- Email: [email protected]
Effects of peening on mechanical properties in friction stir welded 2195 aluminum alloy joints
Materials Science and Engineering A 492 (2008) 168–176
Contents lists available at ScienceDirect
Materials Science and Engineering A journal homepage: www.elsevier.com/locate/msea
Effects of peening on mechanical properties in friction stir welded 2195 aluminum alloy joints Omar Hatamleh ∗ Structures Branch, NASA – Johnson Space Center, Houston, TX 77058, United States
a r t i c l e
i n f o
Article history: Received 19 December 2007 Received in revised form 5 March 2008 Accepted 6 March 2008 Keywords: Friction stir welding Laser peening Shot peening Mechanical properties 2195
a b s t r a c t The effects of surface treatment techniques like laser and shot peening on the mechanical properties were investigated for friction stir welded 2195 aluminum alloy joints. The loading in the tensile specimens was applied in a direction perpendicular to the weld direction. The peening effects on the local mechanical properties through the different regions of the weld were characterized using a digital image correlation technique assuming an iso-stress condition. This assumption implies that the stress is uniform over the cross-section and is equal to the average stress. The surface strain and average stress were used giving an average stress–strain curve over the region of interest. The extension of the iso-stress assumption to calculate local stress–strain curves in surface treated regions is a novel approach and will help to understand and improve the local behavior at various regions across the weld resulting in a sound welding process. The surface and through-thickness residual stresses were also assessed using the X-ray diffraction and the contour methods. The laser peened samples displayed approximately 60% increase in the yield strength of the material. In contrast, shot peening exhibited only modest improvement to the tensile properties when compared to the unpeened FSW specimens. The result that laser peening is superior to shot peening because of the depth of penetration is original since this superiority has not been presented before regarding mechanical properties performance. Published by Elsevier B.V.
1. Introduction High strength aluminum alloys are used extensively in the aerospace industries due to their strength and light weight. However, these aerospace aluminum alloys are traditionally considered to be unweldable using conventional fusion welds [1] because of the dentretic structure formed with the fusion welding [2]. Dentretic structure typical of a fusion weld joint can seriously degrade the mechanical properties of the welded joint [3]. Since its invention by the Welding Institute in 1991 [4], friction stir welding (FSW) has emerged as a promising solid state process with encouraging results. This welding technique has potential significant application in different fields including the automotive and aerospace industries [5], and has resulted in welded joints being used in critical load bearing structures and structurally demanding applications [6]. The FSW technique employs a non-consumable cylindrical pin that rotates at high speeds and is then plunged into butting edges of the work pieces to be joined [2]. This process transforms the metal into a plastic state at a temperature below the melting temperature of the material [7], and then mechanically stirs the
∗ Tel.: +1 281 483 0286; fax: +1 281 244 5918. E-mail address: [email protected]. 0921-5093/$ – see front matter. Published by Elsevier B.V. doi:10.1016/j.msea.2008.03.017
material together under pressure to form a welded joint. Since FSW is considered a solid state welding process, significant differences compared to conventional welding may be expected in terms of heat affected zone (HAZ) size and microstructure, and residual stress fields around the weld [8]. The FSW consists of a nugget, the thermo-mechanical affected zone (TMAZ), and a heat affected zone. FSW takes place at a low temperature level compared to fusion welding; therefore, residual stresses may be considerably less than those in fusion welds. Nevertheless, the rigid clamping configuration required to clamp the parts during the FSW process along with the heating cycle the material experiences during welding, can result in higher residual stresses in the weld [9,10]. These residual stresses, along with the reduction in properties from the welding process are likely to affect the mechanical properties and therefore influence the in-service performance of structural components [11]. The weld strength in some cases can be improved by post-weld heat treatment. However, this is not always an option in welded structural components. Consequently, laser shock peening was investigated as a mean for improving the tensile properties in FSW. Laser shock peening is an efficient surface treatment technique that has been proven to be capable of improving the fatigue and mechanical properties of a number of metallic materials [12–15]. The laser peening process (Fig. 1) provides high energy laser pulses
O. Hatamleh / Materials Science and Engineering A 492 (2008) 168–176
Fig. 1. Laser peening process.
that are fired at the surface of a metal coated with an ablative layer, and covered with a thin layer of transparent material (usually water). As the surface layer ablates, the water confines the evaporating material and the vapor is ionized to form plasma. The subsequent laser energy is absorbed by the plasma and generates a high intensity shock wave that impinges on the metal. When the peak pressure of the shock wave is greater than the dynamic yield strength of the material, it produces extensive plastic deformation in the metal. The laser peening process generates a compressive residual stress at the surface that can be significantly deeper than for conventional shot peening [15–17]. The actual depths of the laser-peening induced stresses will vary depending on the material properties of the peened parts and the processing conditions chosen [18]. Several investigations [5,19–20] have assessed the local tensile properties on FSW, however none have assessed the effects of laser peening on the tensile properties on FSW. Since the overall tensile stress of the weld is connected to the local mechanical properties in the weld region; understanding and improving the local behavior at various regions across the weld will result in a sound welding process. In this study, shot peening and laser peening were used to introduce compressive residual stresses into FSW AA 2195. The influence of the peening on the mechanical properties of the FSW specimens were characterized and assessed. Since conventional transverse tensile testing only provides the overall strain experienced by the sample, the local strains and equivalent tensile properties were evaluated at different regions of the weld using a digital image correlation (DIC) system. 2. Experimental setup Aluminum alloy (AA) 2195-T8 plates with a 1.25 cm thickness were used in this research. The aluminium plates were welded together by a single pass using a rotational speed of 300 rpm in the counterclockwise direction and a translation speed of 15 cm/min. The welds were oriented parallel to the rolling direction. The mechanical properties for the base material are as shown in Table 1. Both the root and the crown sides of the weld were tested to evaluate the quality of the weld. The samples were inspected visually afterward with no crack indications revealed. The tensile specimens were either shot-peened or laserpeened. Unpeened FSW samples were also tested and used Table 1 Tensile properties for the as received AA 2195-T8 Material
0.2% Yield stress (MPa)
Ultimate strength (MPa)
2195-T8
503
537
169
as a baseline. To optimize the shot peening process, Peenstress, which is a software developed at the Metal Improvement Company, was used. Based on this evaluation, the samples were shot-peened with 0.059 mm glass beads, with an Almen intensity of 0.008–0.012 A and a 200% coverage rate. The impinging shots were fired at the surface of the specimens at an angle in order to avoid collision with the rebounding beads. For the laser peening process, different peening layers were used in this study in an effort to identify the optimum number of peening layers capable of producing superior tensile properties. Prior to the laser peening process, the surface of the specimens intended for peening was covered with a 0.22 mm thick aluminium tape, which was replaced between layers of peening. A 1 mm thick laminar layer of flowing water was used as a tamping layer. The laser peening was applied using a square laser spot with a laser power density of 5 GW/cm2 and 18 ns in duration. The spots within a layer were overlapped 3%. All four sides of the gauge section in the specimens were peened using the same conditions. The surface residual stresses were measured using the X-ray diffraction (XRD) technique. In XRD, the strain in the crystal lattice is measured assuming that the crystal lattice is linearly distorted. The atomic spacing (d) between crystallographic planes that are equal will vary consistently with their psi ( ) angle, where the angle is defined as the “angle between the surface normal and the normal to the crystallographic planes from which the X-ray peak is diffracted” [21]. Therefore, to determine the magnitude of residual stresses, the lattice strains are assessed in various directions and a plot of sin2 vs. εø is derived (where ø is the angle between a reference direction and the direction of stress measurement in the plane). εø is the strain in the ø and directions defined by [22]: ε =
1+v ( sin2 E
)−
v (1 + 2 ) E
(1)
where : Poisson’s ratio; ø : surface stress at an ø angle with a principal stress direction; E: modulus of elasticity; 1 , 2 : principal stresses. Then from the sin2 vs. εø plot, residual stresses are established through the following relation: =
mE 1+v
(2)
where m = slope of the sin2 vs. εø plot. Residual stress were acquired using a Philips X’Pert PW3040 MRD X-ray diffractometer, equipped with a pole figure goniometer, operating at 40 kV and 45 mA, and employing Ni filtered Cu K-alpha radiation. The measurements were taken using 2 scans from 77◦ to 79◦ , with 0.01◦ per step and 1 s per step, (3 1 1) peak positions at 10 different tilt angles. The through-thickness residual stresses for the FSW sample were measured using the contour method [23] on the plane outlined in Fig. 2. To perform the technique, the specimen was cut along the measurement plane with an EDM wire. In order to minimize movement during the cutting process, the specimen was fixed to a rigid backing plate. The deformed surface shape resulting from the relaxed residual stresses was measured on both cutting surfaces using a coordinate measuring machine (CMM). The displacements from both cutting surfaces were then averaged and noise in the measurements was filtered from average displacements by fitting to a smooth analytical surface. Finally, the original residual stresses were calculated from the measured contour using a finite element model (FEM). Tensile testing was performed at room temperature on a 200 kN servo-hydraulic universal testing machine using a constant crosshead speed of 0.1 mm/min. The transverse tensile specimens
170
O. Hatamleh / Materials Science and Engineering A 492 (2008) 168–176
Fig. 3. Metallographic cross-section of 2195 friction stir weld.
