–3650
Fig. 6. Microhardness–depth distribution in (a) Ti–2.5Cu (SHT) and (b) Ti–2.5Cu (SHT + A) after various surface treatments (SP= shot peening, BB= ball-burnishing, LSP= laser shock peening, USP= ultrasonic shot peening, SHT= solid solution heat treatment, and A = aging).
Fig. 7. As seen in Fig. 7b, the FWHM in the virgin region in Ti–2.5Cu (SHT + A) is relatively larger than that in Ti–2.5Cu (SHT). This is explained by the increase of Ti2Cu precipitation as a structural defect. Furthermore, it was observed that the FWHM distributions close to the surface agree with the microhardness distribution. The reason for some slight differences in the penetration depth between the microhardness and FWHM could be explained by the data scattering of microhardness due to the existence of eutectoid components (α + Ti2Cu) in the grain boundaries. In the present study, the contribution of nanocrystalline grains to FWHM was ignored because no significant changes of the grain size (20 μm) were observed close to the surface after the surface treatments. Fig. 8a and b shows the near-surface microstructures after SP and BB which produced higher work hardening in Ti–2.5Cu (SHT), respectively. 3.3. Residual stress The residual stress–depth distributions in Ti–2.5Cu (SHT) and in Ti–2.5Cu (SHT + A) after the various surface treatments are illustrated in Fig. 9a and b, respectively. The response of Ti–2.5Cu after different heat treatments to the various surface treatments and the influence of each surface treatment on the residual stress distribution were discussed. The values of residual stresses at the surface and the maximum compressive residual stresses are listed in Table 2. Obviously, the surface and maximum residual stresses in Ti–2.5Cu (SHT) is relatively lower than that in Ti–2.5Cu (SHT + A). This is explained by higher
Fig. 7. FWHM-depth distribution in (a) Ti–2.5Cu (SHT) and (b) Ti–2.5Cu (SHT + A) after various surface treatments (SP = shot peening, BB = ball-burnishing, LSP = laser shock peening, USP = ultrasonic shot peening, SHT = solid solution heat treatment, and A = aging).
modulus of elasticity and yield strength of Ti–2.5Cu (SHT + A). This is a result of relatively larger potential maxima of the induced residual stress by plastic deformation. It was observed that LSP induced the lowest residual stress at the surface, while USP induced lower residual stress than SP and BB and higher than LSP (see Table 2). This is attributed to the magnitude of work hardening induced at the surface as shown in Figs. 6 and 7. Lower work hardening resulted in lower flow stress and thus lower residual stress. However, it was found that BB produced higher residual stress at the surface compared to SP in both Ti–2.5Cu (SHT) and Ti–2.5Cu (SHT + A) despite the fact that BB produced lower work hardening than SP up to 100 μm approximately. This is opposed to the results found in Ref. [2], where SP or BB was applied only on one surface of a 10 mm thick specimen. In contrast, in the present work, both surfaces of the 5 mm thick specimen were shot peened or ballburnished. Therefore, the higher compressive residual stress distribution (as induced by BB) in one side could influence on the residual stress state in the other side by generating relatively higher balancing tensile residual stress and severe elastic bending [9]. Furthermore, changes of the crystallographic texture close to the surface after different surface treatments may also influence on the elastic anisotropy and elastic lattice strain [26–29]. Consequently, further investigations are needed to study these influences.
E. Maawad et al. / Surface & Coatings Technology 205 (2011) 3644–3650
3649
Fig. 8. An example of near-surface microstructures after (a) shot peening (SP) and (b) ball-burnishing (BB) in Ti–2.5Cu (SHT).
The highest maximum residual stresses were induced by BB in Ti– 2.5Cu (SHT) and Ti–2.5Cu (SHT + A) as listed in Table 2. Furthermore, SP, LSP and USP produced nearly similar maximum residual stresses, if the error bars are taken into account. Although the zero crossing depth could not be measured due to the limited energy range and corresponding penetration depth provided by ED, it seems that LSP produced much deeper compressive layer compared to SP and USP, This is also attributed to the work hardening. The same Almen intensity used in SP and USP could be the reason why the residual stress distributions were nearly similar despite the larger and harder balls used in USP compared to the shots used in SP. This indicates that the residual stress distribution is significantly influenced by the Almen intensity or the kinetic energy used in the present condition rather than the peening media parameters. Indeed, compressive residual stress has little effect on fatigue crack nucleation, but can drastically retard fatigue micro-crack propagation [24,30,31] by overcoming the crack opening stress. In conclusion, the greater amount and penetration depths of both work hardening and compressive residual stresses after BB as opposed to SP are thought to be the reason for the more pronounced improvement of the high cycle fatigue (HCF) strength of Ti–2.5Cu after BB [2]. The influence of surface layer characteristics after LSP or USP on the HCF performance of Ti–2.5Cu is currently being investigated. 4. Summary
Fig. 9. Residual stress–depth distribution in (a) Ti–2.5Cu (SHT) and (b) Ti–2.5Cu (SHT+ A) after various surface treatments (SP = shot peening, BB= ball-burnishing, LSP = laser shock peening, USP = ultrasonic shot peening, SHT = solid solution heat treatment, and A = aging).
