Journal of Materials Science & Technology  2019 , 35 (8): 1555-1562 https://doi.org/10.1016/j.jmst.2019.03.036

Orginal Article

Improving fatigue performance of Ti-6Al-4V alloy via ultrasonic surface rolling process

Chengsong Liu, Daoxin Liu*, Xiaohua Zhang, Dan Liu, Amin Ma, Ni Ao, Xingchen Xu

Corrosion and Protection Research Laboratory, Northwestern Polytechnical University, Xi'an 710072, China

Corresponding authors:   *Corresponding author.E-mail address: liudaox@nwpu.edu.cn (D. Liu).

Received: 2018-12-19

Revised:  2019-03-5

Accepted:  2019-03-12

Online:  2019-08-05

Copyright:  2019 Editorial board of Journal of Materials Science & Technology Copyright reserved, Editorial board of Journal of Materials Science & Technology

More

Abstract

The effect of a gradient nanostructured (GNS) surface layer obtained by ultrasonic surface rolling process (USRP) on the fatigue behavior of Ti-6Al-4 V alloy has been studied in this paper. Microstructure, surface topography, surface roughness and residual stress measurements were performed to characterize the surface under different conditions. Rotating bending fatigue tests were carried out to evaluate the fatigue behavior of different treatments. The results present a remarkable fatigue performance enhancement for the Ti-6Al-4 V alloy with a GNS surface layer obtained by application of USRP with respect to the untreated condition, notwithstanding its considerable surface roughness due to severe ultrasonic impacts and extrusions. Mechanical surface polishing treatment further enhances the beneficial effects of USRP on the fatigue performance. The significantly improved fatigue performance can mainly be ascribed to the compressive residual stress. Simultaneously, the GNS surface layer and surface work hardening have a synergistic effect that accompanies the effect of compressive residual stress.

Keywords: Ultrasonic surface rolling process ; Fatigue ; Compressive residual stress ; Gradient nanostructured surface layer ; Ti-6Al-4V alloy

0

PDF (3411KB) Metadata Metrics Related articles

Cite this article Export EndNote Ris Bibtex

Chengsong Liu, Daoxin Liu, Xiaohua Zhang, Dan Liu, Amin Ma, Ni Ao, Xingchen Xu. Improving fatigue performance of Ti-6Al-4V alloy via ultrasonic surface rolling process[J]. Journal of Materials Science & Technology, 2019, 35(8): 1555-1562 https://doi.org/10.1016/j.jmst.2019.03.036

1. Introduction

Titanium alloys are extensively used in the aerospace field, such as for compressor fan blades/disks, in view of their low density, high specific strength, good corrosion resistance and excellent thermal stability [1]. However, these components are commonly subjected to fatigue and fail, due to the fact that flaws or cracks always initiate on the surface under the rotational and high-frequency vibrational conditions [2]. Therefore, it is imperative to improve the surface properties of metallic materials by the surface deformation strengthening technology. Conventional shot peening (SP) can be expected to effectively improve the surface properties and fatigue resistance, owing to the high level of compressive residual stress and surface work hardening obtained by SP [3,4].

However, there are other surface deformation strengthening technologies based on surface severe plastic deformation (S2PD), in addition to conventional SP, that have been previously developed to enhance the surface and mechanical properties of different materials, notably severe shot peening (SSP) [5,6], surface mechanical attrition treatment (SMAT) [7], surface mechanical rolling treatment (SMRT) [8], surface mechanical grinding treatment (SMGT) [9,10], laser shock processing (LSP) [11], ultrasonic shot peening (USP) [12,13] and ultrasonic peening treatment (UPT) [14,15]. With these technologies, a gradient nanostructured (GNS) surface layer can be formed on the surface of metallic materials via S2PD. The GNS surface layer can usually be divided into several layers, such as the nanocrystallization layer, refined structure layer, deformation coarse-grained layer and strain-free coarse-grained substrate, from the top surface into the interior [8,16,17]. Previous studies have demonstrated that the formation of GNS surface layers with compressive residual stress and work hardening remarkably enhanced the wear [9], corrosion [12] and fatigue [[5], [6], [7], [8],11,13,14] resistance of mechanical components. For example, the low cycle fatigue (LCF) and high cycle fatigue (HCF) strength of AISI 316 stainless steel was enhanced by SMAT [7]. The fatigue limit of GNS Cu specimens was improved by 75% compared with untreated specimens [10]. The fatigue life of surface nanocrystalline K403 nickel-alloy could be improve effectively by 2.44 times under the stress level of 425 MPa [11]. Simultaneously, the LCF life of Ti-6Al-4 V alloy with the GNS surface layer could be enhanced more than 4 times under the lowest strain amplitude of ±0.60% [13].

