- Email: [email protected]
Influences of single laser tracks' space on the rolling fatigue contact of gray cast iron
Optics & Laser Technology 72 (2015) 15–24
Contents lists available at ScienceDirect
Optics & Laser Technology journal homepage: www.elsevier.com/locate/optlastec
Influences of single laser tracks' space on the rolling fatigue contact of gray cast iron Zhi–kai Chen a, Ti Zhou a, Peng Zhang a, Hai-feng Zhang a,b, Wan-shi Yang a, Hong Zhou a,n, Lu-quan Ren c a
The Key Lab of Automobile Materials, The Ministry of Education, Jilin University, Changchun 130025, PR China The Department of Mechanical and Automotive Engineering, Changchun University, Changchun 130025, PR China c The Key Lab of Terrain Machinery Bionics Engineering, The Ministry of Education, Jilin University, Changchun 130025, PR China b
art ic l e i nf o
a b s t r a c t
Article history: Received 7 January 2015 Received in revised form 16 March 2015 Accepted 16 March 2015
To improve the fatigue wear resistance of gray cast iron, the surface is modified by Nd:YAG laser to imitate the unique surface of soil creatures (alternative soft and hard phases). After laser treatment, the remelting region is the named unit which is mainly characterized of compact and refinement grains. In the present work, the influence of the unit space on the fatigue wear resistance is experimentally studied. The optimum space is proven to be 2 mm according to the tested results and two kinds of delamination are observed on samples' worn surface. Subsequently, the mechanisms of fatigue wear resistance improvement are suggested: (i) for microscopic behavior, the bionic unit not only delays the initiation of microcracks, but also significantly obstructs the propagation of cracks; (ii) for macroscopic behavior, the hard phase resists the deformation and the soft phase releases the deformation. & 2015 Elsevier Ltd. All rights reserved.
Keywords: Rolling contact fatigue Laser surface remelting Unit space
1. Introduction The rolling contact fatigue (RCF) is a common failure type of the components which are subjected to the repeated stress, such as railroads, rolling guide ways, gears, etc. Various surface damages occur and also cracks develop in the machine units, thus leading to loss of serviceability of the units under roller contact. Typical damage in roller contact fatigue under dry friction or boundary lubrication includes pitting, spalling, and cracking and 'squat' like damage that in most cases evolves into a main crack through the roller body [1]. Since the failure of these components, it is necessary to change or scrap the damaged components for the safety and quality of production. Not only large cost would be spent on these works, but also huge waste would be caused. Therefore, it is necessary to improve the fatigue wear resistance (FWR) of materials based on the considerations of economical and effective. Previously, many papers have focused on the RCF behaviors of ferrous materials, such as rail steel [2] and nodular cast iron [3], and many works have been done to improve the FWR of ferrous materials, for example bearing steel [4] and austempered ductile iron [5], which are verified to thoroughly understand the RCF behaviors and significantly prolong the service life. However, few studies on gray cast iron (GCI) have been performed and the effective approaches to n
Corresponding author. Tel.: þ 86 431 8509 4427; fax: þ 86 431 8509 5592. E-mail address: [email protected] (H. Zhou).
http://dx.doi.org/10.1016/j.optlastec.2015.03.008 0030-3992/& 2015 Elsevier Ltd. All rights reserved.
promote the FWR of GCI have not been reported. GCI is one of the most widely used engineering materials with the characteristics of low cost, cast-ability, good wear resistance and toughness. The unique feature of GCI is prominence in shock-absorbance as flake graphite disperses in pearlite matrix. On the contrast, the existence of flake graphite leads GCI to be more complex than other ferrous materials because the concentrated stress prefers to occur at the graphite tip. Consequently, the improvement of FWR and the studying of RCF behaviors of GCI are imperative. In decades, Nd:YAG laser is widely used in industrial manufacture because it has a shorter wavelength and thus normally does not need to coat the surface that is to be treated compared with CO2 laser. Additionally, laser surface melting (LSM) has its advantages over the conventional surface processes, including minimum distortion, high hardness, narrow heat-affected zone and more easily controllable, and the energy can be delivered to component surfaces by fiber optics. According to Grum and Sturm [6], who had proven the harden mechanism to be the occurrence of refined martensite and ledeburite in melt zone, the laser processing therefore could be considered to be useful for surface modification. However, Iino and Shimoda [7] had proposed that the occurrence of tempered effect would be investigated as the overlapping laser track resulting in the reduction of microhardness, and therefore the way treated by laser on the whole surface is impossible to promote the FWR of GCI. Enlightened by surface of some soil creatures whose apparently common characteristic is soft–hard alternative, the group of Zhou
16
Z. Chen et al. / Optics & Laser Technology 72 (2015) 15–24
[8–10] has devoted to improving the materials' properties according to imitate the surface by laser surface remelting. Although many properties of materials have been confirmed to signally improve, including thermal fatigue resistance [8], wear resistance [9] and tensile resistance [10], the research works on FWR of GCI have not been carried out. The authors' previous works [11] have verified that the type of soft–hard alternative is positive to FWR and the distribution angles of units dramatically affect the FWR. The diverse angles of laser tracks with constant horizontal space corresponded to the change of hard phase space in rolling direction. However, the influence of the change on RCF has not been explicitly introduced in previous works. This paper aims to exclusively investigate the fatigue behavior of GCI with laser track in different spaces. The surface characteristic of treated sample is modified with laser tracks which are continuous in width while they are isolated in length of sample. Subsequently, according to observations and detections, the mechanisms of FWR improvement are discussed and the fatigue behaviors are suggested.
laser processing. Optical microscope, scanning electron microscopy (SEM) and X-ray diffraction (XRD) were utilized for these investigations. The microhardness was measured in loading of 0.2 kg with a holding time of 10 s by Vickers microhardness test (Model 5104, manufactured by Buehler Co. Ltd., USA). Additionally, the tensile test was performed on the sample with single laser track along tensile direction to learn the reinforcement of unit. Fatigue wear tests were carried out by home-made machine in the air atmosphere, as shown in Fig. 2. The rolling bearing and an eccenter were connected by a metal rod, and the samples were mounted at the back of loading station that was always supported by 10 rollers (φ5 11). The constant loading of 80 N would be transformed to samples' surface through the narrow contact area, in which the generated pressure was well below that necessary to cause yielding. When the eccenter rotated, the roller bearing was actuated on the samples' surface back and forth. The rotating speed of the eccenter was set at 30 rpm, corresponding to horizontal velocity (v) of the roller bearing expressed as
v = ω. l. sin φ
(1)
2. Experimental 2.1. Experimental materials For this work a common GCI, cut from machine tool by electric spark machine (DK7732) was used. And its chemical composition, microhardness and tensile properties are illustrated in Table 1. In terms of the original microstructure, the flake graphite is surrounded by pearlite and ferrite, and the percentage of pearlite is more than that of ferrite. What is more, the average width of flake graphite in 'A' type is about 4 mm and the length is uncertain. 2.2. Sample preparation Samples (120 mm 15 mm 5 mm) were cut by electric spark machine, and the corners were ground to arc-shape to embed into the load station of the wear experimental system. To enable the effect of space (s) on the RCF life of treated samples to be studied, the samples were divided into six groups according to the unit space and each group contained three indiscriminate samples for repetitive test. Among the classification, the group without treatment called smooth sample was marked No. 1 and the groups with units space (s) ranged between 1 mm and 5 mm were marked Nos. 2–6, respectively. The stripes unit perpendicular to the rolling direction was processed by Nd:YAG laser in the same parameters. The strips were to form the geometrical non-smooth surface to imitate the rough surface of soil animals. It should be noted that the remelting region was named units in the paper. The sketch of treated sample was shown in Fig. 1. As known, the percentage of treated region increased by the decrease of unit space. The surface of samples was performed by Nd:YAG laser with wavelength of 1.06 μm and a maximum export power of 300 W, using a circular pattern Gaussian beam and set under an argon gas shield with a flow of 5 L/min. The working-bench moved at 0.5 mm/s. The parameters of laser were the pulse duration of 3 ms, frequency of 7 Hz, and the energy of a single laser point was about 2.28 J. After laser deposition, the size parameters of unit were listed in Table 2. To identify the microstructure of the unit, the standard method of metallography was performed on the transverse section after the
where ω represented the angular velocity of drive wheel, l was the eccentric distance (35 mm) that indicated the amplitude of back and forth path to be 35 mm, and φ was the deflection angle ranging 0–360° as shown in Fig. 2. Depending on the calculation of v, the positive and negative were implied the forth and back of roller, respectively. Prior to wear tests, the specimens were mechanically polished with fine emery paper sequentially, followed by ultrasonic cleaning in acetone. The weights of samples were measured by the electronic balance with a precision of 0.1 mg. Every 10 h, the samples were interrupted at regular intervals for mass loss measurement, and the above procedure was followed besides polishing to record the mass until 60 h. Additionally, the final roughness of the specimens after
Fig. 1. The schematic of the treated sample and the form of unit distribution.