Fig. 2. Tensile sample used for testing.
tensile test. The local strain data was mapped to the corresponding global stress levels by assuming that the transversely loaded FSW specimens were considered a composite material loaded in an iso-stress configuration [19,20]. Using this assumption, local constitutive stress–strain relationships were obtained. 3. Results
consisted of conventional dog bone coupons and were 20 cm long with a gage length of 8.5 cm and a gage width of 1.25 cm (Fig. 2). The coupons were oriented such that the weld was in the center of the specimen and the load was applied perpendicular to the weld direction. Mechanical properties obtained in the transverse tensile test of the FSW weld generally represent the weakest region of the weld. In that configuration, the elongation constitutes an average strain over the whole gage length, which includes the different weld regions. This, in return, does not provide an insight into the correlation between the intrinsic tensile properties and localized microstructure [9]. The local tensile properties at the different regions of the weld were measured using a digital image correlation system. DIC is a powerful system for inspecting complex materials and geometries for their deformation and strain during loading. This tool is capable of performing full-field, non-contact strain measurements. The strain is measured using a three-dimensional video correlation methods and high-resolution digital cameras. A random or regular pattern with good contrast is applied to the surface of the test object and is deformed along with the object. As the specimen is deformed under load, the deformation is recorded by the cameras and evaluated using digital image processing. The initial image processing defines a set of unique correlation areas known as macro-image facets, typically 5–20 pixels across. These facets are then tracked in each successive image with sub-pixel accuracy [24,25]. Therefore, the intrinsic tensile properties for various locations across the weld zone were also characterized by the
3.1. Weld microstructure and hardness The specimen used for metallographic investigation was cut and sectioned in a direction normal to the welding direction and then subjected to several successive steps of grinding and polishing before etching. The micrographs showing different regions of the weld and the corresponding grain sizes for 2195 are shown in Figs. 3 and 4. The FSW sample showed no evidence of porosity or other kinds of defects. The weld nugget seems wider on the crown region of the weld because of the upper surface contact with the tool shoulder. The cross-section also revealed the classical formation of the elliptical “onion rings” structure in the center of the weld. The structure at the nugget is fine and equiaxed with grain sizes significantly smaller than the parent material grain due to the higher temperature and extensive plastic deformation. It was also noted from Fig. 4 that the grains on the top were relatively larger compared to the bottom grains. That was attributed to the heat sink effect of the backing plate. The grain structure at the TMAZ region for the 2195 was elongated with some considerable distortions due to the mechanical action from the welding tool. The HAZ is unaffected by the mechanical effects from the welding tool and the grain structure in that region resembles the parent material grain structure. Kumar et al. and Niskanen et al. [26,27] investigated the precipitates in 2195 and indicated that two main precipitates of 2195 are T1
Fig. 4. Grain sizes at the top and bottom of the weld.
O. Hatamleh / Materials Science and Engineering A 492 (2008) 168–176
171
Fig. 5. The microhardness profile across the weld region for FSW 2195.
(Al2CuLi) and (Al2Cu), which appears at dislocations, subgrains, and grain boundaries. Meanwhile, equivalent phase T2 (Al6CuLi3) precipitates along grain boundaries. During the precipitation of , T1, and T2, a Cu and Li depleted precipitate free zone is formed along grain boundaries and subgrain boundaries due to the absorption of Cu and Li atoms to T1, T2, and phase [26,27]. It is found that T1 precipitates within the grains and coarse equivalent T2 exists at the grain boundaries. To assess the hardness across the weld region, hardness measurements shown in Fig. 5 were taken using a microhardness machine. Softening was noted throughout the weld zone, probably due to coarsening and dissolution of strengthening precipitates induced by the thermal cycle of the FSW. For example, Li et al. and Oertelt et al. [28,29] reported that in 2195 the strengthening precipitates like T1 were no longer present in the weld nugget. The T1 is considered to be the primary strengthening precipitate in the 2195 alloy [30]. Away from the weld nugget, hardness levels increased with increasing distance as precipitation hardening became more effective. It was also noticed from the hardness profiles that the softened region at the top of the weld was relatively wider than the one on the bottom. That is attributed to the specimen thermal characteristics during welding. The top region of the weld generally exhibits higher temperatures compared to the bottom region due to the heating from the tool shoulder, while the bottom surface is in contact with the backing plate, which acts as a heat sink. The lower heat input at the bottom of the plate can significantly reduce the extent of metallurgical transformations, such as re-precipitation and coarsening of precipitates, that take place during welding.
by the shot peening. The surface residual stresses at different distances in the weld nugget region are outlined in Fig. 7. The stresses are shown for the various peening conditions used in this study. In general, all the peening techniques resulted in high compressive stresses, however, shot peening produced the highest surface compressive residual stresses. It was also noticed that the residual stresses in the longitudinal and transverse direction did not vary by an appreciable amount. The through thickness residual stresses obtained by means of the contour method are shown in Figs. 8–10. The magnitude of the measured subsurface residual stresses in the weld were relatively small as shown in Fig. 8 and is mainly attributed to the small size of the specimen that was cut from the plate after welding. When a small specimen is machined from a large plate, a significant amount of the residual stresses they contain are relaxed and reduced in magnitude. Although the shot peened samples exhibited high magnitudes of compressive residual stresses on the surface as illustrated in Figs. 6 and 7, no appreciable subsurface compressive residual stresses were noticed in Fig. 9. The laser peened sample processed with a single layer is shown in Fig. 10. Laser peening
3.2. Residual stress The surface residual stresses at the center of the weld obtained through X-ray diffraction are shown in Fig. 6. The residual stresses for the unpeened specimens were in compression for this region of the weld and the highest compressive residual stresses on the surface were obtained by shot peening, due to the cold work action on the surface. It was also noted that laser peening the samples with one layer did not significantly change the surface stresses at this location compared to the unpeened ones, while the six layers of peening produced residual stresses around 60% of those obtained
Fig. 6. Residual stresses at the center of the weld.
172
O. Hatamleh / Materials Science and Engineering A 492 (2008) 168–176
Fig. 7. Residual stresses for different peening conditions across the weld.