Recently, a novel surface deformation strengthening technology, known as ultrasonic surface rolling process (USRP), has attracted increasing attention for surface modification of metallic materials, especially forming the GNS surface layer via S2PD [[18], [19], [20], [21]]. During USRP, repeated impacts with a high vibration frequency (≥20 kHz) are applied to the specimen surface through a rolling working tip attached to an ultrasonic apparatus [18,20]. Thus, obvious grain refinement and even surface nanocrystallization, compressive residual stress and work hardening can be observed on the specimen surface, which is ascribed to S2PD produced by the repeated high frequency impacts and extrusions [19,21]. Compared with the conventional SP, USRP equipment is usually assembled with computer numerical control (CNC) lathe or milling machine, and the strike intensity can be accurately controlled. Thus, this treatment method has high repeatability and accuracy for industrial applications. Currently, the USRP has been successfully applied to improve the mechanical and fatigue properties of steels [20,22], aluminum alloy [23] and titanium alloys [18,24,25].

The enhancement of the fatigue performance for the GNS specimens has generally been ascribed to the combined effects of the compressive residual stress, microstructure refinement and surface work hardening [26]. However, the main factor for improving fatigue performance is not clear enough. Huang et al. [8] studied the relations between the microstructure, mechanical properties and the fatigue performance of GNS 316 L stainless steel and found that the GNS surface layer was the major factor in inhibiting the initiation of fatigue cracks, and compressive residual stress was insignificant to the fatigue enhancement. However, little has been done with the dominant factor for enhancing the fatigue performance of materials in the USRP. S2PD methods, for example SSP [5,6] and UPT [14], have both beneficial and detrimental effects. On the one hand, the induced GNS surface layer, compressive residual stress and work hardening layer can provide beneficial effects for improving the fatigue strength. On the other hand, surface defects and high surface roughness introduced during the S2PD have a detrimental effect, leading to the relaxation of surface compressive residual stress and the initiation of fatigue cracks at the surface defects. For instance, the fatigue limit of cast iron specimens with a GNS surface layer created by means of SSP exhibited no significant improvement in comparison to the fatigue limit of conventional SP [6]. S355 steel specimens with increasing fold defects induced by UPT showed a significant decrease of the fatigue life [14]. However, there is a lack of understanding about the surface defects introduced during USRP, and the effects of those defects on the fatigue behavior have not been systematically studied.

Based on the above backgrounds, a GNS surface layer is prepared on the Ti-6Al-4 V alloy specimens by application of the USRP to improve their fatigue performance. In particular, the correlations between the microstructure, residual stress, work hardening, surface defects and the fatigue behavior of the GNS specimens are clarified. The main factor for the USRP enhancing the fatigue performance of Ti-6Al-4 V alloy is also studied.

2. Experimental

2.1. Materials

The original material in this work was an annealed-state Ti-6Al-4 V alloy with the following chemical composition (wt%): 6.67 Al, 4.2 V, 0.1 Fe, 0.03 C, 0.015 N, 0.03 H, 0.14 O and balance Ti. The material was vacuum-annealed at 890 °C for 1 h and subsequently air cooled. The corresponding microstructure composed of equiaxed primary α phase grains (with a grain size of approximately 5-20 μm) and lamellar α + β colonies (transformed β phase); the scanning electron microscope (SEM) image of original material is shown in Fig. 1(a). Fig. 1(b) presents the transmission electron microscope (TEM) image of the original Ti-6Al-4 V alloy before the USRP. As depicted in Fig. 1(b), the grains show a relatively large size with clear boundaries. Only small number of dislocations can be found in the grains, but meanwhile some dislocations accumulate near the grain boundaries. The main mechanical properties are shown in Table 1.