Table 2 Dimensions of unit's cross section. Width (mm)
Depth (mm)
Area (mm2)
Fit equation of unit's edge
17 0.1
0.337 0.03
0.22
Y ¼1.32x2 0.33
Table 1 Nominal chemical composition (wt%), the micro-hardness (HV0.2) and tensile properties (MPa) of the base material. C
Si
Mn
P
S
Fe
Hardness
Tensile strength
Fracture strength
3–3.2
1.5–1.8
0.7–0.9
o 0.15
o 0.12
Bal.
260
135
90
Z. Chen et al. / Optics & Laser Technology 72 (2015) 15–24
17
Fig. 2. The sketch of the man-made fatigue wear system.
polishing and before wear test was recorded. At the end of test, the average mass loss of individual group was utilized to evaluate the FWR and the worn surface was observed by optical micrograph and 3-D surface profile scanning. 2.3. Finite element analysis What is more, to explain the FWR future, the numerical simulation was utilized to exhibit the distribution of stress on the surface under the experimental conditions. To simplify the calculation, the single roller was taken into account with loading of 8 N, and the roller was regarded as rigid body. For the calculation, only the pure rolling contact cases were considered, i.e., the friction coefficient between the sample and roller was assumed to be zero [12] and the linear velocity was equal to the horizontal velocity of rolling bearing. The parameters used in the numerical simulation are listed in Table 3. The refined meshes were defined in the remelting region to improve the accuracy of calculation. It should be noted that the residual stress caused by the laser processing [13,14] was not considered in the calculation.
3. Results 3.1. The microstructure and micro-hardness of bionic units After laser treatment, the microstructure of bionic unit is observed. Fig. 3a shows the cross-sectional SEM morphology of unit with low magnification and it is easy to identify three zones: transformed zone (TZ), remelting zone (RZ) and matrix which are shown in Fig. 3b, c and d with high magnification, respectively. As shown in Fig. 3b, the interface is clearly observed because of the distinct microstructure, and the RZ and TZ are combined by metallurgical bonding. Additionally, there is a little un-dissolution G in TZ and it consists of acicular martensite. Fig. 3c demonstrates the micrograph of microstructure in RZ, in which a complete dissolution of the G and re-solidification of the ledeburite are presented. As a supplement, ledeburite, the mixture of cementite and martensite, reveals the reticular structure. Fig. 3d indicates that the original phase of matrix is mainly characterized of the pearlite (P) which is surrounded by ferrite (F) and flake graphite (G). After laser irradiation, the phase compositions are detected by XRD, as presented in Fig. 4. The diffractogram shows that the new phases are formed, such as martensite, austenite (retain austenite in room temperature) and Fe7C3. As the heat is transformed in depth, the absorbed heat by material element is gradually reduced in depth leading to the absolute melting zone and partially melted zone, which are the RZ and TZ respectively mentioned above. Actually, a zone in which the phase transition takes place without remelting is named heat affect zone (HAZ), and it is categorized into TZ. In RZ, F experienced the hypo-eutectic reaction to transform to retained austenite (Ar) and martensite (M) during the remelting process. Moreover, the cementite close to rich carbon
Table 3 The parameters used in numerical simulation. Elasticity modulus (GPa) PM 130 Unit 150
Passion ratio
Density (g/cm3) Friction coefficient
0.23 0.25
7.0 7.4
0
zone is transformed to metastable carbides (Fe7C3) at the duration of resolidification. Material cannot be absolutely melted because of the insufficient heat that is proven by the retain graphite and the low cooling rate then induces the formation of coarse martensite in TZ. Another aspect of the pattern exhibits broadening width of diffraction peaks, demonstrating crystal grains in the whole unit zone very fine and compact. It can be explained as a considerable amount of high cooling rate during solidification caused by the laser processing. Due to phase transition, the micro-hardness is considered to be remarkably various upon the microstructure. The average surface micro-hardness is measured to be 680 HV0.2. The measurement of unit is depicted in Fig. 5 as a function of the distance from surface. The plot feature demonstrates that the hardness decreases from the top to the bottom of unit since the energy is gradually reduced with the distance from surface increases. According to the plot, the microhardness in RZ and TZ is markedly different in which the hardness in RZ varies from 620 to 750 HV0.2 and in TZ ranges from 320 HV0.2 to 620 HV0.2 with rapid decrease. The exhibition of different values of micro-hardness is with the occurrence of fine grain in RZ while the relatively coarse particle is observed in TZ, as shown in Fig. 3b. By calculation, the average microhardness of unit is about 560 HV0.2, which is increased by 115%, in contrast to that of the substrate with the average microhardness of 260 HV0.2. It is interesting to point that the peak hardness occurs at the shallow layer, about 0.08 mm. In terms of the fact that the existence of carbide could significantly increase the microhardness, it is reasonable to conjecture the formation of a razor-thin decarburized layer on surface. In conclusion, the existence of refined grain, arising from rapid solidification of the melt, can obviously increase the hardness of units. 3.2. Results of fatigue wear tests 3.2.1. The mass loss and the mass loss rate of samples The histogram shown in Fig. 6a is the average recorded mass losses for groups at the end of test. According to the figure, it can be clearly seen that the mass loss (ML) of untreated sample is always higher than that of the treated samples which intensively demonstrates the laser treatment to be effective on FWR improvement. Additionally, among the treated samples, the ML decreases with the decrease of unit space, attaining a minimum at s¼2 mm, while for less values of unit space (s¼1), the ML increases slightly. Therefore, it can be inferred that there must be a balance between the development of surface damages and the space of reinforced region. With respect to Fig. 6b, all the plots suggested a maximum ML rate at 20 h
18
Z. Chen et al. / Optics & Laser Technology 72 (2015) 15–24
Fig.3. The micrographs of sample cross-section: (a) The cross-sectional SEM micrograph of unit, (b) the interface of RZ and TZ (1000 ), (c) the microstructure of RZ (2000 ) and (d) the matrix (1000 ).