Fig. 8. Two-dimensional map of the measured residual stresses in the unpeened FSW sample.
weld nugget exhibited the lowest tensile properties when compared to other locations across the weld. This may be attributed to the fact that the original structure in this region is over-aged and there is not enough solute left in the material. The transverse tensile properties in welded samples were strongly related to hardness profile in the welds, as shown in Fig. 5. This trend was also reported by Sato et al. and Liu et al. [31,32] for FSW that had no defects. This region of the weld will be relatively ineffective in inhibiting dislocation motion and the resultant strain localization in the softened area of the weld will result in lower mechanical properties. In addition to these microstructural changes, the residual stresses in the weld region may have had an effect on the tensile properties in the samples.
Fig. 9. Two-dimensional map of the measured residual stresses in the shot peened FSW sample.
introduced significant levels of compressive residual stresses near the surface. These stresses had a drastic effect on the overall residual stress state and produced high levels of compensating tensile residual stress at the center of the specimen. 3.3. Mechanical properties The tensile properties at different locations across the weld region are illustrated in Fig. 11. From the graph, it is clear that the
Fig. 10. Two-dimensional map of the measured residual stresses in the laser peened (single layer) FSW sample.
Fig. 11. Tensile properties at different regions of the weld for an unpeened FSW 2195 AA.
O. Hatamleh / Materials Science and Engineering A 492 (2008) 168–176
173
Fig. 12. Tensile properties at the weld nugget under different peening conditions.
Fig. 13. Tensile properties at the advancing side of the TMAZ under different peening conditions.
Due to the heat generated during the welding process, the dislocation density in the nugget is lowered from the parent material. Since the strengthening precipitates appear to have solutionized during the welding process [28,29], this indicated that the temperature during joining was above the solution temperature of the hardening precipitates. It is also clear that the tensile properties at the weld interface were lower than their correspondent properties in the HAZ. The properties at the TMAZ on both sides of the weld (advancing, and retreating) were similar. Since the yield strength of the transversely loaded FSW specimens are less than the yield strength of the base metal, the base metal experienced predominantly elastic strain throughout the test [5]. The mechanical properties for all the welded specimens where lower than the base material, as is the case for other precipitation hardened aluminum alloys [30]. Fig. 12 shows the tensile properties at the weld nugget region for different peening conditions. The shot peened specimens demonstrated a slight increase (3%) in the yield stress (0.2% offset). The graph indicates that tensile properties kept increasing as the laser peening layers increased. Eventually, the tensile properties for the laser peened specimens processed with six layers indicated an increase in yield stress of around 60% when compared to the unpeened FSW samples. This specific peening condition also resulted in an 11% increase to the ultimate tensile strength. The tensile properties for the advancing and retreating sides of the TMAZ are presented in Figs. 13 and 14 under different peening conditions. For this region of the weld, the shot peening resulted in tensile properties slightly lower that the specimens processed with a single layer of laser peening; however, the tensile properties were significantly lower than the ones produced using six layers of laser peening. Similar to Fig. 11, the graph also demonstrates that tensile properties increased as the laser peening layers increased. Fig. 15 illustrates the tensile properties at the HAZ region for different peening conditions. In contrast to the other areas of the weld discussed earlier, this graph indicates no improvement in tensile properties was achieved for this location of the weld. This is because this region of the weld mainly exhibited elastic strain throughout the test, as evident by the stain distribution graphs shown in Fig. 16. Plastic strains in the FSW specimen were mainly concentrated in the softer regions of the weld. Specimens that are cut transverse to the weld direction usually fail at lower strains than for the parent material. This is normally
explained by strain localization in areas softened by the welding process [5]. The strain distribution across the weld for different samples before failure is shown in Fig. 16. The figure illustrates the highest strain localization taking place in the weld nugget. This region was shown earlier in Fig. 5 to be the softest area of the weld; therefore, the strain concentration will be highest in this region. It was also noted from Fig. 16, that the specimens peened with six layers of laser peening failed at a lower strain compared to the rest of the samples. The strain at fracture for that sample corresponded to 8.1%. The yield stresses (0.2% offset) were determined at different regions of the weld from the local stress–strain curves, and are shown in Fig. 17. The highest benefits from the peening process were achieved at the weld nugget region (weakest region of the weld). The yield stress values seem to be consistent with the hardness profile in Fig. 5, where the softened weld nugget zone has a considerably lower value than the base material. The yield stress increased in value as the distance from the weld centerline increased. On the other hand, the improvement in the yield stress due to peening started to diminish as the dis-
Fig. 14. Tensile properties at the retreating side of the TMAZ under different peening conditions.
174
O. Hatamleh / Materials Science and Engineering A 492 (2008) 168–176
Fig. 15. Tensile properties at the HAZ region of the weld under different peening conditions.
Fig. 17. The yield stress for various peening conditions at different areas across the weld.
Fig. 16. Strain distribution across the weld for different samples before failure.
O. Hatamleh / Materials Science and Engineering A 492 (2008) 168–176
175
Fig. 18. The yield stress (0.2% offset) for different peening conditions.
Fig. 20. Line plot for the through thickness residual stresses at the center of the weld.
Fig. 19. The ultimate tensile strength for different peening conditions.
tance from the weld increased. The graph also confirms the big increase in tensile properties due to the laser peening using the six layers. The global mechanical properties for the different peening conditions are shown in Figs. 18 and 19. It is clear from Fig. 18, as in the previous graphs, that significant increases in the yield stress were obtained from six layers of laser peening. In contrast, the shot peened samples did not show a measurable change to the yield stress. Fig. 19 also shows the significant increase in ultimate tensile stress from using six layers of laser peening. 4. Discussion The increase in tensile properties from the laser peening process can mainly be attributed to the high level of compressive residual stresses introduced during the high energy peening, and to the increase in hardness from the peening [33]. Fig. 20 illustrates the residual stresses through the thickness of laser peened sample processed with six layers. The measurements were acquired in the weld nugget region using the contour method. The results from the graph outline the high levels of compressive residual stresses at the surface/subsurface, followed by a tensile region in the center of the specimen to balance the compressive residual stresses on the surface. At that specific section measured in the weld nugget, the sum of the tensile and compressive stresses was equivalent to −33 MPa. That level of compressive residual stress is not significant to explain the high increase in tensile properties from the laser peening process. Therefore, other potential contributors to the increase in tensile properties were examined. Recent work by Ali et al. [34] measured the strain hardening as a function of depth from the surface on shot peened aluminium alloy FSW components. The results indicate an increase in hardness at the