Fig. 1.   SEM (a) and TEM (b) images of original Ti-6Al-4V alloy before USRP.

Table 1   Main mechanical properties of Ti-6Al-4 V alloy used in this work.

MaterialYield strength (MPa)Tensile strength (MPa)Elongation (%)Reduction in area (%)
Ti-6Al-4V101010801441

New window

2.2. Fatigue specimen preparation and fatigue testing

The shape and dimension of the fatigue specimen are given in Fig. 2, and the working segments of the fatigue specimens (with a gauge of 40 mm in the middle part) were longitudinally polished using a lathe machine. The USRP experiments were carried out on a self-built platform based on a CNC lathe. The schematic diagram and working principle of the USRP are shown in Ref. [20] in detail. The working tip is a freely scrollable rolling WC/Co ball with a diameter, hardness, and surface roughness of 14 mm, 80 HRC, and Ra 0.1 μm, respectively. The basic USRP parameters of the present work are as follows: 120 rpm for the lathe rotational speed, 600 N for the static force, 0.1 mm/r for the feeding rate, 10 μm for the ultrasonic vibration amplitude, 20 kHz for the ultrasonic vibration frequency and 24 for the repeated processing number. The USRP treated and untreated base material specimens were labeled as USRP and BM, respectively. Fatigue tests under different conditions, with a test frequency of 50 Hz and a stress ratio of -1, were conducted by using a PQ-6 type rotating bending fatigue tester at room temperature.

Fig. 2.   Shape and dimension of fatigue specimen of Ti-6Al-4 V alloy (unit: mm).

2.3. Measurement apparatuses and methods

The microstructures, corresponding surface morphologies, and fracture features were characterized via a JSM-6390 SEM. To characterize the nature of the microstructures of the BM and USRP specimens, TEM and the corresponding selected area electron diffraction (SAED) were carried out using a 200 kV Tecnai G2 F20 microscope. A 3D optical stereomicroscope (C130 type) was applied to observe the 3D surface morphologies and surface roughness values. The microhardness was measured using a MHV-1000 type semiautomatic digital microhardness testing system at a 25 g test load and 20 s dwell time equipped with a Knoop diamond indenter. Three parallel measurements were performed starting from the top surface towards interior material and the results were averaged at each depth. The state of the residual stress on the specimens was measured using an Xstress 3000 type X-ray diffractometer (radiation CuKα, irradiated area 3.14 mm2, sin2Ψ method (Ψ: tilt angle), and Bragg angle of 142° in the {213} plane of α-Ti). The residual stress distribution with different depth was conducted after successive layer removal using a mixture of HNO3 and HF solution.

3. Results

3.1. Microstructure of USRP surface layer

The cross-sectional microstructure of the USRP specimen is shown in Fig. 3. After USRP, a gradient plastic deformation layer of approximately 335 μm and obvious plastic flow can be observed on the treated surface (Fig. 3(a)). Furthermore, the maximum amount of impact energy is accumulated on the surface, so that the plastic deformation at the specimen surface is expected to be the largest (Fig. 3(b)). The primary equiaxed α phase grains of the deformation layer evolve into compressed and elongated lamellar grains near the surface under repeated ultrasonic impacts (Fig. 3(b)). And the grain size at the short axis on the near surface are in the range of 1-5 μm (Fig. 3(b)). Meanwhile, deflection angle θ of lamellar α phase at the depth of 0 $\widetilde{3}$0 μm is approximately 72° along the direction of plastic flow (Fig. 3(b)). In addition, the grains of the α phase are also elongated in the subsurface (Fig. 3(c)-(e)). Nevertheless, the grain size of the α phase at the short axis gradually increases, and the deflection angle θ gradually reduces with the increasing depth (Fig. 3(c)-(e)), which result from the fact that the impact energy gradually attenuates with increasing depth. Then, the equiaxed grains of the substrate are found at a depth of <335 μm below the treated surface (Fig. 3(f)).