Fig. 5. The plot of unit's hardness in depth in loading of 0.2 kg with a holding time of 10 s. Fig. 4. The XRD diffractogram of smooth and treated samples.
and the stable wear behaviors until the end. This can be explained as the high initial roughness of the sample's surfaces significantly affects the fatigue wear behavior as the contact pressure profile [15] or the occurrence of weak region during the sample preparation, and the serious wear is removed under the steady-state wear condition as a result of material removal or plastic deformation [16]. For another aspect, the peak ML rates of treated samples are always less than that of parent material. Similarly, the peak ML rate decreases in sequence with space decrease, ranging 5–2 mm and however the peak ML rate still increases for s¼1 mm. Namely, the sample with less ML shows the better FWR because the metal removal acts as the result of fatigue failure. According to the worn surface of samples by 3-D surface profile scanning, it is obviously seen that no serious damages (spalling, large delamination and metallic flake) occur at the unit part in which only less and slight pits are detected (Fig. 7a). On the other hand, the serious defects are investigated on the original part of
treated sample (Fig. 7b) and however they are much slighter than that on untreated sample (Fig. 7c). Comparing the original part with unit part of treated sample, more and larger particles are torn fallen from the surface of original part. In conclusion, since the existence of unit, the surface damages are remarkably reduced and the main defects focus on the untreated parts for treated samples. 3.2.2. The enhanced degree of fatigue wears resistance The average Ml of samples in each group was chosen for tests under the same test conditions. Since the ML of all treated samples (MLtreated) are less than that of the smooth one (MLuntreated), the ML decrease percent (here-in after referred to as a 'MLDP') is taken to assess the enhanced degree of FWR for different groups. The calculation formula is as follows:
MLDP =
ML treated−ML untreated × 100% ML untreated
(2)
Z. Chen et al. / Optics & Laser Technology 72 (2015) 15–24
19
Fig. 6. The results of fatigue wear tests: (a) the mass losses and (b) the mass loss rates of samples vs. the test time.
Fig. 7. The worn surface profile by optical microscope: (a) unit region, (b) untreated part of treated sample and (c) untreated sample.
The MLDPs given by Eq. (2) show that the value of Nos. 2, 3, 4, 5 and 6 are 33.1%, 40.2%, 34%, 25.4% and 13.4% respectively. According to the dates, it can be seen that the MLDP of No. 3 is the biggest to imply the optimum FWR, attaining about 40.2%. What is more, the degree of FWR improvement is generally remarkable for the large treated part, while it would indicate the opposite tendency when the treated area reduces to a threshold value (s¼ 2 mm). Therefore, we deduce that there are some adverse factors to affect the FWR against the treated area. In terms of the MLDPs, it is confirmed that laser surface treatment is significantly useful for the improvement of fatigue wears resistance. Moreover, it intensively
indicates the space of units to be an important factor on the improvement of FWR. 3.2.3. Fatigue wear behavior The curves shown in Fig.8 are drawn with the stress and strain as functions of the distance in width (along AA′) to infer the distributions of strain and stress introduced by finite element analysis. It definitely validates that the maximum stress and strain are shown at both the sides of roller where the serious damages prefer to emerge. According to the worn track of insertion in Fig. 8, the dark area occurs at the roller's edges which have been ensured to be the serious damage and gradually weaken to the center of the contact part (gray
20
Z. Chen et al. / Optics & Laser Technology 72 (2015) 15–24
area) in width. The worn surface is well agreed with the simulation analysis. With respect to the roughness, the measured values of dark area and gray area are ranged 285–318 nm and 295–325 nm, respectively, which significantly suggested a decrease compared to the original surface roughness of 417 nm. What is more, since the stress at gray area is smaller than that at the dark area during the rolling contact, the gray area could be regarded as the precursor of dark area. And the roughness of dark area and gray area is considered to be almost similar that proves the roughness of worn track to be about 300 nm at the duration of steady-state wear condition. Generally, the common failures of RCF are in variations of types, including cracks, pits, spalling and delamination. Complementarily, two types of delamination are suggested in this work: the one is characterized by the part edge to be cliff and the rest edge to be inclined from the bottom to surface which results from the nearly closed crack defined as type I; another one is all the edge to be cliff resulting from the entirely closed crack defined as type II. Fig. 9 displays the worn surface observations of samples at the end of wear test. It can be seen that serious damages (cracks, spalling and large scale delamination) are observed on the surface of untreated sample (Fig. 9a). In Fig. 9b, the delamination of type I is clearly illustrated and
Fig.8. The plots of strain and stress of the surface against the distance in width (AA'). The insertion of figure is the image of the appearance of worn surface.
it looks like the flake metallics are torn from the surface. Looking at the tip region of the principal crack (indicated by the solid arrow in Fig. 9b), the partial crack removal is evident. Furthermore, the shape of type I is revealed in a fanned-out appearance which is well agreed with the points of Zakharov and Goryacheva [17] who proposed that microcracks developed at a small angle to the surface resulting in fine angled cracks developing on circumference of the defect. Additionally, no spalling or delamination is observed at the center of contact while the so-called defects appearing on the surface are almost cracks, owing to the stress on this region being weaker than that at the both sides (see Fig. 9c). Therefore, it is believable that serious defects (spalling, delamination) are resulted from the propagation and nucleation of cracks by repetitive rolling contact.
4. Discussions 4.1. Resistance of microcrack initiation by treated unit As stated, serious defects are resulted from the microcracks' behaviors, such as nucleation, propagation and closure under rolling contact. Thus, delaying the initiation of crack is considered to be an effective way to improve FWR of GCI. Since the flake graphite is regarded as the natural hole or crack, their presence can lead to the notch effect in the matrix. Previous works have pointed that fatigue failures in service invariably occur at stress concentrations [18], and therefore at the tips of graphite where concentration stress generally prefers to arise is the origin of micro-crack, as shown in Fig. 10a. Graphite is relatively softer than the matrix materials and it cyclically softened to the matrix resulting in a concentration of the plastic strain. Moreover, the graphite lamellae act as stress raiser producing self-microcracking, or interface debonding at very low stress levels and plastic deformation in the matrix [19]. Hence, graphite is a potential threat to GCI component in RCF conditions. The performance of remelting zone (Fig. 2c) is in substantial agreement with that of Benyounis [20] who proposed the graphite to be completely dissolved after the laser processing. By disappearance of graphite, the FWR of material is significantly improved. Moreover, due to the
Fig. 9. Worn surface after fatigue wear test: (a) dark area of untreated sample, (b) dark area of treated sample and (c) gray area of untreated sample.