center of the weld, the TMAZ, and the parent material. However the hardness depth did not exceed 0.4 mm from the surface. Another investigation by Montross et al. [35] on the effects of laser peening on hardness in aluminium alloys indicate that strain hardening levels were achieved through a depth of 2 mm from the surface of the peened samples. Other studies also indicated a depth of hardening by laser peening of aluminium alloys around 1–2 mm [36] which is consistent with the observations in this study (Fig. 20). The strain hardening which can be explained by the generation of dislocations under the effect of the plastic deformation from peening, is likely to increase the flow resistance of the material to plastic deformation [34]. Since laser peening was able to produce compressive residual stresses significantly deeper than the ones produced by shot peening, the hardness increase in the deeper layers of the shocked sample will result in material plastic deformation and a consequent increase in dislocation density at a deeper thickness. This will result in the noted increase to the tensile properties in the laser peened samples. The decrease in elongation noted from laser peening the specimens with six layers as shown in Fig. 16 is often observed in strain hardened metals. Similar observation was noted by Clauer [37] on laser peened fusion welded 6061-T6 aluminum alloy. The iso-stress assumption used in this study suggests that the different regions of the weld are arranged in series and that the cross-section at any location in the specimen is homogeneous [19,20]. Therefore, this assumption implies that the stress is uniform over the cross-section and is equal to the average stress. The accuracy of the measured tensile properties is therefore determined by the degree of non-homogeneity and residual stress levels at all cross-sections to which the load is applied. Because of the through the weld thickness property gradients that exists in FSW, Lockwood and Reynolds [20] conducted a series of tensile tests on reduced thickness specimens to investigate the validity of this assumption on FSW. Their work included different regions across the weld thickness, and their results indicated that the thick specimen properties closely match the thin specimen properties and seem to justify the iso-stress approximation for friction stir welds. As a result, the iso-stress assumption should be approximately valid in the as-welded specimen. The extension of the iso-stress assumption to calculate local stress–strain curves in surface treated regions is a novel approach and will help to understand and improve the local behavior at var-
176
O. Hatamleh / Materials Science and Engineering A 492 (2008) 168–176
ious regions across the weld resulting in a sound welding process. Nevertheless, since peening the samples only changes the properties near the surface of the processed samples, they are now composed of a composite in parallel, and the results may imply an average stress–strain curve over the region of interest. Because the average will be biased by the depth of the peened layer relative to the thickness, assuming an iso-stress condition in the case of the peened samples may reduce the accuracy of the results. However, the current data obtained in this study may be considered more qualitative than quantitative comparisons, but can still shed some light into the improvement in different regions of the weld due to the peening process. 5. Summary The peening effects on the global and local mechanical properties through the different regions of the FSW were characterized and assessed for 2195 aluminum alloy joints. The tensile coupons were treated with different peening intensities and were machined such that the loading was applied in a direction perpendicular to the weld direction. Since conventional transverse tensile testing only provides the overall strain experienced by the sample, the local strains and equivalent tensile properties were evaluated at different regions of the weld using a digital image correlation system. The peening effects on the local mechanical properties through the different regions of the weld were characterized using a digital image correlation technique assuming an iso-stress condition. This assumption implies that the stress is uniform over the cross-section and is equal to the average stress. The surface and through-thickness residual stresses were also assessed using the X-ray diffraction and contour methods. The laser peened samples displayed an approximate increase of 60% to the yield strength of the material. On the other hand, shot peening exhibited only a slight improvement to the tensile properties when compared to the unpeened FSW specimens. The increase in mechanical properties from the laser peening was mainly attributed to the strain hardening which can be explained by the generation of dislocations under the effect of the plastic deformation from the high energy laser peening. The resulting increase in dislocations tends to increase the flow resistance of the material to plastic deformation. The result that laser peening is superior to shot peening because of the depth of penetration is novel since this superiority has not been presented before. References [1] A. Etter, T. Baudin, N. Fredj, R. Penelle, Mater. Sci. Eng. A 445–446 (2007) 94–99. [2] P. Pao, S. Gill, C. Feng, K. Sankaran, Scripta Mater. 45 (2001) 605–612.
[3] C. Rhodes, M. Mahoney, W. Bingel, R. Spurling, W. Bampton, Scripta Mater. 36 (1) (1997) 69–75. [4] W.M. Thomas, E.D. Nicholas, J.C. Needham, M.G. Church, P. Templesmith, C.J. Dawes, International Patent Application No. PCT/GB92/02203 and GB Patent Application No. 9125978.8, 1991. [5] M. Peel, et al., Acta Mater. 51 (2003) 4791–4801. [6] P. Cavaliere, Mater. Character 57 (2) (2006) 100–104. [7] P. Staron, M. Kocak, S. Williams, A. Wescott, Physica B: Condens. Matter 350 (1–3 (Suppl. 1)) (2004) e491–e493. [8] G. Bussu, P.E. Irving, Int. J. Fatigue 25 (2003) 77–88. [9] R.S. Mishra, Z.Y. Ma, Mater. Sci. Eng. R 50 (2005) 1–78. [10] C. Dale Donne, E. Lima, J. Wegener, A. Pyzalla, T. Buslaps, Third International Symposium on Friction Stir Welding, Kobe, Japan, 2001. [11] M. Milan, J. Tarpani, W. Bose Filho, 18th International Congress of Mechanical Engineering, Ouro Preto, MG, November 6–11, 2005. [12] J.M. Yang, Y.C. Her, N. Han, A. Clauer, Mater. Sci. Eng. A298 (2001) 296–299. [13] C. Rubio-Gonzalez, J.L. Ocana, G. Gomez-Rosas, C. Molpeceres, M. Paredes, A. Banderas, J. Porro, M. Morales, Mater. Sci. Eng. A 386 (2004) 291–295. [14] O. Hatamleh, J. Mater. Eng. Performance (2007). [15] O. Hatamleh, J. Lyons, R. Forman, Int. J. Fatigue 29 (March (3)) (2007) 421–434. [16] C. Montross, T. Wei, L. Ye, G. Clark, Y. Mai, Int. J. Fatigue 24 (2002) 1021– 1036. [17] P. Peyre, R. Fabbro, P. Merrien, H.P. Lieurade, Mater. Sci. Eng. A210 (1996) 102–113. [18] K. Gregory, H.J. Rack, D. Eylon (Eds.), Surface Performance of Titanium, TMS, Warrendale, PA, 1996, pp. 217–230. [19] W.D. Lockwood, Borislav Tomas, A.P. Reynolds, Mater. Sci. Eng., A 323 (January (1–2)) (2002) 349–354. [20] W.D. Lockwood, A.P. Reynolds, Mater. Sci. Eng. A 339 (January (1–2)) (2003) 35–42. [21] Ruud, Clay Olaf, Carpenter, F. Davis, Operators Manual for the D-1000-A Stress Analyzer, Denver X-Ray Instruments, Incorporated, Englewood, 1985. [22] M.E. Hilley (Ed.), J.A. Larson (assoc. Ed.), C.F. Jatczak (assoc. Ed.), R.E. Ricklefs (assist. Ed.), Residual Stress Measurement by X-ray Diffraction SAE J784a, Society of Automotive Engineers, Inc., New York, 1971. [23] M.B. Prime, J. Eng. Mater. Technol. 123 (2001) 162–168. [24] J. Tyson, T. Schmidt, K. Galanulis, Proceedings of SEM 2002 Annual Conference, Milwaukee, WI, June, 2002. [25] T. Schmidt, J. Tyson, K. Galanulis, Exp. Techniques 27 (3) (May/June 2003) 41–44. [26] K. Kumar, S. Brown, J. Pickens, Acta Mater. 44 (5) (1996) 1899–1915. [27] P. Niskanen, T. Sanders, J. Kinker, Corros. Sci. 22 (4) (1982) 283–304. [28] Z. Li, J. Arbegast, P. Hartley, E. Meletis, Proc. 5’th Int. Conf on Trends in Weldin Research, ASM International, Materials Park, OH, 1999, pp. 568–573. [29] G. Oertelt, S. Babu, S. David, E. Kenik, Weld J. 80 (2001) 71–79. [30] H. Kroninger, A. Reynolds, Fatigue Fracture Eng. Mater. Struct. 25 (2002) 283–290. [31] Y. Sato, H. Kokawa, Key Eng. Mater. 345–346 (2007) 1493–1496. [32] H. Liu, M. Maeda, H. Fujii, K. Nogi, J. Mater. Sci. Lett. 22 (January (1)) (2003) 41–43. [33] A.M. Marc, E. Lawrence (Eds.), Shock Waves And High-Strain-Rate Phenomena In Metals, Murr. Plenum Publishing Corporation, 233 Spring Street, New York, N.Y. 10013, 1981. [34] X. Ali, C.A. An, M.W. Rodopoulos, P. Brown, A. O’Hara, S. Levers, Gardiner, Int. J. Fatigue 29 (August (8)) (2007) 1531–1545. [35] C.S. Montross, M. Brandt, M.V. Swain, Surf. Eng. 17 (6) (2001) 477–482. [36] A.H. Clauer, B.P. Fairand, Proceedings of the ASM Conference on Applications of Lasers in Material Processing, Washington D.C., 1979. [37] A.H. Clauer, B.P. Fairand, B.A. Wilcox, Metall. Trans. A 8 (1977) 1871–1876.