Fig. 3.   Cross-sectional SEM morphologies of USRP specimen.

Fig. 4 shows the TEM micrographs at a depth of 10 μm below the USRP specimen surface. The coarse grains of original material have been refined into roughly nanoscale equiaxed grains, indicating the characteristics of recrystallization during the USRP (Fig. 4(a) and (b)) [27]. The corresponding SAED pattern (the inset in Fig. 4(a)), within a 500 nm diameter selected area, is characterized by approximately complete continuous rings, proving that there are random misorientations between the recrystallized fine grains. In addition, the average grain size is 91.5 nm at 10 μm below the surface based on the statistical analysis. The formation of nanograins can be ascribed to the high energy generated by the continuous impacts and extrusions at high speeds [20]. According to research results of the authors [18] and other scholars [25], the GNS surface layers can be formed in Ti-6Al-4 V and TC11 titanium alloys by means of the USRP. The equiaxed nanograin layers just below the top surface were 0-50 μm thick when using different S2PD methods [11,17,18,25,28]. Therefore, one can conclude that a GNS surface layer has been formed on the surface of Ti-6Al-4 V alloy during the USRP in this work, and the thickness of equiaxed nanograin layer on the top surface is at least 10 μm.

Fig. 4.   TEM micrographs at depth of 10 μm below USRP specimen surface: (a) bright-field image with inset showing corresponding SAED pattern; (b) dark-field image.

3.2. Surface morphology and roughness

The 3D topographies, surface topographies and cross-sectional topographies of the BM and USRP specimens are presented in Fig. 5, and the surface roughness values of the corresponding specimens are compared in Fig. 6. It can be seen from Fig. 5(a)-(c) that obvious mechanical polishing marks can be found on the surface of the BM specimen. And the surface roughness average values of Ra and Rz for the BM specimens are 0.416 μm and 3.207 μm, respectively (Fig. 6). As a comparison, the USRP specimen shows a smoother and more compact surface morphology in most regions (Fig. 5(e)). Nevertheless, surface defects, including surface delamination (Fig. 5(e)) and a so-called fold defect (Fig. 5(f)), are found on the treated surface due to the severe ultrasonic impacts and extrusions. Fold defects, as reported by Liu et al., usually appear on the surface of UPT specimens [29]. During UPT, a part of the surface material is extruded to adjacent parts along the travel direction of the needle, and those adjacent interface overlaps to each other and close together due to the S2PD, resulting in a fold defect [14]. Accordingly, the surface roughness average values of Ra and Rz for the USRP specimens are 72.6% and 39.1% higher than those of the BM specimens, respectively, due to the appearance of surface defects (Fig. 6).

Fig. 5.   3D topographies (a, d), surface topographies (b, e) and cross-sectional topographies (c, f) with treatments of BM (a-c) and USRP (d-f).

Fig. 6.   Surface roughness of BM and USRP specimens.

3.3. Microhardness and residual stress

The variations in microhardness and residual stress distribution along the depth axis for the USRP specimen are shown in Fig. 7. The USRP specimen shows an increment of 44.7% in the surface microhardness, where the surface microhardness shows a value of 521 H K at the top surface and then gradually reduces until reaching the untreated BM value of approximately 360 H K at a depth of about 500 μm. This gradient hardness distribution is beneficial to improving the load carrying capacity [30] and suppressing fatigue crack initiation [31]. The reason for such an increase in the microhardness after the USRP can be attributed to dislocation multiplication and microstructure refinement [19]. It is also observed that a significant level of compressive residual stress field with the depth of approximately 650 μm is introduced by USRP at the specimen surface and subsurface region. The value of surface compressive residual stress is 864.7 MPa. Besides, the maximum value of compressive residual stress is 1155 MPa at the subsurface depth of about 125 μm. In comparison with those of conventional SP and wet peening, the values and depth of the compressive residual stress are rather high in the GNS surface layer obtained by USRP [31,32].

Fig. 7.   Variations in residual stress and microhardness distribution of USRP specimen.