Z. Chen et al. / Optics & Laser Technology 72 (2015) 15–24
21
Fig. 10. (a) The origin of crack initiation and (b) the route of crack propagation.
phase transformation and refinement of grain, the treated portions with high hardness exhibit excellent deformation resistance and meanwhile the matrix acts as a soft phase to cushion shear stress coming from the roller, which reduces the possibility of stress concentrations. 4.2. Resistance of microcrack propagation by treated unit As graphite plays the role of a bridge to connect the adjacent micro-cracks (Fig. 10b), the disappearance of graphite is positive to the resistance of crack propagation. The unit containing amount of fine crystal combined with the quantity of grain boundary increase is also significantly beneficial for the resistance of crack propagation, as described in Ref. [21] that the fatigue cracks' sensitivity depends on the matrix strength and the grain boundary can prevent crack propagation. Reasonably, the fatigue cracks propagate in a zig-zag manner which is difficult to cross or along the fine and reinforced crystal particle as the zig-zag route. In comparison with linear propagation, the crack much more frequent deflection of prolonged route increases the absorbed energy and will contribute to greatly lower fatigue crack growth rate. In addition, owing to the residual stress incurred by accumulated plastic deformation which drives the propagation of cracks [22], the increase of plastic deformation resistance which corresponds to enhanced yield strength of materials can markedly postpone the propagation behavior. After laser remelting, refinement of the interlamellar spacing increases the yield strength of the steel according to Hall– Petch type relationship [23] as
σs = σ0 + K y·d−1/2
(3)
where ss is the yield stress, s0 represents the lattice friction stress required to move individual dislocations, ky is a constant value which depends on the contribution of measured grain boundary to yield strength, and d is the average diameter of grains. According to the formula, the yield strength of materials increases in inverse to grains' dimension. Indeed, the tensile test shows that the tensile strength and fracture strength of treated one are 195 MPa and 145 MPa which increase by 44.4% and 61.1%, respectively, compared to those of original sample. Due to grain refinement, the unit acts as a phase to contribute to high strength of sample's surface, that is, the crack propagation is evidently hindered by unit. With respect to the hindering effect of unit during the fatigue wear test, four types of cracks are revealed for discussion and marked a, b, c and d, respectively, as presented in Fig. 11. When the crack 'a' arrives the unit, the crack cannot be driven to further propagate as inadequate force and is subsequently stopped at the edge of unit. Additionally, when the crack ‘c’ reaches the region of unit, the crack is divided into two branches as response because the refinement grain
Fig. 11. The principle of the unit hinders the propagation of crack.
in unit makes the crack difficult to through the unit. The branches however propagate in different directions in which the left one connects the crack 'b' to form the semi-closed crack in the matrix region and the right one connects the crack 'd' to afford sufficient energy to deeply propagate in the unit. Consequently, the unit makes cracks propagation complex which contributes to delay the cracks' propagation compared with the linear manner in matrix. In conclusion, the initiation and propagation of cracks from the substrate region are significantly suppressed by bionic unit, which is considered to be the reason of FWR enhancement by LSM. 4.3. Behaviors of fatigue failure According to Fig. 7, a few and slight damage is observed on unit region while the large scale spalling and delamination are detected on the matrix region of both treated and untreated samples. During the fatigue wear processing, the cracks are driven by the accumulation force to form circular crack resulting in failure. Since the coarse grain size and the flake graphite are distributed in the matrix, the cracks are relatively easy to gather to longer cracks leading to the subsequently serious damages. By the propagation of cracks, the damage is resulted from the renegade metallic from the matrix along the crack when the cracks are associated to incompletely closed crack. This type of defect is defined as type I and the depth is about 4000 nm, as shown in Fig. 12a in detail. The phenomenon has been reported that the first directed towards the depth with an inclination and a branch begins to rise towards the contact surface inclined by about 70° to the initial direction, causing material removal [18]. Differently, as the existence of refined grain size in unit, the numerous grain boundaries make the propagation of cracks harder and more frequent deflection which is prone to form entirely closed crack. The followed occurrence of defect is resulted from the completely closed crack which is called
22
Z. Chen et al. / Optics & Laser Technology 72 (2015) 15–24
Fig. 12. The different types of delamination on surface: (a) type I of delamination on untreated sample and (b) type II of delamination on unit region.
Z. Chen et al. / Optics & Laser Technology 72 (2015) 15–24
Fig. 13. The schematic of the occurrence of non-uniform deformation: (a) before and (b) after rolling contact.
delamination of type II, as Fig. 12b explicitly shows. A comparison of two types of delamination is that type II not only takes more time and energy to form the closed crack, but is also smaller and shallower (about 1500 nm in depth) in dimension than type I in RCF failure. The treated region is proportional to the enhancement degree of FWR in theory, and the experimental results however do not agree with the assumption. According to the previous research works that the RCF results from the plastic flow [24] and the maximum shear stress occurs at sub-surface [25], the fatigue micro-crack would be considered to be initial at the sub-surface. As known, the plastic deformation takes place with the exhaustion of material ductility and the degree of plastic deformation is also diverse with the difference of material ductility. Fig. 13 demonstrates the status of parent material and treated portion before (Fig. 13a) and after (Fig. 13b) rolling contact. Since the surface is modified with the reinforced unit, the assumption of non-uniform deformation between hard phase (unit) and soft phase (matrix) is reasonable in RCF condition, and the maximum non-uniform deformation does generally occur at the position where the maximum shear stress emerges beneath surface. By the occurrence of non-uniform deformation, the huge residual stress would take place at the interface where the micro-crack preferentially occurs after thousands of repetitive rolling contact. Once the crack forms, it will be driven to grow at interior of sample until a critical length and then toward the surface resulting in the surface damages of metallic debris removal (mass loss). Therefore, it is necessary to decrease the degree of the non-uniform deformation for reducing the possibility of crack initiation. As increase of unit yield stress, the degree of non-uniform deformation would be aggravated under the same magnitude of shear stress and hence relatively large fraction of soft phase is designed to reduce or even eliminate the phenomenon of non-uniform deformation. In this paper, different fractions of soft phase come true by adjusting the spaces of units to attain the balance of plastic deformation. Additionally, formation of soft zone caused by tempering effect and low compressive residual stresses in the overlapping tracks are reported to produce a long narrow crack along the edge of the overlapping tracks and cause premature failure with the adjustment of the unit track space [7,26]. The discussion can well explain the result that the ML of No. 2 is higher than that of No. 3, although No. 2 is with larger treated percentage. By combination of the discussion and result of ML, it can be concluded that when the s is more than 2 mm, the percentage of remelting area (reinforced region) is the dominant factor on the FWR improvement, while the development of damage becomes the prominent factor for so2 mm. Hence, the influence of the space is considered to be reasonable on RCF and the optimum space is 2 mm according to the wear results. What is more, Young et al. stated that during plastic deformation, the strength of the ductile phase is enhanced via plastic constraint by the surrounding reinforced hard phase [27] and the deformation of hard phase can be released due to
23
the excellent ductility of soft phase. This structure in this case that the soft substrate regions are mingled by the hard phase with relatively higher deformation resistance is similar to and accords the principle of bionic approach. Halling [28] suggested that a stick area exists in the contact area in which the particles would be broken away from surface by stick force. After LSM, because of the increased interaction of refined crystal, the stick force is suggested to be released or exhausted more or less. Thus, the fatigue life is considered to be prolonged. What is more, the traction coefficient is subsequently reduced after laser treatment no matter under the lubricated condition or the dry condition, which has been indicated by Franklin et al. [2]. According to Fletcher [25] who manifests that the maximum orthogonal shear stress increases with the increase of traction coefficient, the conclusion therefore can be stated that the way of LSM could significantly reduce the shear stress of materials correspondingly to improve the FWR. In fact, the optimum space is obtained just under the laboratorial condition. In actual application, the shear stress is far more than that in laboratory which interprets that the space is not commonly used in various applied loadings, because of the reason that gray cast iron presented a non-linear load–displacement ratio even at small strain levels, implying that Hooke's law cannot be applied to these materials [29]. The space of laser track optimization in this paper only gives effective surface morphology guidance for processing related parts and suggests that the effect factor should be taken into consideration in engineering.