Contents lists available at ScienceDirect
Materials Science and Engineering A journal homepage: www.elsevier.com/locate/msea
Effects of peening on mechanical properties in friction stir welded 2195 aluminum alloy joints Omar Hatamleh ∗ Structures Branch, NASA – Johnson Space Center, Houston, TX 77058, United States
a r t i c l e
i n f o
Article history: Received 19 December 2007 Received in revised form 5 March 2008 Accepted 6 March 2008 Keywords: Friction stir welding Laser peening Shot peening Mechanical properties 2195
a b s t r a c t The effects of surface treatment techniques like laser and shot peening on the mechanical properties were investigated for friction stir welded 2195 aluminum alloy joints. The loading in the tensile specimens was applied in a direction perpendicular to the weld direction. The peening effects on the local mechanical properties through the different regions of the weld were characterized using a digital image correlation technique assuming an iso-stress condition. This assumption implies that the stress is uniform over the cross-section and is equal to the average stress. The surface strain and average stress were used giving an average stress–strain curve over the region of interest. The extension of the iso-stress assumption to calculate local stress–strain curves in surface treated regions is a novel approach and will help to understand and improve the local behavior at various regions across the weld resulting in a sound welding process. The surface and through-thickness residual stresses were also assessed using the X-ray diffraction and the contour methods. The laser peened samples displayed approximately 60% increase in the yield strength of the material. In contrast, shot peening exhibited only modest improvement to the tensile properties when compared to the unpeened FSW specimens. The result that laser peening is superior to shot peening because of the depth of penetration is original since this superiority has not been presented before regarding mechanical properties performance. Published by Elsevier B.V.
1. Introduction High strength aluminum alloys are used extensively in the aerospace industries due to their strength and light weight. However, these aerospace aluminum alloys are traditionally considered to be unweldable using conventional fusion welds [1] because of the dentretic structure formed with the fusion welding [2]. Dentretic structure typical of a fusion weld joint can seriously degrade the mechanical properties of the welded joint [3]. Since its invention by the Welding Institute in 1991 [4], friction stir welding (FSW) has emerged as a promising solid state process with encouraging results. This welding technique has potential significant application in different fields including the automotive and aerospace industries [5], and has resulted in welded joints being used in critical load bearing structures and structurally demanding applications [6]. The FSW technique employs a non-consumable cylindrical pin that rotates at high speeds and is then plunged into butting edges of the work pieces to be joined [2]. This process transforms the metal into a plastic state at a temperature below the melting temperature of the material [7], and then mechanically stirs the
∗ Tel.: +1 281 483 0286; fax: +1 281 244 5918. E-mail address: [email protected]. 0921-5093/$ – see front matter. Published by Elsevier B.V. doi:10.1016/j.msea.2008.03.017
material together under pressure to form a welded joint. Since FSW is considered a solid state welding process, significant differences compared to conventional welding may be expected in terms of heat affected zone (HAZ) size and microstructure, and residual stress fields around the weld [8]. The FSW consists of a nugget, the thermo-mechanical affected zone (TMAZ), and a heat affected zone. FSW takes place at a low temperature level compared to fusion welding; therefore, residual stresses may be considerably less than those in fusion welds. Nevertheless, the rigid clamping configuration required to clamp the parts during the FSW process along with the heating cycle the material experiences during welding, can result in higher residual stresses in the weld [9,10]. These residual stresses, along with the reduction in properties from the welding process are likely to affect the mechanical properties and therefore influence the in-service performance of structural components [11]. The weld strength in some cases can be improved by post-weld heat treatment. However, this is not always an option in welded structural components. Consequently, laser shock peening was investigated as a mean for improving the tensile properties in FSW. Laser shock peening is an efficient surface treatment technique that has been proven to be capable of improving the fatigue and mechanical properties of a number of metallic materials [12–15]. The laser peening process (Fig. 1) provides high energy laser pulses
O. Hatamleh / Materials Science and Engineering A 492 (2008) 168–176
Fig. 1. Laser peening process.
that are fired at the surface of a metal coated with an ablative layer, and covered with a thin layer of transparent material (usually water). As the surface layer ablates, the water confines the evaporating material and the vapor is ionized to form plasma. The subsequent laser energy is absorbed by the plasma and generates a high intensity shock wave that impinges on the metal. When the peak pressure of the shock wave is greater than the dynamic yield strength of the material, it produces extensive plastic deformation in the metal. The laser peening process generates a compressive residual stress at the surface that can be significantly deeper than for conventional shot peening [15–17]. The actual depths of the laser-peening induced stresses will vary depending on the material properties of the peened parts and the processing conditions chosen [18]. Several investigations [5,19–20] have assessed the local tensile properties on FSW, however none have assessed the effects of laser peening on the tensile properties on FSW. Since the overall tensile stress of the weld is connected to the local mechanical properties in the weld region; understanding and improving the local behavior at various regions across the weld will result in a sound welding process. In this study, shot peening and laser peening were used to introduce compressive residual stresses into FSW AA 2195. The influence of the peening on the mechanical properties of the FSW specimens were characterized and assessed. Since conventional transverse tensile testing only provides the overall strain experienced by the sample, the local strains and equivalent tensile properties were evaluated at different regions of the weld using a digital image correlation (DIC) system. 2. Experimental setup Aluminum alloy (AA) 2195-T8 plates with a 1.25 cm thickness were used in this research. The aluminium plates were welded together by a single pass using a rotational speed of 300 rpm in the counterclockwise direction and a translation speed of 15 cm/min. The welds were oriented parallel to the rolling direction. The mechanical properties for the base material are as shown in Table 1. Both the root and the crown sides of the weld were tested to evaluate the quality of the weld. The samples were inspected visually afterward with no crack indications revealed. The tensile specimens were either shot-peened or laserpeened. Unpeened FSW samples were also tested and used Table 1 Tensile properties for the as received AA 2195-T8 Material
0.2% Yield stress (MPa)
Ultimate strength (MPa)
2195-T8
503
537
169
as a baseline. To optimize the shot peening process, Peenstress, which is a software developed at the Metal Improvement Company, was used. Based on this evaluation, the samples were shot-peened with 0.059 mm glass beads, with an Almen intensity of 0.008–0.012 A and a 200% coverage rate. The impinging shots were fired at the surface of the specimens at an angle in order to avoid collision with the rebounding beads. For the laser peening process, different peening layers were used in this study in an effort to identify the optimum number of peening layers capable of producing superior tensile properties. Prior to the laser peening process, the surface of the specimens intended for peening was covered with a 0.22 mm thick aluminium tape, which was replaced between layers of peening. A 1 mm thick laminar layer of flowing water was used as a tamping layer. The laser peening was applied using a square laser spot with a laser power density of 5 GW/cm2 and 18 ns in duration. The spots within a layer were overlapped 3%. All four sides of the gauge section in the specimens were peened using the same conditions. The surface residual stresses were measured using the X-ray diffraction (XRD) technique. In XRD, the strain in the crystal lattice is measured assuming that the crystal lattice is linearly distorted. The atomic spacing (d) between crystallographic planes that are equal will vary consistently with their psi ( ) angle, where the angle is defined as the “angle between the surface normal and the normal to the crystallographic planes from which the X-ray peak is diffracted” [21]. Therefore, to determine the magnitude of residual stresses, the lattice strains are assessed in various directions and a plot of sin2 vs. εø is derived (where ø is the angle between a reference direction and the direction of stress measurement in the plane). εø is the strain in the ø and directions defined by [22]: ε =
1+v ( sin2 E
)−
v (1 + 2 ) E
(1)
where : Poisson’s ratio; ø : surface stress at an ø angle with a principal stress direction; E: modulus of elasticity; 1 , 2 : principal stresses. Then from the sin2 vs. εø plot, residual stresses are established through the following relation: =
mE 1+v
(2)
where m = slope of the sin2 vs. εø plot. Residual stress were acquired using a Philips X’Pert PW3040 MRD X-ray diffractometer, equipped with a pole figure goniometer, operating at 40 kV and 45 mA, and employing Ni filtered Cu K-alpha radiation. The measurements were taken using 2 scans from 77◦ to 79◦ , with 0.01◦ per step and 1 s per step, (3 1 1) peak positions at 10 different tilt angles. The through-thickness residual stresses for the FSW sample were measured using the contour method [23] on the plane outlined in Fig. 2. To perform the technique, the specimen was cut along the measurement plane with an EDM wire. In order to minimize movement during the cutting process, the specimen was fixed to a rigid backing plate. The deformed surface shape resulting from the relaxed residual stresses was measured on both cutting surfaces using a coordinate measuring machine (CMM). The displacements from both cutting surfaces were then averaged and noise in the measurements was filtered from average displacements by fitting to a smooth analytical surface. Finally, the original residual stresses were calculated from the measured contour using a finite element model (FEM). Tensile testing was performed at room temperature on a 200 kN servo-hydraulic universal testing machine using a constant crosshead speed of 0.1 mm/min. The transverse tensile specimens
170
O. Hatamleh / Materials Science and Engineering A 492 (2008) 168–176
Fig. 3. Metallographic cross-section of 2195 friction stir weld.