2. the stress distributions in SP and USP Ti–2.5Cu are nearly the same, while LSP produced relatively much deeper compressive layer, 3. BB produced a deep layer of high compression compared to SP, USP and LSP, 4. fewer impacts by larger bearing ball (Ø1.5 mm) in USP process resulted in lower cold worked surface layer, and 5. LSP produced the lowest work hardening close to the surface due to the stress created by shock wave propagation rather than cold work as in SP. This could lead to a thermal stability of the residual stress.
Table 2 Residual stress component (parallel to rolling direction) after various surface treatments. Ti–2.5Cu (SHT)
Work hardening and residual stress–depth distributions close to the surface as well as surface roughness of mechanically surface treated Ti–2.5Cu were investigated. The results show that: 1. Different tensile properties and hardness of Ti–2.5Cu (SHT) and Ti– 2.5Cu (SHT+A) result in different residual stress distributions after surface treatments,
SP USP LSP BB
Ti–2.5Cu (SHT + A)
RSsurface (MPa)
RSmax (MPa)
RSsurface (MPa)
RSmax (MPa)
− 460 − 415 − 330 − 665
− 600 − 560 − 550 − 710
− 465 − 435 − 390 − 730
− 625 − 575 − 650 − 850
RSsurface = residual stress at the surface (measured by laboratory X-ray diffraction). RSmax = maximum residual stress (measured by energy-dispersive X-ray diffraction).
3650
E. Maawad et al. / Surface & Coatings Technology 205 (2011) 3644–3650
Acknowledgements The authors are indebted to the German Research Foundation (DFG) for financial support through BR961/5-2 and WA 692/32-2. This work was additionally supported by the European CommunityResearch Infrastructure Action under the FP6 “Structuring the European Research Area” Programme (through the Integrated Infrastructure Initiative “Integrating Activity on Synchrotron and Free Electron Laser Science — Contract R II 3-CT2004-506008”). We would like to thank Mr. Holger Polanetzki of MTU Aero Engines GmbH, Munich, for his help to carry out the USP process. References [1] E. Maawad, S. Yi, H.-G. Brokmeier, L. Wagner, in: K. Tosha (Ed.), Proc. of the 10th Int. Conference on Shot Peening, 2008, p. 499, Tokyo, Japan. [2] E. Maawad, H.-G. Brokmeier, M. Hofmann, Ch. Genzel, L. Wagner, Mater. Sci. Eng. A 527 (2010) 5745. [3] J.-E. Masse, G. Barreau, Surf. Coat. Technol. 70 (1995) 231. [4] X.C. Zhang, Y.K. Zhang, J.Z. Lu, F.Z. Xuan, Z.D. Wang, S.T. Tu, Mater. Sci. Eng. A 527 (2010) 3411. [5] R.K. Nalla, I. Altenberger, U. Noster, G.Y. Liu, B. Scholtes, R.O. Ritchie, Mater. Sci. Eng. A 355 (2003) 216. [6] Y. Sano, M. Obata, T. Kubo, N. Mukai, M. Yoda, K. Masaki, Y. Ochi, Mater. Sci. Eng. A 417 (2006) 334. [7] S.K. Cheong, D.S. Lee, J.H. Lee, in: K. Tosha (Ed.), Proc. of the 10th Int. Conference on Shot Peening, 2008, p. 494, Tokyo, Japan.
[8] J.-M. Duchazeaubeneix, V. Desfontaine, S. Forgues, B. Labelle, in: K. Tosha (Ed.), Proc. of the 10th Int. Conference on Shot Peening, 2008, p. 401, Tokyo, Japan. [9] M. Turski, S. Clitheroe, A.D. Evans, C. Rodopoulos, D.J. Hughes, P. Withers, J. Appl. Phys. A 99 (2010) 549. [10] I. Nikitin, B. Scholtes, H.J. Maier, I. Altenberger, Scr. Mater. 50 (2004) 1345. [11] N.R. Tao, Z.B. Wang, W.P. Tong, M.L. Sui, J. Lu, K. Lu, Acta Mater. 50 (2002) 4603. [12] A.L. Ortiz, J.-W. Tian, L.L. Shaw, P.K. Liaw, Scr. Mater. 62 (2010) 129. [13] L.L. Shaw, J.-W. Tian, A.L. Ortiz, K. Dai, J.C. Villegas, P.K. Liaw, R. Ren, D.L. Klarstrom, Mater. Sci. Eng. A 527 (2010) 986. [14] M.G. Moore, W.P. Evans, SAE Trans. 66 (1958) 340. [15] Ch. Genzel, C. Stock, W. Reimers, Mater. Sci. Eng. A 372 (2004) 28. [16] E. Kröner, Z. Phys. 151 (1958) 504. [17] T. Ungár, Mater. Sci. Eng. 14219 (2000) 1. [18] T. Ungár, Scr. Mater. 51 (2004) 777. [19] J. Pelleg, E. Elish, J. Vac. Sci. Technol. A20 (2001) 745. [20] R. Flitney, Seals and Searling Handbook, fifth ed., 2007, p. 478, chapter 7, General information. [21] B.P. Fairand, A.H. Clauer, B.A. Wilcox, Metall. Trans. A 8A (1977) 119. [22] M. Gerland, M. Hallouin, Mater. Sci. 29 (1994) 345. [23] C.S. Montross, T. Wei, L. Ye, G. Clark, Y.-W. Mai, Int. J. Fatigue 24 (2002) 1021. [24] L. Wagner, Meter. Sci. Eng. A 263 (1999) 210. [25] R. Stephens, A. Fatemi, R.R. Stephens, H.O. Fuchs, Metal Fatigue in Engineering, second ed., Wiley-Interscience, New York, 2001, p. 43. [26] C.M. Brakman, J. Appl. Cryst. 16 (1983) 325. [27] C.M. Brakman, P. Penning, Acta Crystallogr. A44 (1988) 163. [28] Y.D. Wang, R. Lin Peng, X.-L. Wang, R.L. McGreevy, Acta Mater. 50 (2002) 1717. [29] B. Clausen, T. Lorentzen, T. Leffers, Acta Mater. 46 (1998) 3087. [30] L. Wagner, in: J.K. Gregory, H.J. Rack, D. Eylon (Eds.), Surface Performance of Titanium Alloys, TMS, City, 1997, p. 199. [31] P.S. Prevéy, in: M.J. Shepard, P.R. Smith (Eds.), Proc. 6th National Turbine Engine High Cycle Fatigue (HCF) Conference, AFRL/MLLMP, Jacksonville, FL, 2001, p. 1.