3.4. S/N curves

The stress/life (S/N) curves for the BM and USRP specimens are compared in Fig. 8. It is apparent that USRP significantly enhances the fatigue life not only in the LCF regime but also in the HCF regime. The 107 cycles fatigue limit of the BM specimen is 500 MPa. Compared with the untreated condition, the USRP condition increases the fatigue limit by well over 22% ($\widetilde{1}$10 MPa). Moreover, the beneficial effect of the USRP on fatigue life is more significant in HCF regime with a fatigue life of more than 5 × 104 cycles, which is consistent with previous results showing that conventional SP, LSP and UPT improved the fatigue limit of metallic materials [[33], [34], [35], [36]]. The main reasons for this behavior are that crack initiation consumes a greater portion of the fatigue life [34] and less relaxation of compressive residual stress occurs in the low cycle stress HCF regime [3,37]. Meanwhile, the compressive residual stress introduced by the USRP can effectively retard crack initiation and premature propagation [24,38]. Hence, the beneficial influence on fatigue life for the USRP specimens is more remarkable in the HCF regime.

Fig. 8.   S/N curves for BM and USRP conditions. Run-outs are indicated by arrows.

4. Discussion

4.1. Factor separation methods

Under S2PD conditions, for instance the LSP, UPT and USRP, the enhanced fatigue performance can always be attributed to the comprehensive effects of compressive residual stress, microstructures and surface work hardening [26,34,39,40]. The presence of compressive residual stress shows a beneficial effect on suppressing the fatigue crack initiation and slowing down the initial crack propagation [34,39]. The GNS surface layers, with the increased dislocation density and surface work hardening, are helpful in suppressing the initiation of fatigue cracks [8,13,26]. However, the key factor in improving fatigue performance needs to be revealed. Furthermore, the existence of surface defects for the USRP specimens may be harmful to fatigue performance to some extent, and the effects of those defects on fatigue performance also need to be revealed. Therefore, several factor separation methods were developed for the further study in our research. The separation factor methods of the USRP specimens are shown as follows:

(1) Mechanical surface polishing by SiC grit papers, numbers 1000-5000, was carried out to remove surface delamination and fold defects. The total removal depth was approximately 10 μm judging from the cross-sectional microstructure (Fig. 5(f)). The corresponding specimens are labeled as USRP + P. According to the analysis of Fig. 3, Fig. 4 in Section 3.1, a GNS surface layer is also present on the surface of the USRP + P specimens. Measurements show that the surface roughness Ra of the USRP + P specimens is approximately 0.185-0.260 μm. Fig. 9 shows the 3D topography and surface topography of the USRP + P specimen. It can be observed that the USRP + P specimen presents a smooth and compact surface morphology, and surface defects have been basically removed.

Fig. 9.   3D topography (a) and surface topography (b) of USRP + P specimen.

(2) To effectively eliminate residual stress while not affecting the microstructure as much as possible, a residual stress relaxation experiment was conducted using the high stress rotating bending fatigue cyclic loading method [37,41]. Under the process of residual stress relaxation experiment, fatigue tests were stopped after a designed cycle number of 50 under the stress condition of 900 MPa, then the surface residual stress was measured by using a X-ray diffractometer. In order to exclude the effects of surface defects, the residual stress relaxation experiment was conducted on the USRP + P specimens, and corresponding specimens were labeled as USRP + P+F in this work. To verify this method, the BM specimens could also be evaluated by this method, and corresponding specimens were labeled as BM + F.

4.2. Fatigue strengthening mechanism

The surface microhardness and surface residual stress of the different treatment specimens are shown in Fig. 10(a). It is apparent that the polishing treatment has a negligible effect on the surface microhardness, but the surface compressive residual stress has a slight increase in the USRP + P specimens. The reason for this slight increase is that the compressive residual stress in the subsurface within the depth of 125 μm is higher than that of the top surface in Fig. 7. The surface microhardness of the BM and USRP + P specimens would be slightly decreased after the residual stress relaxation experiment, because the relaxation of compressive residual stress causes a decrease in surface microhardness [7]. After the residual stress relaxation experiment, the surface compressive residual stress of the USRP + P+F specimens was reduced by 51.2% compared with that of the USRP + P specimens.