Fig. 2. Tensile sample used for testing.
tensile test. The local strain data was mapped to the corresponding global stress levels by assuming that the transversely loaded FSW specimens were considered a composite material loaded in an iso-stress configuration [19,20]. Using this assumption, local constitutive stress–strain relationships were obtained. 3. Results
consisted of conventional dog bone coupons and were 20 cm long with a gage length of 8.5 cm and a gage width of 1.25 cm (Fig. 2). The coupons were oriented such that the weld was in the center of the specimen and the load was applied perpendicular to the weld direction. Mechanical properties obtained in the transverse tensile test of the FSW weld generally represent the weakest region of the weld. In that configuration, the elongation constitutes an average strain over the whole gage length, which includes the different weld regions. This, in return, does not provide an insight into the correlation between the intrinsic tensile properties and localized microstructure [9]. The local tensile properties at the different regions of the weld were measured using a digital image correlation system. DIC is a powerful system for inspecting complex materials and geometries for their deformation and strain during loading. This tool is capable of performing full-field, non-contact strain measurements. The strain is measured using a three-dimensional video correlation methods and high-resolution digital cameras. A random or regular pattern with good contrast is applied to the surface of the test object and is deformed along with the object. As the specimen is deformed under load, the deformation is recorded by the cameras and evaluated using digital image processing. The initial image processing defines a set of unique correlation areas known as macro-image facets, typically 5–20 pixels across. These facets are then tracked in each successive image with sub-pixel accuracy [24,25]. Therefore, the intrinsic tensile properties for various locations across the weld zone were also characterized by the
3.1. Weld microstructure and hardness The specimen used for metallographic investigation was cut and sectioned in a direction normal to the welding direction and then subjected to several successive steps of grinding and polishing before etching. The micrographs showing different regions of the weld and the corresponding grain sizes for 2195 are shown in Figs. 3 and 4. The FSW sample showed no evidence of porosity or other kinds of defects. The weld nugget seems wider on the crown region of the weld because of the upper surface contact with the tool shoulder. The cross-section also revealed the classical formation of the elliptical “onion rings” structure in the center of the weld. The structure at the nugget is fine and equiaxed with grain sizes significantly smaller than the parent material grain due to the higher temperature and extensive plastic deformation. It was also noted from Fig. 4 that the grains on the top were relatively larger compared to the bottom grains. That was attributed to the heat sink effect of the backing plate. The grain structure at the TMAZ region for the 2195 was elongated with some considerable distortions due to the mechanical action from the welding tool. The HAZ is unaffected by the mechanical effects from the welding tool and the grain structure in that region resembles the parent material grain structure. Kumar et al. and Niskanen et al. [26,27] investigated the precipitates in 2195 and indicated that two main precipitates of 2195 are T1
Fig. 4. Grain sizes at the top and bottom of the weld.
O. Hatamleh / Materials Science and Engineering A 492 (2008) 168–176
171
Fig. 5. The microhardness profile across the weld region for FSW 2195.
(Al2CuLi) and (Al2Cu), which appears at dislocations, subgrains, and grain boundaries. Meanwhile, equivalent phase T2 (Al6CuLi3) precipitates along grain boundaries. During the precipitation of , T1, and T2, a Cu and Li depleted precipitate free zone is formed along grain boundaries and subgrain boundaries due to the absorption of Cu and Li atoms to T1, T2, and phase [26,27]. It is found that T1 precipitates within the grains and coarse equivalent T2 exists at the grain boundaries. To assess the hardness across the weld region, hardness measurements shown in Fig. 5 were taken using a microhardness machine. Softening was noted throughout the weld zone, probably due to coarsening and dissolution of strengthening precipitates induced by the thermal cycle of the FSW. For example, Li et al. and Oertelt et al. [28,29] reported that in 2195 the strengthening precipitates like T1 were no longer present in the weld nugget. The T1 is considered to be the primary strengthening precipitate in the 2195 alloy [30]. Away from the weld nugget, hardness levels increased with increasing distance as precipitation hardening became more effective. It was also noticed from the hardness profiles that the softened region at the top of the weld was relatively wider than the one on the bottom. That is attributed to the specimen thermal characteristics during welding. The top region of the weld generally exhibits higher temperatures compared to the bottom region due to the heating from the tool shoulder, while the bottom surface is in contact with the backing plate, which acts as a heat sink. The lower heat input at the bottom of the plate can significantly reduce the extent of metallurgical transformations, such as re-precipitation and coarsening of precipitates, that take place during welding.
by the shot peening. The surface residual stresses at different distances in the weld nugget region are outlined in Fig. 7. The stresses are shown for the various peening conditions used in this study. In general, all the peening techniques resulted in high compressive stresses, however, shot peening produced the highest surface compressive residual stresses. It was also noticed that the residual stresses in the longitudinal and transverse direction did not vary by an appreciable amount. The through thickness residual stresses obtained by means of the contour method are shown in Figs. 8–10. The magnitude of the measured subsurface residual stresses in the weld were relatively small as shown in Fig. 8 and is mainly attributed to the small size of the specimen that was cut from the plate after welding. When a small specimen is machined from a large plate, a significant amount of the residual stresses they contain are relaxed and reduced in magnitude. Although the shot peened samples exhibited high magnitudes of compressive residual stresses on the surface as illustrated in Figs. 6 and 7, no appreciable subsurface compressive residual stresses were noticed in Fig. 9. The laser peened sample processed with a single layer is shown in Fig. 10. Laser peening
3.2. Residual stress The surface residual stresses at the center of the weld obtained through X-ray diffraction are shown in Fig. 6. The residual stresses for the unpeened specimens were in compression for this region of the weld and the highest compressive residual stresses on the surface were obtained by shot peening, due to the cold work action on the surface. It was also noted that laser peening the samples with one layer did not significantly change the surface stresses at this location compared to the unpeened ones, while the six layers of peening produced residual stresses around 60% of those obtained
Fig. 6. Residual stresses at the center of the weld.
172
O. Hatamleh / Materials Science and Engineering A 492 (2008) 168–176
Fig. 7. Residual stresses for different peening conditions across the weld.