5. Conclusion The following conclusions can be drawn from this investigation: After the laser processing, the treated region with an average hardness of 680 HV0.2 is created, which is increased by 162%, in contrast to that of the substrate with the average microhardness of 260 HV0.2. The remelting zone is mainly characterized of compact and refinement grains, such as retain austenite, martensite, Fe3C, etc. and there is no graphite in melt pool. Additionally, the tensile properties of treated sample are significantly enhanced such that tensile strength and fracture strength of the treated one are 195 MPa and 145 MPa which are to increase by 44.4% and 61.1%, respectively, compared to those of the original sample. For the result of Finite Element, the stress arising at the both sides of contact region is relatively higher than that at the center part. The worn surface profile agrees well with simulations that serious damages occur at both the sides of contact region while a number of cracks take place at the center of contact. Additionally, the roughness of worn track is about 300 nm at the duration of steady-state wear condition. According to the MLDP, the optimum unit space is 2 mm, which attains 40.2%. Depending on the worn surface, the defects on the matrix of treated samples are lesser than that on original sample, but it is more serious than that on unit region. However, it cannot suggest that the FWR is in direct proportion to the treated area. When the s is more than 2 mm, the percentage of remelting area (reinforced region) is the dominant factor, while the development of damage becomes the prominent factor for s o2 mm. For fatigue failures, there are two types of delamination in the present work: the one is characterized by the part edge to be cliff and the rest edge to be inclined from the bottom to surface and another one is all the edges to be cliff. Subsequently, the mechanisms of FWR improvement according to the results are pointed out that (i) for microscopic behavior, the bionic units not only delay the initiation of microcracks, but also significantly obstruct the propagation of cracks; (ii) for macroscopic behavior,
24
Z. Chen et al. / Optics & Laser Technology 72 (2015) 15–24
the hard phase resists the deformation and the soft phase releases the deformation.
Acknowledgment This article was supported by Project 985 – High Performance Materials of Jilin University, the Project 985 – Bionic Engineering Science and Technology Innovation and National Natural Science Foundation of China (No. 51275200). Moreover, the authors would like to express their thanks to the conscientious editor Marc Sentis and the anonymous reviewers on the manuscript revision.
References [1] Datsyshyn O, Panasyuk V. Pitting of the rolling bodies contact surface. Wear 2001;25:1347–55. [2] Franklin F, Weeda GJ, Kapoor A, Hiensch E. Rolling contact fatigue and wear behaviour of the infrastar two-material rail. Wear 2005;258:1048–54. [3] Šamec B, Potrč I, Šraml M. Low cycle fatigue of nodular cast iron used for railway brake discs. Eng Fail Anal 2011;18:1424–34. [4] Polonsky I, Chang T, Keer L, Sproul W. A study of rolling-contact fatigue of bearing steel coated with physical vapor deposition TiN films: coating response to cyclic contact stress and physical mechanisms underlying coating effect on the fatigue life. Wear 1998;215:191–204. [5] Colombo DA, Echeverría MD, Laino S, Dommarco RC, Massone JM. Rolling contact fatigue resistance of PVD CrN and TiN coated austempered ductile iron. Wear 2013;308:35–45. [6] Grum J, Šturm R. Comparison of measured and calculated thickness of martensite and ledeburite shells around graphite nodules in the hardened layer of nodular iron after laser surface remelting. Appl Surf Sci 2002;187:116–23. [7] Iino Y, Shimoda K. Effect of overlap pass tempering on hardness and fatigue behaviour in laser heat treatment of carbon steel. J Mater Sci Lett 1987;6:1193–4. [8] Tong X, Zhou H, Zhang ZH, Sun N, Shan HY, Ren LQ. Effects of surface shape on thermal fatigue resistance of biomimetic treated cast iron. Mater Sci Eng A 2007;467:97–103. [9] Zhou H, Guo QC, Lin PY, Zhang W, Zhang XL, Ren LQ. Influence of H13 steel unit on wear behavior of vermicular cast iron. Appl Surf Sci 2008;255:3394–9. [10] Wang CW, Zhou H, Zhang ZH, Jing ZJ, Cong DL, Meng C, Ren LQ. Tensile property of low carbon steel with gridding units. Appl Surf Sci 2013;273:128– 34.
[11] Chen ZK, Lu SC, Song XB, Zhang HF, Yang WS, Zhou H. Effects of bionic units on the fatigue wear of gray cast iron surface with different shapes and distributions. J Opt Laser Technol 2015;66:166–74. [12] Zhang XC, Xu BS, Tu ST, Xuan FZ, Wang HD, Wu YX. Fatigue resistance and failure mechanisms of plasma-sprayed CrC–NiCr cermet coatings in rolling contact. Int J Fatigue 2009;31:906–15. [13] Grum J, Šturm R. Influence of laser surface melt-hardening conditions on residual stresses in thin plates. Surf Coat Technol 1998;100–101:455–8. [14] Grum J, Šturm R. Influence of laser remelting process parameters on residual stresses in nodular cast iron 500-7. Mater Manuf Process 2000;15:815–27. [15] Donzella G, Faccoli M, Ghidini A, Mazzù A, Roberti R. The competitive role of wear and RCF in a rail steel. Eng Fract Mech 2005;72:287–308. [16] García A, Cadenas M, Fernández MR, Noriega A. Tribological effects of the geometrical properties of plasma spray coatings partially melted by laser. Wear 2013;305:1–7. [17] Zakharov SM, Goryacheva IG. Rolling contact fatigue defects in freight car wheels. Wear 2005;58:1142–7. [18] Taylor D, Hughes M, Allen D. Notch fatigue behaviour in cast irons explained using a fracture mechanics approach. Int J Fatigue 1996;18:439–45. [19] Bertolino G, Perez-Ipiña JE. Geometrical effects on lamellar graycast iron fracture toughness. J Mater Process Technol 2006;179:202–6. [20] Benyounis KY, Fakronb OMA, Abboudb JH, Olabic AG, Hashmic MJS. Surface melting of nodular cast iron by Nd–YAG laser and TIG. J Mater Process Technol 2005;170:127–32. [21] Ochi Y, Masaki K, Matsumura T, Sekino T. Effect of shot-peening treatment on high cycle fatigue property of ductile cast iron. Int J Fatigue 2001;23:441–8. [22] Ringsberg J, Loo-Morrey M, Josefson B, Kapoor A, Beynon JH. Prediction of fatigue crack initiation for rolling contact fatigue. Int J Fatigue 2000;22:205–15. [23] Hugo RC, Kung H, Weertman JR, Mitra R, Knapp JA, Follstaedt DM. In situ TEM tensile testing of DC magnetron sputtered and pulsed laser deposited Ni thin films. Acta Mater 2003;51:1937–43. [24] Garnham JE, Davis CL. The role of deformed rail microstructure on rolling contact fatigue initiation. Wear 2008;265:1363–72. [25] Fletcher DI. The influence of lubrication on the fatigue of pearlitic rail steel [Ph.D. thesis]. UK: Department of Mechanical Engineering, University of Sheffield; 1999. [26] Van Brussel B, Hegge H, De Hosson JTM, Delhez R, De Keijser TH, Van der Pers N. Development of residual stress and surface cracks in laser treated low carbon steel. Scr Metall Mater 1991;25:779–84. [27] Young CH, Bhadeshia H. Strength of mixtures of bainite and martensite. Mater Sci Technol 1994;10:209–14. [28] Halling J. Principles of tribology. New York: The Macmillan; 1975. p. 174–201. [29] Haenny L, Zambelli G. The stiffness and modulus of elasticity of graycast irons. J Mater Sci Lett 1983;2:239–42.