Contents lists available at ScienceDirect
Surface & Coatings Technology j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / s u r f c o a t
Investigation on the surface and near-surface characteristics of Ti–2.5Cu after various mechanical surface treatments E. Maawad a,⁎, H.-G. Brokmeier a,b, L. Wagner a, Y. Sano c, Ch. Genzel d a
Institute of Materials Science and Engineering, Clausthal University of Technology, Agricolastr. 6, D-38678 Clausthal-Zellerfeld, Germany Helmholtz-Zentrum Geesthacht Centre for Materials and Coastal Research, D-21502 Geesthacht, Germany Toshiba Corporation, 8 Shinsugita-cho Isogo-ku Yokohama 235-8523, Japan d Helmholtz-Zentrum Berlin (BESSY), Albert-Einstein-Str. 15, D-12489 Berlin, Germany b c
a r t i c l e
i n f o
Article history: Received 14 September 2010 Accepted in revised form 5 January 2011 Available online 12 January 2011 Keywords: Mechanical surface treatment Shot peening Ball-burnishing Laser shock peening Ultrasonic shot peening Residual stress
a b s t r a c t As a result of the variety of mechanical surface treatments, studies of surface and near-surface characteristics are becoming increasingly important in a variety of industrial fields. This is to achieve more gains by balancing between optimum conditions and costs. In the present study, shot peening (SP), ball-burnishing (BB), laser shock peening (LSP) and ultrasonic shot peening (USP) processes were performed on the α-titanium alloy Ti– 2.5Cu after two different heat treatments. The influence of such surface treatments on the surface and nearsurface characteristics, such as residual stress, work hardening or dislocation density and surface roughness, was studied. The depth profiles of residual stress and full width at half maximum (FWHM) were obtained by using energy-dispersive X-ray diffraction. Results show that the BB process produced the highest and deepest maximum residual stress and LSP produced the lowest work hardening close to the surface. © 2011 Elsevier B.V. All rights reserved.
1. Introduction Mechanical surface treatments are widely used to prolong fatigue life of many engineering components in the industry. This is attributed to the changes of surface and near-surface characteristics by producing plastic deformation and compressive residual stress [1]. Shot peening (SP) is most commonly used due to its low cost and can be used on small or large areas depending on requirements. However, there are several techniques which could be competitive such as ballburnishing (BB), laser shock peening (LSP) and ultrasonic shotpeening (USP). BB leads to a smoother surface and deeper compressive residual stress than SP process [2]. Disadvantages of this process are difficulty in tooling development and poor access to some locations, for example, on engine components. Compared to the SP process, LSP can produce deeper compressive residual stresses [3]. Therefore, it is used mainly for increasing fatigue strength [4]. Furthermore, the compressive residual stresses are produced with less work hardening than that with SP, allowing for less thermal relaxation of these stresses when the part is subjected to high temperatures [5]. In addition, the process, effect and application
Abbreviations: SP, shot peening; BB, ball-burnishing; LSP, laser shock peening; USP, ultrasonic shot peening. ⁎ Corresponding author. IWW, Agricolastr. 6, D-38678 Clausthal-Zellerfeld, Germany. Tel.: + 49 5323 72 2758; fax: + 49 5323 72 2766. E-mail address: [email protected] (E. Maawad). 0257-8972/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.surfcoat.2011.01.001
of LSP without coating were studied [6]. It was concluded that LSP without coating prevents stress corrosion cracking (SCC) and prolongs fatigue life. On the other hand, the major disadvantage of LSP is that it currently requires a laser system specially designed for production runs with high average power and large pulse energy. Recently, USP has been studied to improve the durability of structural parts [7]. In USP, unlike SP where the shot is propelled by compressed air, the shot is energized by a sonotrode vibrating at ultrasonic frequency. The main characteristic of such process could give a smooth surface compared to SP because of using polished bearing ball with a large diameter. Furthermore, it can give much energy in short time due to using high frequency [8]. There are several researches studying the influence of various surface treatments on the surface and near-surface characteristics in different alloys. Turski et al. [9] investigated on the surface roughness, microstructure, level of plastic work and through thickness residual stress distribution in austenitic steel AISI 304 after different surface treatments. It was concluded that the samples show compressive stresses to a significantly greater depth for the LSP, ultrasonic impact treatment and water jet cavitation peening samples compared to the more conventional SP treated sample. The influences of deep rolling (DR) and LSP on the fatigue behavior of Ti–6Al–4V and austenitic steel AISI 304 were investigated by R.K. Nalla et al. [5] and I. Nikitin et al. [10], respectively. It was found that DR induced larger near-surface compressive residual stresses and degree of work hardening compared to LSP. Recently, a new family of surface severe plastic deformation processes (SPD) using high energy balls, widely known as the surface