Fig. 8. Two-dimensional map of the measured residual stresses in the unpeened FSW sample.
weld nugget exhibited the lowest tensile properties when compared to other locations across the weld. This may be attributed to the fact that the original structure in this region is over-aged and there is not enough solute left in the material. The transverse tensile properties in welded samples were strongly related to hardness profile in the welds, as shown in Fig. 5. This trend was also reported by Sato et al. and Liu et al. [31,32] for FSW that had no defects. This region of the weld will be relatively ineffective in inhibiting dislocation motion and the resultant strain localization in the softened area of the weld will result in lower mechanical properties. In addition to these microstructural changes, the residual stresses in the weld region may have had an effect on the tensile properties in the samples.
Fig. 9. Two-dimensional map of the measured residual stresses in the shot peened FSW sample.
introduced significant levels of compressive residual stresses near the surface. These stresses had a drastic effect on the overall residual stress state and produced high levels of compensating tensile residual stress at the center of the specimen. 3.3. Mechanical properties The tensile properties at different locations across the weld region are illustrated in Fig. 11. From the graph, it is clear that the
Fig. 10. Two-dimensional map of the measured residual stresses in the laser peened (single layer) FSW sample.
Fig. 11. Tensile properties at different regions of the weld for an unpeened FSW 2195 AA.
O. Hatamleh / Materials Science and Engineering A 492 (2008) 168–176
173
Fig. 12. Tensile properties at the weld nugget under different peening conditions.
Fig. 13. Tensile properties at the advancing side of the TMAZ under different peening conditions.
Due to the heat generated during the welding process, the dislocation density in the nugget is lowered from the parent material. Since the strengthening precipitates appear to have solutionized during the welding process [28,29], this indicated that the temperature during joining was above the solution temperature of the hardening precipitates. It is also clear that the tensile properties at the weld interface were lower than their correspondent properties in the HAZ. The properties at the TMAZ on both sides of the weld (advancing, and retreating) were similar. Since the yield strength of the transversely loaded FSW specimens are less than the yield strength of the base metal, the base metal experienced predominantly elastic strain throughout the test [5]. The mechanical properties for all the welded specimens where lower than the base material, as is the case for other precipitation hardened aluminum alloys [30]. Fig. 12 shows the tensile properties at the weld nugget region for different peening conditions. The shot peened specimens demonstrated a slight increase (3%) in the yield stress (0.2% offset). The graph indicates that tensile properties kept increasing as the laser peening layers increased. Eventually, the tensile properties for the laser peened specimens processed with six layers indicated an increase in yield stress of around 60% when compared to the unpeened FSW samples. This specific peening condition also resulted in an 11% increase to the ultimate tensile strength. The tensile properties for the advancing and retreating sides of the TMAZ are presented in Figs. 13 and 14 under different peening conditions. For this region of the weld, the shot peening resulted in tensile properties slightly lower that the specimens processed with a single layer of laser peening; however, the tensile properties were significantly lower than the ones produced using six layers of laser peening. Similar to Fig. 11, the graph also demonstrates that tensile properties increased as the laser peening layers increased. Fig. 15 illustrates the tensile properties at the HAZ region for different peening conditions. In contrast to the other areas of the weld discussed earlier, this graph indicates no improvement in tensile properties was achieved for this location of the weld. This is because this region of the weld mainly exhibited elastic strain throughout the test, as evident by the stain distribution graphs shown in Fig. 16. Plastic strains in the FSW specimen were mainly concentrated in the softer regions of the weld. Specimens that are cut transverse to the weld direction usually fail at lower strains than for the parent material. This is normally
explained by strain localization in areas softened by the welding process [5]. The strain distribution across the weld for different samples before failure is shown in Fig. 16. The figure illustrates the highest strain localization taking place in the weld nugget. This region was shown earlier in Fig. 5 to be the softest area of the weld; therefore, the strain concentration will be highest in this region. It was also noted from Fig. 16, that the specimens peened with six layers of laser peening failed at a lower strain compared to the rest of the samples. The strain at fracture for that sample corresponded to 8.1%. The yield stresses (0.2% offset) were determined at different regions of the weld from the local stress–strain curves, and are shown in Fig. 17. The highest benefits from the peening process were achieved at the weld nugget region (weakest region of the weld). The yield stress values seem to be consistent with the hardness profile in Fig. 5, where the softened weld nugget zone has a considerably lower value than the base material. The yield stress increased in value as the distance from the weld centerline increased. On the other hand, the improvement in the yield stress due to peening started to diminish as the dis-
Fig. 14. Tensile properties at the retreating side of the TMAZ under different peening conditions.
174
O. Hatamleh / Materials Science and Engineering A 492 (2008) 168–176
Fig. 15. Tensile properties at the HAZ region of the weld under different peening conditions.
Fig. 17. The yield stress for various peening conditions at different areas across the weld.
Fig. 16. Strain distribution across the weld for different samples before failure.
O. Hatamleh / Materials Science and Engineering A 492 (2008) 168–176
175
Fig. 18. The yield stress (0.2% offset) for different peening conditions.
Fig. 20. Line plot for the through thickness residual stresses at the center of the weld.
Fig. 19. The ultimate tensile strength for different peening conditions.
tance from the weld increased. The graph also confirms the big increase in tensile properties due to the laser peening using the six layers. The global mechanical properties for the different peening conditions are shown in Figs. 18 and 19. It is clear from Fig. 18, as in the previous graphs, that significant increases in the yield stress were obtained from six layers of laser peening. In contrast, the shot peened samples did not show a measurable change to the yield stress. Fig. 19 also shows the significant increase in ultimate tensile stress from using six layers of laser peening. 4. Discussion The increase in tensile properties from the laser peening process can mainly be attributed to the high level of compressive residual stresses introduced during the high energy peening, and to the increase in hardness from the peening [33]. Fig. 20 illustrates the residual stresses through the thickness of laser peened sample processed with six layers. The measurements were acquired in the weld nugget region using the contour method. The results from the graph outline the high levels of compressive residual stresses at the surface/subsurface, followed by a tensile region in the center of the specimen to balance the compressive residual stresses on the surface. At that specific section measured in the weld nugget, the sum of the tensile and compressive stresses was equivalent to −33 MPa. That level of compressive residual stress is not significant to explain the high increase in tensile properties from the laser peening process. Therefore, other potential contributors to the increase in tensile properties were examined. Recent work by Ali et al. [34] measured the strain hardening as a function of depth from the surface on shot peened aluminium alloy FSW components. The results indicate an increase in hardness at the