Fig. 10.   (a) Surface microhardness and surface residual stress of different treatment specimens and (b) fatigue life of corresponding specimens in Fig. 10(a).

The fatigue life of different treatment specimens, obtained from three parallel tests at a maximum stress level of 700 MPa, are shown in Fig. 10(b). It can be observed that the relaxation of residual stress has no significant effect on the fatigue life of the BM specimens, indicating that this method is suitable for the separation of residual stress factor. In terms of the USRP specimens, despite the higher surface roughness, the fatigue life improved considerably by almost 23.5 times with respect to the fatigue life of the BM specimens. The mechanical surface polishing treatment results in a considerable improvement in the fatigue life from 764,410 cycles of USRP specimens to 3,576,893 cycles of USRP + P specimens, an increment of 3.7 times. It is also noted that the fatigue life of the USRP + P specimens, with few surface defects and lower surface roughness, show a more remarkable increase of 113.8 times in comparison to the fatigue life of the BM specimens. A similar phenomenon was also found in the SSP and re-peened severe shot peening (RSSP) treatments of cast iron and low-alloy steel, in which RSSP could cause a more regular surface state and thus further improve the fatigue life of the SSP specimens [5,6]. These results seem to indicate that surface defects have an important influence on the fatigue performance. Bagherifard et al. [5,42] also showed that the obvious relaxation of surface compressive residual stress had occurred on the SSP specimens, and the fatigue crack initiated from surface defects, although the GNS surface layer was formed on the surface of SSP treated specimens. Kim et al. [43] suggested that the presence of microdamage in the case of SP, such as surface and internal cracks, induced not only compressive residual stress relaxation on the surface but also fatigue life degradation. As can be observed in Fig. 10(a) and (b), after the residual stress relaxation experiment, compressive residual stress is relaxed by approximately 51.2% on the top surface layer of the USRP + P specimens. Accordingly, the fatigue life of the USRP + P+F specimens showed an 80.5% fatigue life reduction compared to the fatigue life of the USRP + P specimens. Hence, a conclusion might be reached that the compressive residual stress is a key factor in improving the fatigue performance. However, a considerable increase of almost 21.4 times in fatigue life is observed for the USRP + P+F specimens with respect to that of the BM specimens, which shows that the GNS surface layer and surface work hardening also have significant contributions to the improvement of fatigue life. It is noteworthy that the compressive residual stress remaining after the residual stress relaxation experiment should also have beneficial effects on the enhancement of the fatigue performance. Meanwhile, the GNS surface layer and surface work hardening have a synergistic effect on the enhancement of the fatigue performance. The compressive residual stress can retard the fatigue crack initiation and close the cracks [44]. Moreover, compressive residual stress can reduce the driving force for crack growth, resulting from the balance of a portion of applied tensile stress [40]. Additionally, compressive residual stress will reduce the stress intensity factor, thereby effectively decreasing the crack growth rate [40]. In the coarse grain BM, the movement and rearrangement of dislocations leads to the formation of persistent slip bands (PSBs) [45]. The extrusions and intrusions of PSBs result in fatigue crack initiation [46,47]. The GNS surface microstructure and surface work hardening will help prevent the nucleation of fatigue cracks and the formation of PSBs and then exert a significant beneficial effect on crack initiation [8]. At the same time, the GNS surface layers are expected to provide an ideal microstructure for delaying or preventing crack growth under the fatigue condition [40,48,49].

As discussed above, the beneficial effects of the USRP on improving fatigue performance could be ascribed to the comprehensive effects of the compressive residual stress, GNS surface layer and surface work hardening. The compressive residual stress is a dominant factor in determining the fatigue performance. Simultaneously, the GNS surface layer and surface work hardening have the synergistic effects that accompany the effect of compressive residual stress. This is different from the research results about SMRT on the fatigue behavior of GNS AISI 316 L stainless steel [8], which revealed that the GNS surface layer was a key factor in improving the fatigue performance.