E. Maawad et al. / Surface & Coatings Technology 205 (2011) 3644–3650
mechanical attrition treatment (SMAT), has been developed to induce in workpieces the desired nanocrystalline surface layer [11]. SMAT produces a thicker nanocrystalline surface layer, a deeper plastically deformed region with greater abundance of twins, faults and dislocations, and a thicker surface region with larger residual compressive stresses than SP [12,13]. These differences make SMAT more effective in improving the fatigue properties than SP. Ti–2.5Cu has many applications including sheets, forgings and extrusions for fabricating components such as bypass ducts of gas-turbine engines as well as airframes in aeronautical industry. As a result of the variety of material responses to various surface treatments, the main aim of the present work is to study the influence of SP, BB, USP or LSP on the surface and near-surface characteristics in Ti–2.5Cu, such as residual stress, surface roughness and work hardening. 2. Material and methods 2.1. Material The Ti–2.5Cu alloy was received as a 10 mm thick rolled plate. Four specimens were cut from this plate with a dimension of 20 × 20 × 5 mm3 perpendicular to the rolling direction followed by two different heat treatments. Solution heat treatment (SHT) was done at 805 °C for 1 h followed by water-quenching. The other was SHT followed by double aging (SHT + A) by annealing the material at 400 °C for 8 h and then annealing at 475 °C for 8 h. The microstructures of Ti–2.5Cu after the two conditions are illustrated in Fig. 1.
3645
Table 1 Tensile properties and hardness of Ti–2.5Cu.
E (GPa) σy (MPa) UTS (MPa) El (%) εf = ln (Ao/Af)
SHT
SHT+A
105 520 610 28 0.62
110 685 770 16.4 0.57
E = modulus of elasticity, σy = yield stress, UTS = ultimate tensile stress, EL= elongation, and εf = true fracture strain.
It was observed that both microstructures consist of α grains and stringers of the eutectoid component α + Ti2Cu. Threaded cylindrical tensile specimens were machined having a gage length and a diameter of 20 and 4 mm, respectively. Tensile properties and hardness of Ti–2.5Cu with different heat treatments are listed in Table 1. Ti–2.5Cu (SHT + A) shows a pronounced increase in modulus of elasticity, yield strength, ultimate tensile stress and hardness due to the precipitation hardening influence of Ti2Cu. In contrast, the ductility decreased after aging. 2.2. Mechanical surface treatments The specimens were mechanically surface treated on both surfaces with 20 × 20 mm2 by applying SP, BB, LSP or USP. SP was performed at TU Clausthal in Germany using cast steel (S330) having an average shot diameter of 0.8 mm and a hardness of 460 HV. While USP was applied at MTU Aero Engines in Germany using 100Cr6 bearing steel balls with a diameter of 1.5 mm and a hardness of 700–800 HV. The main reason for using 0.8 mm shots in SP and 1.5 mm balls in USP is their higher mass and therefore the higher compressive residual stresses that can be achieved. Peening was performed to roughly 100% coverage in SP and USP using the same Almen intensity of 0.20 mmA. The influence of the different peening media parameters with the same Almen intensity used in SP and USP, such as speed, diameter and hardness, on the surface and near-surface characteristics was considered when discussing the experimental results. Furthermore, Ti–2.5Cu specimens were ball-burnished at TU Clausthal using a conventional lathe and a hydrostatic tool by which a hard metal ball (Ø6 mm) is pressed with a pressure of 300 bar onto the specimen surface. LSP was carried out at Toshiba Corp. in Japan using a compact Q-switched and frequency-doubled Nd:YAG laser with a power density of 5 GW cm− 2, a pulse duration of 8 ns and a spot size of 0.8 mm in diameter. Water was used as a tamping material limiting the thermal expansion of plasma gas. LSP was done without coating having some characteristics which were mentioned elsewhere [6]. 2.3. Residual stress and full width at half maximum
Fig. 1. Microstructure of (a) Ti–2.5Cu (SHT) and (b) Ti–2.5Cu (SHT + A), average grain size ≈ 20 μm (SHT = solid solution heat treatment, A = aging).