center of the weld, the TMAZ, and the parent material. However the hardness depth did not exceed 0.4 mm from the surface. Another investigation by Montross et al. [35] on the effects of laser peening on hardness in aluminium alloys indicate that strain hardening levels were achieved through a depth of 2 mm from the surface of the peened samples. Other studies also indicated a depth of hardening by laser peening of aluminium alloys around 1–2 mm [36] which is consistent with the observations in this study (Fig. 20). The strain hardening which can be explained by the generation of dislocations under the effect of the plastic deformation from peening, is likely to increase the flow resistance of the material to plastic deformation [34]. Since laser peening was able to produce compressive residual stresses significantly deeper than the ones produced by shot peening, the hardness increase in the deeper layers of the shocked sample will result in material plastic deformation and a consequent increase in dislocation density at a deeper thickness. This will result in the noted increase to the tensile properties in the laser peened samples. The decrease in elongation noted from laser peening the specimens with six layers as shown in Fig. 16 is often observed in strain hardened metals. Similar observation was noted by Clauer [37] on laser peened fusion welded 6061-T6 aluminum alloy. The iso-stress assumption used in this study suggests that the different regions of the weld are arranged in series and that the cross-section at any location in the specimen is homogeneous [19,20]. Therefore, this assumption implies that the stress is uniform over the cross-section and is equal to the average stress. The accuracy of the measured tensile properties is therefore determined by the degree of non-homogeneity and residual stress levels at all cross-sections to which the load is applied. Because of the through the weld thickness property gradients that exists in FSW, Lockwood and Reynolds [20] conducted a series of tensile tests on reduced thickness specimens to investigate the validity of this assumption on FSW. Their work included different regions across the weld thickness, and their results indicated that the thick specimen properties closely match the thin specimen properties and seem to justify the iso-stress approximation for friction stir welds. As a result, the iso-stress assumption should be approximately valid in the as-welded specimen. The extension of the iso-stress assumption to calculate local stress–strain curves in surface treated regions is a novel approach and will help to understand and improve the local behavior at var-
176
O. Hatamleh / Materials Science and Engineering A 492 (2008) 168–176
ious regions across the weld resulting in a sound welding process. Nevertheless, since peening the samples only changes the properties near the surface of the processed samples, they are now composed of a composite in parallel, and the results may imply an average stress–strain curve over the region of interest. Because the average will be biased by the depth of the peened layer relative to the thickness, assuming an iso-stress condition in the case of the peened samples may reduce the accuracy of the results. However, the current data obtained in this study may be considered more qualitative than quantitative comparisons, but can still shed some light into the improvement in different regions of the weld due to the peening process. 5. Summary The peening effects on the global and local mechanical properties through the different regions of the FSW were characterized and assessed for 2195 aluminum alloy joints. The tensile coupons were treated with different peening intensities and were machined such that the loading was applied in a direction perpendicular to the weld direction. Since conventional transverse tensile testing only provides the overall strain experienced by the sample, the local strains and equivalent tensile properties were evaluated at different regions of the weld using a digital image correlation system. The peening effects on the local mechanical properties through the different regions of the weld were characterized using a digital image correlation technique assuming an iso-stress condition. This assumption implies that the stress is uniform over the cross-section and is equal to the average stress. The surface and through-thickness residual stresses were also assessed using the X-ray diffraction and contour methods. The laser peened samples displayed an approximate increase of 60% to the yield strength of the material. On the other hand, shot peening exhibited only a slight improvement to the tensile properties when compared to the unpeened FSW specimens. The increase in mechanical properties from the laser peening was mainly attributed to the strain hardening which can be explained by the generation of dislocations under the effect of the plastic deformation from the high energy laser peening. The resulting increase in dislocations tends to increase the flow resistance of the material to plastic deformation. The result that laser peening is superior to shot peening because of the depth of penetration is novel since this superiority has not been presented before. References [1] A. Etter, T. Baudin, N. Fredj, R. Penelle, Mater. Sci. Eng. A 445–446 (2007) 94–99. [2] P. Pao, S. Gill, C. Feng, K. Sankaran, Scripta Mater. 45 (2001) 605–612.
[3] C. Rhodes, M. Mahoney, W. Bingel, R. Spurling, W. Bampton, Scripta Mater. 36 (1) (1997) 69–75. [4] W.M. Thomas, E.D. Nicholas, J.C. Needham, M.G. Church, P. Templesmith, C.J. Dawes, International Patent Application No. PCT/GB92/02203 and GB Patent Application No. 9125978.8, 1991. [5] M. Peel, et al., Acta Mater. 51 (2003) 4791–4801. [6] P. Cavaliere, Mater. Character 57 (2) (2006) 100–104. [7] P. Staron, M. Kocak, S. Williams, A. Wescott, Physica B: Condens. Matter 350 (1–3 (Suppl. 1)) (2004) e491–e493. [8] G. Bussu, P.E. Irving, Int. J. Fatigue 25 (2003) 77–88. [9] R.S. Mishra, Z.Y. Ma, Mater. Sci. Eng. R 50 (2005) 1–78. [10] C. Dale Donne, E. Lima, J. Wegener, A. Pyzalla, T. Buslaps, Third International Symposium on Friction Stir Welding, Kobe, Japan, 2001. [11] M. Milan, J. Tarpani, W. Bose Filho, 18th International Congress of Mechanical Engineering, Ouro Preto, MG, November 6–11, 2005. [12] J.M. Yang, Y.C. Her, N. Han, A. Clauer, Mater. Sci. Eng. A298 (2001) 296–299. [13] C. Rubio-Gonzalez, J.L. Ocana, G. Gomez-Rosas, C. Molpeceres, M. Paredes, A. Banderas, J. Porro, M. Morales, Mater. Sci. Eng. A 386 (2004) 291–295. [14] O. Hatamleh, J. Mater. Eng. Performance (2007). [15] O. Hatamleh, J. Lyons, R. Forman, Int. J. Fatigue 29 (March (3)) (2007) 421–434. [16] C. Montross, T. Wei, L. Ye, G. Clark, Y. Mai, Int. J. Fatigue 24 (2002) 1021– 1036. [17] P. Peyre, R. Fabbro, P. Merrien, H.P. Lieurade, Mater. Sci. Eng. A210 (1996) 102–113. [18] K. Gregory, H.J. Rack, D. Eylon (Eds.), Surface Performance of Titanium, TMS, Warrendale, PA, 1996, pp. 217–230. [19] W.D. Lockwood, Borislav Tomas, A.P. Reynolds, Mater. Sci. Eng., A 323 (January (1–2)) (2002) 349–354. [20] W.D. Lockwood, A.P. Reynolds, Mater. Sci. Eng. A 339 (January (1–2)) (2003) 35–42. [21] Ruud, Clay Olaf, Carpenter, F. Davis, Operators Manual for the D-1000-A Stress Analyzer, Denver X-Ray Instruments, Incorporated, Englewood, 1985. [22] M.E. Hilley (Ed.), J.A. Larson (assoc. Ed.), C.F. Jatczak (assoc. Ed.), R.E. Ricklefs (assist. Ed.), Residual Stress Measurement by X-ray Diffraction SAE J784a, Society of Automotive Engineers, Inc., New York, 1971. [23] M.B. Prime, J. Eng. Mater. Technol. 123 (2001) 162–168. [24] J. Tyson, T. Schmidt, K. Galanulis, Proceedings of SEM 2002 Annual Conference, Milwaukee, WI, June, 2002. [25] T. Schmidt, J. Tyson, K. Galanulis, Exp. Techniques 27 (3) (May/June 2003) 41–44. [26] K. Kumar, S. Brown, J. Pickens, Acta Mater. 44 (5) (1996) 1899–1915. [27] P. Niskanen, T. Sanders, J. Kinker, Corros. Sci. 22 (4) (1982) 283–304. [28] Z. Li, J. Arbegast, P. Hartley, E. Meletis, Proc. 5’th Int. Conf on Trends in Weldin Research, ASM International, Materials Park, OH, 1999, pp. 568–573. [29] G. Oertelt, S. Babu, S. David, E. Kenik, Weld J. 80 (2001) 71–79. [30] H. Kroninger, A. Reynolds, Fatigue Fracture Eng. Mater. Struct. 25 (2002) 283–290. [31] Y. Sato, H. Kokawa, Key Eng. Mater. 345–346 (2007) 1493–1496. [32] H. Liu, M. Maeda, H. Fujii, K. Nogi, J. Mater. Sci. Lett. 22 (January (1)) (2003) 41–43. [33] A.M. Marc, E. Lawrence (Eds.), Shock Waves And High-Strain-Rate Phenomena In Metals, Murr. Plenum Publishing Corporation, 233 Spring Street, New York, N.Y. 10013, 1981. [34] X. Ali, C.A. An, M.W. Rodopoulos, P. Brown, A. O’Hara, S. Levers, Gardiner, Int. J. Fatigue 29 (August (8)) (2007) 1531–1545. [35] C.S. Montross, M. Brandt, M.V. Swain, Surf. Eng. 17 (6) (2001) 477–482. [36] A.H. Clauer, B.P. Fairand, Proceedings of the ASM Conference on Applications of Lasers in Material Processing, Washington D.C., 1979. [37] A.H. Clauer, B.P. Fairand, B.A. Wilcox, Metall. Trans. A 8 (1977) 1871–1876.