4.3. Fracture morphology

Fatigue fracture morphologies of the BM, BM + F, USRP, USRP + P and USRP + P+F specimens at a maximum stress level of 700 MPa are shown in Fig. 11. All of the fatigue fractures displayed three typical zones (the insets in Fig. 11(a)-(e)): the crack initiation zone (Region 1), the fast propagation zone (Region 2), and the instantaneous fracture zone (Region 3) [50]. In addition, all of the fatigue fractures present the characteristics of a single fatigue crack source. The fatigue cracks of the BM and BM + F specimens both initiate from the surface at some stress concentration zone, such as mechanical polishing marks and surface defects, and the fatigue cracks expand immediately inward along a fan-shaped path until early fatigue fracture (Fig. 11(a) and (b)). The main reason for this behavior is that the maximum alternating stress exists on the specimen surface under rotating bending fatigue loading conditions. As shown in Fig. 11(c), the fatigue crack initiation of the USRP specimen with a GNS surface layer starts from surface defects due to stress concentration phenomena of those defects, which was expected to initiate from subsurface layer. This experimental result is consistent with results obtained by other scholars [5,6,14]. Moreover, in terms of the USRP specimen, the early propagation direction of the fatigue crack has a certain inclination angle to the surface, which is directly related to the compressive residual stress in the surface layer [51]. After surface defects were removed using the polishing treatment, the fatigue crack source of the USRP + P specimen was shifted to subsurface with a depth approximately 780 μm from the surface (Fig. 11(d)). The shift of the fatigue crack source toward the subsurface is commonly accompanied by a significant increase in fatigue life [3]. In contrast, fatigue crack source of the USRP + P+F specimen shifted to the surface again after the residual stress relaxation experiment (Fig. 11(e)).

Fig. 11.   Fatigue fracture morphologies with different conditions under maximum stress level of 700 MPa.

5. Conclusions

In this study, the effects of the USRP on the fatigue behavior of Ti-6Al-4 V alloy were studied. On the basis of the experimental analysis, some conclusions can be drawn as follows:

(1) A GNS surface layer, with a surface equiaxed nanograin layer thickness of at least 10 μm, is synthesized on the surface of Ti-6Al-4 V alloy by mean of the USRP. The average grain size is 91.5 nm at 10 μm from the surface.

(2) Several surface defects, for instance surface delamination and fold defects, are observed on the surface of the USRP specimens. Therefore, the average surface roughness value Ra for the USRP specimens shows an increment of 72.6% compared to that of BM specimens. The maximum value of the microhardness exists in the surface (521 H K), more than 44.7% that of the BM value. Furthermore, the compressive residual stress layer is about 650 μm thick at the surface after the USRP. The value of surface compressive residual stress is 864.7 MPa.

(3) Fatigue results show that the fatigue performance is obviously enhanced not only in the LCF regime but also in the HCF regime for the USRP specimens, although those specimens present a higher surface roughness with some surface defects. The 107 cycle fatigue limit improves from 500 MPa for the BM specimen to 610 MPa for the USRP specimen, an increase of 22%.

(4) Under a maximum stress of 700 MPa, the fatigue life improves 23.5 times in the USRP compared with that of the untreated condition. Mechanical surface polishing treatment (USRP + P) further enhances the beneficial effects of the USRP on fatigue performance, with increments of 3.7 times and 113.8 times that of the USRP and BM specimens respectively. After the residual stress relaxation experiment, the corresponding fatigue life of the USRP + P+F condition presents a decrease of 80.5% compared to that of the USRP + P condition.

(5) The fatigue strengthening mechanism of the USRP on the Ti-6Al-4 V alloy could be ascribed to the comprehensive effects of the compressive residual stress, GNS surface layer and surface work hardening. The compressive residual stress is a key factor in determining the fatigue performance. Simultaneously, the GNS surface layer and surface work hardening have a synergistic effect that accompanies the effect of compressive residual stress.

Acknowledgement

The work is financially supported by the National Natural Science Foundation of China (No. 51771155).

The authors have declared that no competing interests exist.


/