Residual stress measurements were performed by hard X-ray diffraction using synchrotron radiation at BESSY-II in Berlin. The characteristic of the used beamline EDDI offers a white X-ray beam with an energy range of 10–80 keV. The primary beam cross-section was set to 0.5 × 0.5 mm2. The angular divergence in the diffracted beam was restricted by a double slit system with apertures of 0.03 × 5 mm2 to Δθ ≤ 0.005°. The scattering angle was chosen to 8° considering the energy of X-ray. Energy-dispersive X-ray diffraction (ED) gives complete diffraction spectra for a fixed detector position. Any Bragg reflection was obtained by a different X-ray energy (wavelength) which means the signal of any reflection belongs to a different depth in the sample as schematically shown in Fig. 2. Due to the limited usable energy range provided by the 7T multipole wiggler which extends from about 10 keV to 80 keV, the maximum information depth for titanium accessible in reflection mode experiments is about 100 μm. In order to get a stress distribution in deeper region,
3646
E. Maawad et al. / Surface & Coatings Technology 205 (2011) 3644–3650
Fig. 2. Scheme showing a relation between energy (E) and penetration depth (τ).
three specimens were prepared. For the two of them, layer removal in a step of 100–150 μm was applied by electropolishing. The correction on the measured residual stress was made for the electropolished specimens using the equation described elsewhere [14]. Residual stresses were evaluated by means of the sin2ψ method, where ψ is the tilting angle, in steps of Δψ = 4° up to 80°. A modified multiwavelength approach [15] for any energy line E(hkl) gives an average penetration depth τ(hkl) (Eq. (1)). τðhklÞ ¼ ðτðhklÞmin þ τðhklÞmax Þ=2
ð1Þ
where τ(hkl)min and τ(hkl)max are the minimum and the maximum penetration depths corresponding to the maximum (ψmin) and minimum (ψmax) tilting angles, respectively. The diffraction elastic constants of alpha-reflections were calculated by the Kröner-model [16]. Residual stresses at the surface were determined by using laboratory X-ray diffraction (D-5000 at TU Clausthal) using Cu-Kαradiation with a wavelength of 1.54 Å. The {21.3}-Bragg peak was used. The moduli of elasticity of 105 GPa for Ti–2.5Cu (SHT) and 110 GPa for Ti–2.5Cu (SHT + A) with Poisson's ratio of 0.33 were used to calculate the diffraction elastic constants (S1 and 1/2 S2). The sin2ψ method was also used to evaluate the surface residual stresses with tilting angles range of −60° to +60° in steps of Δsin2ψ = 0.125. Work hardening or dislocation density distribution can be evaluated by means of X-ray diffraction peak broadening using characteristic parameters of individual peak profiles, especially the full width at half maximum (FWHM) and the integral breadths. It is well known that besides the instrumental contribution, there are two main types of broadening: the size and the strain components. The former one depends on the finite size of the coherent diffraction domains and the latter is caused by any lattice imperfection (point, line or plane defects). The strain field of linear defects, such as dislocations, is of long-range character, therefore their diffraction effects cluster around the fundamental Bragg reflections [17,18]. In the present work, FWHM was also determined by using ED. The (10.3) reflection was used due to its high multiplicity factor (n = 12) and corresponding reliability of the results as well as good statistics. The FWHM of (10.3) reflection was calculated by using the Pseudo-Voigt function [19] as shown in Fig. 3.
Fig. 3. (10.3) peak fitted by Pseudo-Voigt function to define FWHM.
Vickers tester, a nominal force of 100 gf (HV0.1) and a loading time of 10 s. The hardness testers in the Duramin series conform to the standard (DIN EN ISO 6507). The average of three measurements was taken at each depth to construct the hardness–depth profiles. 3. Results and discussions 3.1. Surface roughness The surface roughness (Rz) after the various surface treatments was determined and compared to the electropolished (EP) sample as a reference. As seen in Fig. 4, the surface roughness of shot peened Ti– 2.5Cu is much higher than that of the electropolished reference, while a remarkable improvement on the surface roughness was observed after BB (Fig. 5a) compared to that after SP (Fig. 5c). Furthermore, the surface roughness after USP became smoother than that after SP. As seen in Fig. 5b, wider dimples with smaller amplitude were produced after USP than that after SP (Fig. 5c). The reason of this result was considered that the balls in the case of USP are set into a random motion inside a component-specific peening chamber to act on the component, while the shot stream is nearly perpendicular to the surface in the case of SP. Therefore, the influence of the shot velocity on the indent depth or roughness is more significant in SP compared to that in USP. Furthermore, the bearing balls used for USP have higher spherical accuracy and smoother surface than the shots used in SP process. Contrary to what is expected for LSP with coating, a rough surface with linear furrows appeared after LSP without coating (Fig. 5d). When no protective laser-absorbent coating was used on the specimen, severe surface melting and vaporization would be induced
2.4. Surface roughness The surface roughness of the various conditions was determined by means of an electronic contact (stylus) profilometer instrument (Perpethometer). The average absolute value of the five highest peaks and the five lowest valleys over the evaluation length (Rz) was reported (DIN 4768). The average of three roughness measurements was taken. The parameter (Rz) was used rather than the average roughness comparing all the peaks and valleys to the mean line (Ra). Quite different surfaces could have the same Ra and consequently perform in a different manner [20]. 2.5. Microhardness Microhardness was determined by means of a Struers Duramin tester using a square base pyramid shaped indenter for testing in a
Fig. 4. Surface roughness after various surface treatments (SP = shot peening, BB = ball-burnishing, LSP = laser shock peening, and USP = ultrasonic shot peening).