Contents lists available at ScienceDirect
Optics & Laser Technology journal homepage: www.elsevier.com/locate/optlastec
Influences of single laser tracks' space on the rolling fatigue contact of gray cast iron Zhi–kai Chen a, Ti Zhou a, Peng Zhang a, Hai-feng Zhang a,b, Wan-shi Yang a, Hong Zhou a,n, Lu-quan Ren c a
The Key Lab of Automobile Materials, The Ministry of Education, Jilin University, Changchun 130025, PR China The Department of Mechanical and Automotive Engineering, Changchun University, Changchun 130025, PR China c The Key Lab of Terrain Machinery Bionics Engineering, The Ministry of Education, Jilin University, Changchun 130025, PR China b
art ic l e i nf o
a b s t r a c t
Article history: Received 7 January 2015 Received in revised form 16 March 2015 Accepted 16 March 2015
To improve the fatigue wear resistance of gray cast iron, the surface is modified by Nd:YAG laser to imitate the unique surface of soil creatures (alternative soft and hard phases). After laser treatment, the remelting region is the named unit which is mainly characterized of compact and refinement grains. In the present work, the influence of the unit space on the fatigue wear resistance is experimentally studied. The optimum space is proven to be 2 mm according to the tested results and two kinds of delamination are observed on samples' worn surface. Subsequently, the mechanisms of fatigue wear resistance improvement are suggested: (i) for microscopic behavior, the bionic unit not only delays the initiation of microcracks, but also significantly obstructs the propagation of cracks; (ii) for macroscopic behavior, the hard phase resists the deformation and the soft phase releases the deformation. & 2015 Elsevier Ltd. All rights reserved.
Keywords: Rolling contact fatigue Laser surface remelting Unit space
1. Introduction The rolling contact fatigue (RCF) is a common failure type of the components which are subjected to the repeated stress, such as railroads, rolling guide ways, gears, etc. Various surface damages occur and also cracks develop in the machine units, thus leading to loss of serviceability of the units under roller contact. Typical damage in roller contact fatigue under dry friction or boundary lubrication includes pitting, spalling, and cracking and 'squat' like damage that in most cases evolves into a main crack through the roller body [1]. Since the failure of these components, it is necessary to change or scrap the damaged components for the safety and quality of production. Not only large cost would be spent on these works, but also huge waste would be caused. Therefore, it is necessary to improve the fatigue wear resistance (FWR) of materials based on the considerations of economical and effective. Previously, many papers have focused on the RCF behaviors of ferrous materials, such as rail steel [2] and nodular cast iron [3], and many works have been done to improve the FWR of ferrous materials, for example bearing steel [4] and austempered ductile iron [5], which are verified to thoroughly understand the RCF behaviors and significantly prolong the service life. However, few studies on gray cast iron (GCI) have been performed and the effective approaches to n
Corresponding author. Tel.: þ 86 431 8509 4427; fax: þ 86 431 8509 5592. E-mail address: [email protected] (H. Zhou).
http://dx.doi.org/10.1016/j.optlastec.2015.03.008 0030-3992/& 2015 Elsevier Ltd. All rights reserved.
promote the FWR of GCI have not been reported. GCI is one of the most widely used engineering materials with the characteristics of low cost, cast-ability, good wear resistance and toughness. The unique feature of GCI is prominence in shock-absorbance as flake graphite disperses in pearlite matrix. On the contrast, the existence of flake graphite leads GCI to be more complex than other ferrous materials because the concentrated stress prefers to occur at the graphite tip. Consequently, the improvement of FWR and the studying of RCF behaviors of GCI are imperative. In decades, Nd:YAG laser is widely used in industrial manufacture because it has a shorter wavelength and thus normally does not need to coat the surface that is to be treated compared with CO2 laser. Additionally, laser surface melting (LSM) has its advantages over the conventional surface processes, including minimum distortion, high hardness, narrow heat-affected zone and more easily controllable, and the energy can be delivered to component surfaces by fiber optics. According to Grum and Sturm [6], who had proven the harden mechanism to be the occurrence of refined martensite and ledeburite in melt zone, the laser processing therefore could be considered to be useful for surface modification. However, Iino and Shimoda [7] had proposed that the occurrence of tempered effect would be investigated as the overlapping laser track resulting in the reduction of microhardness, and therefore the way treated by laser on the whole surface is impossible to promote the FWR of GCI. Enlightened by surface of some soil creatures whose apparently common characteristic is soft–hard alternative, the group of Zhou
16
Z. Chen et al. / Optics & Laser Technology 72 (2015) 15–24
[8–10] has devoted to improving the materials' properties according to imitate the surface by laser surface remelting. Although many properties of materials have been confirmed to signally improve, including thermal fatigue resistance [8], wear resistance [9] and tensile resistance [10], the research works on FWR of GCI have not been carried out. The authors' previous works [11] have verified that the type of soft–hard alternative is positive to FWR and the distribution angles of units dramatically affect the FWR. The diverse angles of laser tracks with constant horizontal space corresponded to the change of hard phase space in rolling direction. However, the influence of the change on RCF has not been explicitly introduced in previous works. This paper aims to exclusively investigate the fatigue behavior of GCI with laser track in different spaces. The surface characteristic of treated sample is modified with laser tracks which are continuous in width while they are isolated in length of sample. Subsequently, according to observations and detections, the mechanisms of FWR improvement are discussed and the fatigue behaviors are suggested.