E. Maawad et al. / Surface & Coatings Technology 205 (2011) 3644–3650
3647
Fig. 5. Images of the mechanically treated surface by (a) BB, (b) USP, (c) SP and (d) LSP (SP = shot peening, BB = ball-burnishing, LSP = laser shock peening, and USP = ultrasonic shot peening).
by LSP with high-power and long-duration laser pulses [21,22]. This can result in resolidified droplets and craters leading to very rough surfaces [23]. To prevent the surface from such damage, laser pulses with much lower energy and shorter duration were used in this study. 3.2. Work hardening/dislocation density The microhardness–depth distributions of Ti–2.5Cu after various mechanical surface treatments are shown in Fig. 6. It was found that the bulk hardness in Ti–2.5Cu (SHT) is 250 HV0.1 approximately (Fig. 6a), while the bulk hardness in Ti–2.5Cu (SHT + A) is about 320 HV0.1 (Fig. 6a). This increase of the hardness is explained by the precipitation hardening influence. The maximum work hardening induced by SP, BB, LSP and USP was observed at the surface which gradually decreased in the near-surface regions as shown in Fig. 6a and b. In both conditions of the heat treatments, it was found that the repeated dimpling at the surface by SP to achieve uniform surface coverage resulted in the highest cold worked layer or dislocation density at the surface. In contrast, LSP produced the lowest work hardening at the surface, although each point on the surface was hit 18 times (laser spot area × irradiation density) by the laser pulse. This is a result of the stress created by shock wave propagation rather than cold work as in SP. This lower magnitude of work hardening after LSP supports the previous work hardening results of laser peened hypoeutectoid steel [3]. In this study, laser beam with a small footprint (Ø0.8 mm) and a lower power density (5 GW cm− 2)
generated a less planar pressure wave to propagate shallower into the specimen producing plastic deformation with a depth of 0.45 and 0.6 mm from the treated surface in the case of SHT and SHT + A, respectively. In contrast, it was found in Ref. [9] that a deeper plastic deformation (2 mm from the treated surface) was produced after LSP with a larger footprint (3 × 3 mm) and a higher power density (10 GW cm− 2) resulting in a higher pressure in 304 austenitic stainless steel. It should be also taken into account that the tensile properties of Ti–2.5Cu and 304 austenitic stainless steel are different when comparing the influence of LSP parameters. This feature of LSP leads to a thermal stability of residual stress close to the surface. It was found that the work hardening induced by BB slightly decreased close to the surface (up to 100 μm in depth) compared to SP, while it increased in deeper region. This is due to less shearing between the ball and the specimen surface while using lubricant as well as due to larger ball size of 6 mm in diameter. The increase of work hardening or dislocation density in the near-surface region retards dramatically fatigue crack nucleation [24]. This is a result of a decrease of dislocation mobility which results in a limited number of slip planes [25] and corresponding slip band intrusions and extrusions. On the other hand, it accelerates fatigue micro-cracks propagation. Furthermore, it was observed that USP produced relatively higher work hardening than LSP and lower than SP and BB. This is a result of fewer impacts by large balls required to achieve the surface coverage. The work hardening behavior close to the surface after these surface treatments was supported by determining FWHM as shown in
3648
E. Maawad et al. / Surface & Coatings Technology 205 (2011) 3644–3650
Fig. 6. Microhardness–depth distribution in (a) Ti–2.5Cu (SHT) and (b) Ti–2.5Cu (SHT + A) after various surface treatments (SP= shot peening, BB= ball-burnishing, LSP= laser shock peening, USP= ultrasonic shot peening, SHT= solid solution heat treatment, and A = aging).
Fig. 7. As seen in Fig. 7b, the FWHM in the virgin region in Ti–2.5Cu (SHT + A) is relatively larger than that in Ti–2.5Cu (SHT). This is explained by the increase of Ti2Cu precipitation as a structural defect. Furthermore, it was observed that the FWHM distributions close to the surface agree with the microhardness distribution. The reason for some slight differences in the penetration depth between the microhardness and FWHM could be explained by the data scattering of microhardness due to the existence of eutectoid components (α + Ti2Cu) in the grain boundaries. In the present study, the contribution of nanocrystalline grains to FWHM was ignored because no significant changes of the grain size (20 μm) were observed close to the surface after the surface treatments. Fig. 8a and b shows the near-surface microstructures after SP and BB which produced higher work hardening in Ti–2.5Cu (SHT), respectively. 3.3. Residual stress The residual stress–depth distributions in Ti–2.5Cu (SHT) and in Ti–2.5Cu (SHT + A) after the various surface treatments are illustrated in Fig. 9a and b, respectively. The response of Ti–2.5Cu after different heat treatments to the various surface treatments and the influence of each surface treatment on the residual stress distribution were discussed. The values of residual stresses at the surface and the maximum compressive residual stresses are listed in Table 2. Obviously, the surface and maximum residual stresses in Ti–2.5Cu (SHT) is relatively lower than that in Ti–2.5Cu (SHT + A). This is explained by higher
Fig. 7. FWHM-depth distribution in (a) Ti–2.5Cu (SHT) and (b) Ti–2.5Cu (SHT + A) after various surface treatments (SP = shot peening, BB = ball-burnishing, LSP = laser shock peening, USP = ultrasonic shot peening, SHT = solid solution heat treatment, and A = aging).
modulus of elasticity and yield strength of Ti–2.5Cu (SHT + A). This is a result of relatively larger potential maxima of the induced residual stress by plastic deformation. It was observed that LSP induced the lowest residual stress at the surface, while USP induced lower residual stress than SP and BB and higher than LSP (see Table 2). This is attributed to the magnitude of work hardening induced at the surface as shown in Figs. 6 and 7. Lower work hardening resulted in lower flow stress and thus lower residual stress. However, it was found that BB produced higher residual stress at the surface compared to SP in both Ti–2.5Cu (SHT) and Ti–2.5Cu (SHT + A) despite the fact that BB produced lower work hardening than SP up to 100 μm approximately. This is opposed to the results found in Ref. [2], where SP or BB was applied only on one surface of a 10 mm thick specimen. In contrast, in the present work, both surfaces of the 5 mm thick specimen were shot peened or ballburnished. Therefore, the higher compressive residual stress distribution (as induced by BB) in one side could influence on the residual stress state in the other side by generating relatively higher balancing tensile residual stress and severe elastic bending [9]. Furthermore, changes of the crystallographic texture close to the surface after different surface treatments may also influence on the elastic anisotropy and elastic lattice strain [26–29]. Consequently, further investigations are needed to study these influences.
E. Maawad et al. / Surface & Coatings Technology 205 (2011) 3644–3650
3649
Fig. 8. An example of near-surface microstructures after (a) shot peening (SP) and (b) ball-burnishing (BB) in Ti–2.5Cu (SHT).
The highest maximum residual stresses were induced by BB in Ti– 2.5Cu (SHT) and Ti–2.5Cu (SHT + A) as listed in Table 2. Furthermore, SP, LSP and USP produced nearly similar maximum residual stresses, if the error bars are taken into account. Although the zero crossing depth could not be measured due to the limited energy range and corresponding penetration depth provided by ED, it seems that LSP produced much deeper compressive layer compared to SP and USP, This is also attributed to the work hardening. The same Almen intensity used in SP and USP could be the reason why the residual stress distributions were nearly similar despite the larger and harder balls used in USP compared to the shots used in SP. This indicates that the residual stress distribution is significantly influenced by the Almen intensity or the kinetic energy used in the present condition rather than the peening media parameters. Indeed, compressive residual stress has little effect on fatigue crack nucleation, but can drastically retard fatigue micro-crack propagation [24,30,31] by overcoming the crack opening stress. In conclusion, the greater amount and penetration depths of both work hardening and compressive residual stresses after BB as opposed to SP are thought to be the reason for the more pronounced improvement of the high cycle fatigue (HCF) strength of Ti–2.5Cu after BB [2]. The influence of surface layer characteristics after LSP or USP on the HCF performance of Ti–2.5Cu is currently being investigated. 4. Summary
Fig. 9. Residual stress–depth distribution in (a) Ti–2.5Cu (SHT) and (b) Ti–2.5Cu (SHT+ A) after various surface treatments (SP = shot peening, BB= ball-burnishing, LSP = laser shock peening, USP = ultrasonic shot peening, SHT = solid solution heat treatment, and A = aging).
2. the stress distributions in SP and USP Ti–2.5Cu are nearly the same, while LSP produced relatively much deeper compressive layer, 3. BB produced a deep layer of high compression compared to SP, USP and LSP, 4. fewer impacts by larger bearing ball (Ø1.5 mm) in USP process resulted in lower cold worked surface layer, and 5. LSP produced the lowest work hardening close to the surface due to the stress created by shock wave propagation rather than cold work as in SP. This could lead to a thermal stability of the residual stress.
Table 2 Residual stress component (parallel to rolling direction) after various surface treatments. Ti–2.5Cu (SHT)
Work hardening and residual stress–depth distributions close to the surface as well as surface roughness of mechanically surface treated Ti–2.5Cu were investigated. The results show that: 1. Different tensile properties and hardness of Ti–2.5Cu (SHT) and Ti– 2.5Cu (SHT+A) result in different residual stress distributions after surface treatments,
SP USP LSP BB
Ti–2.5Cu (SHT + A)
RSsurface (MPa)
RSmax (MPa)
RSsurface (MPa)
RSmax (MPa)
− 460 − 415 − 330 − 665
− 600 − 560 − 550 − 710
− 465 − 435 − 390 − 730
− 625 − 575 − 650 − 850
RSsurface = residual stress at the surface (measured by laboratory X-ray diffraction). RSmax = maximum residual stress (measured by energy-dispersive X-ray diffraction).