laser processing. Optical microscope, scanning electron microscopy (SEM) and X-ray diffraction (XRD) were utilized for these investigations. The microhardness was measured in loading of 0.2 kg with a holding time of 10 s by Vickers microhardness test (Model 5104, manufactured by Buehler Co. Ltd., USA). Additionally, the tensile test was performed on the sample with single laser track along tensile direction to learn the reinforcement of unit. Fatigue wear tests were carried out by home-made machine in the air atmosphere, as shown in Fig. 2. The rolling bearing and an eccenter were connected by a metal rod, and the samples were mounted at the back of loading station that was always supported by 10 rollers (φ5 11). The constant loading of 80 N would be transformed to samples' surface through the narrow contact area, in which the generated pressure was well below that necessary to cause yielding. When the eccenter rotated, the roller bearing was actuated on the samples' surface back and forth. The rotating speed of the eccenter was set at 30 rpm, corresponding to horizontal velocity (v) of the roller bearing expressed as
v = ω. l. sin φ
(1)
2. Experimental 2.1. Experimental materials For this work a common GCI, cut from machine tool by electric spark machine (DK7732) was used. And its chemical composition, microhardness and tensile properties are illustrated in Table 1. In terms of the original microstructure, the flake graphite is surrounded by pearlite and ferrite, and the percentage of pearlite is more than that of ferrite. What is more, the average width of flake graphite in 'A' type is about 4 mm and the length is uncertain. 2.2. Sample preparation Samples (120 mm 15 mm 5 mm) were cut by electric spark machine, and the corners were ground to arc-shape to embed into the load station of the wear experimental system. To enable the effect of space (s) on the RCF life of treated samples to be studied, the samples were divided into six groups according to the unit space and each group contained three indiscriminate samples for repetitive test. Among the classification, the group without treatment called smooth sample was marked No. 1 and the groups with units space (s) ranged between 1 mm and 5 mm were marked Nos. 2–6, respectively. The stripes unit perpendicular to the rolling direction was processed by Nd:YAG laser in the same parameters. The strips were to form the geometrical non-smooth surface to imitate the rough surface of soil animals. It should be noted that the remelting region was named units in the paper. The sketch of treated sample was shown in Fig. 1. As known, the percentage of treated region increased by the decrease of unit space. The surface of samples was performed by Nd:YAG laser with wavelength of 1.06 μm and a maximum export power of 300 W, using a circular pattern Gaussian beam and set under an argon gas shield with a flow of 5 L/min. The working-bench moved at 0.5 mm/s. The parameters of laser were the pulse duration of 3 ms, frequency of 7 Hz, and the energy of a single laser point was about 2.28 J. After laser deposition, the size parameters of unit were listed in Table 2. To identify the microstructure of the unit, the standard method of metallography was performed on the transverse section after the
where ω represented the angular velocity of drive wheel, l was the eccentric distance (35 mm) that indicated the amplitude of back and forth path to be 35 mm, and φ was the deflection angle ranging 0–360° as shown in Fig. 2. Depending on the calculation of v, the positive and negative were implied the forth and back of roller, respectively. Prior to wear tests, the specimens were mechanically polished with fine emery paper sequentially, followed by ultrasonic cleaning in acetone. The weights of samples were measured by the electronic balance with a precision of 0.1 mg. Every 10 h, the samples were interrupted at regular intervals for mass loss measurement, and the above procedure was followed besides polishing to record the mass until 60 h. Additionally, the final roughness of the specimens after
Fig. 1. The schematic of the treated sample and the form of unit distribution.
Table 2 Dimensions of unit's cross section. Width (mm)
Depth (mm)
Area (mm2)
Fit equation of unit's edge
17 0.1
0.337 0.03
0.22
Y ¼1.32x2 0.33
Table 1 Nominal chemical composition (wt%), the micro-hardness (HV0.2) and tensile properties (MPa) of the base material. C
Si
Mn
P
S
Fe
Hardness
Tensile strength
Fracture strength
3–3.2
1.5–1.8
0.7–0.9
o 0.15
o 0.12
Bal.
260
135
90
Z. Chen et al. / Optics & Laser Technology 72 (2015) 15–24
17
Fig. 2. The sketch of the man-made fatigue wear system.
polishing and before wear test was recorded. At the end of test, the average mass loss of individual group was utilized to evaluate the FWR and the worn surface was observed by optical micrograph and 3-D surface profile scanning. 2.3. Finite element analysis What is more, to explain the FWR future, the numerical simulation was utilized to exhibit the distribution of stress on the surface under the experimental conditions. To simplify the calculation, the single roller was taken into account with loading of 8 N, and the roller was regarded as rigid body. For the calculation, only the pure rolling contact cases were considered, i.e., the friction coefficient between the sample and roller was assumed to be zero [12] and the linear velocity was equal to the horizontal velocity of rolling bearing. The parameters used in the numerical simulation are listed in Table 3. The refined meshes were defined in the remelting region to improve the accuracy of calculation. It should be noted that the residual stress caused by the laser processing [13,14] was not considered in the calculation.
3. Results 3.1. The microstructure and micro-hardness of bionic units After laser treatment, the microstructure of bionic unit is observed. Fig. 3a shows the cross-sectional SEM morphology of unit with low magnification and it is easy to identify three zones: transformed zone (TZ), remelting zone (RZ) and matrix which are shown in Fig. 3b, c and d with high magnification, respectively. As shown in Fig. 3b, the interface is clearly observed because of the distinct microstructure, and the RZ and TZ are combined by metallurgical bonding. Additionally, there is a little un-dissolution G in TZ and it consists of acicular martensite. Fig. 3c demonstrates the micrograph of microstructure in RZ, in which a complete dissolution of the G and re-solidification of the ledeburite are presented. As a supplement, ledeburite, the mixture of cementite and martensite, reveals the reticular structure. Fig. 3d indicates that the original phase of matrix is mainly characterized of the pearlite (P) which is surrounded by ferrite (F) and flake graphite (G). After laser irradiation, the phase compositions are detected by XRD, as presented in Fig. 4. The diffractogram shows that the new phases are formed, such as martensite, austenite (retain austenite in room temperature) and Fe7C3. As the heat is transformed in depth, the absorbed heat by material element is gradually reduced in depth leading to the absolute melting zone and partially melted zone, which are the RZ and TZ respectively mentioned above. Actually, a zone in which the phase transition takes place without remelting is named heat affect zone (HAZ), and it is categorized into TZ. In RZ, F experienced the hypo-eutectic reaction to transform to retained austenite (Ar) and martensite (M) during the remelting process. Moreover, the cementite close to rich carbon
Table 3 The parameters used in numerical simulation. Elasticity modulus (GPa) PM 130 Unit 150
Passion ratio
Density (g/cm3) Friction coefficient
0.23 0.25
7.0 7.4
0
zone is transformed to metastable carbides (Fe7C3) at the duration of resolidification. Material cannot be absolutely melted because of the insufficient heat that is proven by the retain graphite and the low cooling rate then induces the formation of coarse martensite in TZ. Another aspect of the pattern exhibits broadening width of diffraction peaks, demonstrating crystal grains in the whole unit zone very fine and compact. It can be explained as a considerable amount of high cooling rate during solidification caused by the laser processing. Due to phase transition, the micro-hardness is considered to be remarkably various upon the microstructure. The average surface micro-hardness is measured to be 680 HV0.2. The measurement of unit is depicted in Fig. 5 as a function of the distance from surface. The plot feature demonstrates that the hardness decreases from the top to the bottom of unit since the energy is gradually reduced with the distance from surface increases. According to the plot, the microhardness in RZ and TZ is markedly different in which the hardness in RZ varies from 620 to 750 HV0.2 and in TZ ranges from 320 HV0.2 to 620 HV0.2 with rapid decrease. The exhibition of different values of micro-hardness is with the occurrence of fine grain in RZ while the relatively coarse particle is observed in TZ, as shown in Fig. 3b. By calculation, the average microhardness of unit is about 560 HV0.2, which is increased by 115%, in contrast to that of the substrate with the average microhardness of 260 HV0.2. It is interesting to point that the peak hardness occurs at the shallow layer, about 0.08 mm. In terms of the fact that the existence of carbide could significantly increase the microhardness, it is reasonable to conjecture the formation of a razor-thin decarburized layer on surface. In conclusion, the existence of refined grain, arising from rapid solidification of the melt, can obviously increase the hardness of units. 3.2. Results of fatigue wear tests 3.2.1. The mass loss and the mass loss rate of samples The histogram shown in Fig. 6a is the average recorded mass losses for groups at the end of test. According to the figure, it can be clearly seen that the mass loss (ML) of untreated sample is always higher than that of the treated samples which intensively demonstrates the laser treatment to be effective on FWR improvement. Additionally, among the treated samples, the ML decreases with the decrease of unit space, attaining a minimum at s¼2 mm, while for less values of unit space (s¼1), the ML increases slightly. Therefore, it can be inferred that there must be a balance between the development of surface damages and the space of reinforced region. With respect to Fig. 6b, all the plots suggested a maximum ML rate at 20 h
18
Z. Chen et al. / Optics & Laser Technology 72 (2015) 15–24
Fig.3. The micrographs of sample cross-section: (a) The cross-sectional SEM micrograph of unit, (b) the interface of RZ and TZ (1000 ), (c) the microstructure of RZ (2000 ) and (d) the matrix (1000 ).