3650
E. Maawad et al. / Surface & Coatings Technology 205 (2011) 3644–3650
Acknowledgements The authors are indebted to the German Research Foundation (DFG) for financial support through BR961/5-2 and WA 692/32-2. This work was additionally supported by the European CommunityResearch Infrastructure Action under the FP6 “Structuring the European Research Area” Programme (through the Integrated Infrastructure Initiative “Integrating Activity on Synchrotron and Free Electron Laser Science — Contract R II 3-CT2004-506008”). We would like to thank Mr. Holger Polanetzki of MTU Aero Engines GmbH, Munich, for his help to carry out the USP process. References [1] E. Maawad, S. Yi, H.-G. Brokmeier, L. Wagner, in: K. Tosha (Ed.), Proc. of the 10th Int. Conference on Shot Peening, 2008, p. 499, Tokyo, Japan. [2] E. Maawad, H.-G. Brokmeier, M. Hofmann, Ch. Genzel, L. Wagner, Mater. Sci. Eng. A 527 (2010) 5745. [3] J.-E. Masse, G. Barreau, Surf. Coat. Technol. 70 (1995) 231. [4] X.C. Zhang, Y.K. Zhang, J.Z. Lu, F.Z. Xuan, Z.D. Wang, S.T. Tu, Mater. Sci. Eng. A 527 (2010) 3411. [5] R.K. Nalla, I. Altenberger, U. Noster, G.Y. Liu, B. Scholtes, R.O. Ritchie, Mater. Sci. Eng. A 355 (2003) 216. [6] Y. Sano, M. Obata, T. Kubo, N. Mukai, M. Yoda, K. Masaki, Y. Ochi, Mater. Sci. Eng. A 417 (2006) 334. [7] S.K. Cheong, D.S. Lee, J.H. Lee, in: K. Tosha (Ed.), Proc. of the 10th Int. Conference on Shot Peening, 2008, p. 494, Tokyo, Japan.
[8] J.-M. Duchazeaubeneix, V. Desfontaine, S. Forgues, B. Labelle, in: K. Tosha (Ed.), Proc. of the 10th Int. Conference on Shot Peening, 2008, p. 401, Tokyo, Japan. [9] M. Turski, S. Clitheroe, A.D. Evans, C. Rodopoulos, D.J. Hughes, P. Withers, J. Appl. Phys. A 99 (2010) 549. [10] I. Nikitin, B. Scholtes, H.J. Maier, I. Altenberger, Scr. Mater. 50 (2004) 1345. [11] N.R. Tao, Z.B. Wang, W.P. Tong, M.L. Sui, J. Lu, K. Lu, Acta Mater. 50 (2002) 4603. [12] A.L. Ortiz, J.-W. Tian, L.L. Shaw, P.K. Liaw, Scr. Mater. 62 (2010) 129. [13] L.L. Shaw, J.-W. Tian, A.L. Ortiz, K. Dai, J.C. Villegas, P.K. Liaw, R. Ren, D.L. Klarstrom, Mater. Sci. Eng. A 527 (2010) 986. [14] M.G. Moore, W.P. Evans, SAE Trans. 66 (1958) 340. [15] Ch. Genzel, C. Stock, W. Reimers, Mater. Sci. Eng. A 372 (2004) 28. [16] E. Kröner, Z. Phys. 151 (1958) 504. [17] T. Ungár, Mater. Sci. Eng. 14219 (2000) 1. [18] T. Ungár, Scr. Mater. 51 (2004) 777. [19] J. Pelleg, E. Elish, J. Vac. Sci. Technol. A20 (2001) 745. [20] R. Flitney, Seals and Searling Handbook, fifth ed., 2007, p. 478, chapter 7, General information. [21] B.P. Fairand, A.H. Clauer, B.A. Wilcox, Metall. Trans. A 8A (1977) 119. [22] M. Gerland, M. Hallouin, Mater. Sci. 29 (1994) 345. [23] C.S. Montross, T. Wei, L. Ye, G. Clark, Y.-W. Mai, Int. J. Fatigue 24 (2002) 1021. [24] L. Wagner, Meter. Sci. Eng. A 263 (1999) 210. [25] R. Stephens, A. Fatemi, R.R. Stephens, H.O. Fuchs, Metal Fatigue in Engineering, second ed., Wiley-Interscience, New York, 2001, p. 43. [26] C.M. Brakman, J. Appl. Cryst. 16 (1983) 325. [27] C.M. Brakman, P. Penning, Acta Crystallogr. A44 (1988) 163. [28] Y.D. Wang, R. Lin Peng, X.-L. Wang, R.L. McGreevy, Acta Mater. 50 (2002) 1717. [29] B. Clausen, T. Lorentzen, T. Leffers, Acta Mater. 46 (1998) 3087. [30] L. Wagner, in: J.K. Gregory, H.J. Rack, D. Eylon (Eds.), Surface Performance of Titanium Alloys, TMS, City, 1997, p. 199. [31] P.S. Prevéy, in: M.J. Shepard, P.R. Smith (Eds.), Proc. 6th National Turbine Engine High Cycle Fatigue (HCF) Conference, AFRL/MLLMP, Jacksonville, FL, 2001, p. 1.