Fig. 5. The plot of unit's hardness in depth in loading of 0.2 kg with a holding time of 10 s. Fig. 4. The XRD diffractogram of smooth and treated samples.
and the stable wear behaviors until the end. This can be explained as the high initial roughness of the sample's surfaces significantly affects the fatigue wear behavior as the contact pressure profile [15] or the occurrence of weak region during the sample preparation, and the serious wear is removed under the steady-state wear condition as a result of material removal or plastic deformation [16]. For another aspect, the peak ML rates of treated samples are always less than that of parent material. Similarly, the peak ML rate decreases in sequence with space decrease, ranging 5–2 mm and however the peak ML rate still increases for s¼1 mm. Namely, the sample with less ML shows the better FWR because the metal removal acts as the result of fatigue failure. According to the worn surface of samples by 3-D surface profile scanning, it is obviously seen that no serious damages (spalling, large delamination and metallic flake) occur at the unit part in which only less and slight pits are detected (Fig. 7a). On the other hand, the serious defects are investigated on the original part of
treated sample (Fig. 7b) and however they are much slighter than that on untreated sample (Fig. 7c). Comparing the original part with unit part of treated sample, more and larger particles are torn fallen from the surface of original part. In conclusion, since the existence of unit, the surface damages are remarkably reduced and the main defects focus on the untreated parts for treated samples. 3.2.2. The enhanced degree of fatigue wears resistance The average Ml of samples in each group was chosen for tests under the same test conditions. Since the ML of all treated samples (MLtreated) are less than that of the smooth one (MLuntreated), the ML decrease percent (here-in after referred to as a 'MLDP') is taken to assess the enhanced degree of FWR for different groups. The calculation formula is as follows:
MLDP =
ML treated−ML untreated × 100% ML untreated
(2)
Z. Chen et al. / Optics & Laser Technology 72 (2015) 15–24
19
Fig. 6. The results of fatigue wear tests: (a) the mass losses and (b) the mass loss rates of samples vs. the test time.
Fig. 7. The worn surface profile by optical microscope: (a) unit region, (b) untreated part of treated sample and (c) untreated sample.
The MLDPs given by Eq. (2) show that the value of Nos. 2, 3, 4, 5 and 6 are 33.1%, 40.2%, 34%, 25.4% and 13.4% respectively. According to the dates, it can be seen that the MLDP of No. 3 is the biggest to imply the optimum FWR, attaining about 40.2%. What is more, the degree of FWR improvement is generally remarkable for the large treated part, while it would indicate the opposite tendency when the treated area reduces to a threshold value (s¼ 2 mm). Therefore, we deduce that there are some adverse factors to affect the FWR against the treated area. In terms of the MLDPs, it is confirmed that laser surface treatment is significantly useful for the improvement of fatigue wears resistance. Moreover, it intensively
indicates the space of units to be an important factor on the improvement of FWR. 3.2.3. Fatigue wear behavior The curves shown in Fig.8 are drawn with the stress and strain as functions of the distance in width (along AA′) to infer the distributions of strain and stress introduced by finite element analysis. It definitely validates that the maximum stress and strain are shown at both the sides of roller where the serious damages prefer to emerge. According to the worn track of insertion in Fig. 8, the dark area occurs at the roller's edges which have been ensured to be the serious damage and gradually weaken to the center of the contact part (gray
20
Z. Chen et al. / Optics & Laser Technology 72 (2015) 15–24
area) in width. The worn surface is well agreed with the simulation analysis. With respect to the roughness, the measured values of dark area and gray area are ranged 285–318 nm and 295–325 nm, respectively, which significantly suggested a decrease compared to the original surface roughness of 417 nm. What is more, since the stress at gray area is smaller than that at the dark area during the rolling contact, the gray area could be regarded as the precursor of dark area. And the roughness of dark area and gray area is considered to be almost similar that proves the roughness of worn track to be about 300 nm at the duration of steady-state wear condition. Generally, the common failures of RCF are in variations of types, including cracks, pits, spalling and delamination. Complementarily, two types of delamination are suggested in this work: the one is characterized by the part edge to be cliff and the rest edge to be inclined from the bottom to surface which results from the nearly closed crack defined as type I; another one is all the edge to be cliff resulting from the entirely closed crack defined as type II. Fig. 9 displays the worn surface observations of samples at the end of wear test. It can be seen that serious damages (cracks, spalling and large scale delamination) are observed on the surface of untreated sample (Fig. 9a). In Fig. 9b, the delamination of type I is clearly illustrated and
Fig.8. The plots of strain and stress of the surface against the distance in width (AA'). The insertion of figure is the image of the appearance of worn surface.
it looks like the flake metallics are torn from the surface. Looking at the tip region of the principal crack (indicated by the solid arrow in Fig. 9b), the partial crack removal is evident. Furthermore, the shape of type I is revealed in a fanned-out appearance which is well agreed with the points of Zakharov and Goryacheva [17] who proposed that microcracks developed at a small angle to the surface resulting in fine angled cracks developing on circumference of the defect. Additionally, no spalling or delamination is observed at the center of contact while the so-called defects appearing on the surface are almost cracks, owing to the stress on this region being weaker than that at the both sides (see Fig. 9c). Therefore, it is believable that serious defects (spalling, delamination) are resulted from the propagation and nucleation of cracks by repetitive rolling contact.
4. Discussions 4.1. Resistance of microcrack initiation by treated unit As stated, serious defects are resulted from the microcracks' behaviors, such as nucleation, propagation and closure under rolling contact. Thus, delaying the initiation of crack is considered to be an effective way to improve FWR of GCI. Since the flake graphite is regarded as the natural hole or crack, their presence can lead to the notch effect in the matrix. Previous works have pointed that fatigue failures in service invariably occur at stress concentrations [18], and therefore at the tips of graphite where concentration stress generally prefers to arise is the origin of micro-crack, as shown in Fig. 10a. Graphite is relatively softer than the matrix materials and it cyclically softened to the matrix resulting in a concentration of the plastic strain. Moreover, the graphite lamellae act as stress raiser producing self-microcracking, or interface debonding at very low stress levels and plastic deformation in the matrix [19]. Hence, graphite is a potential threat to GCI component in RCF conditions. The performance of remelting zone (Fig. 2c) is in substantial agreement with that of Benyounis [20] who proposed the graphite to be completely dissolved after the laser processing. By disappearance of graphite, the FWR of material is significantly improved. Moreover, due to the
Fig. 9. Worn surface after fatigue wear test: (a) dark area of untreated sample, (b) dark area of treated sample and (c) gray area of untreated sample.
Z. Chen et al. / Optics & Laser Technology 72 (2015) 15–24
21
Fig. 10. (a) The origin of crack initiation and (b) the route of crack propagation.
phase transformation and refinement of grain, the treated portions with high hardness exhibit excellent deformation resistance and meanwhile the matrix acts as a soft phase to cushion shear stress coming from the roller, which reduces the possibility of stress concentrations. 4.2. Resistance of microcrack propagation by treated unit As graphite plays the role of a bridge to connect the adjacent micro-cracks (Fig. 10b), the disappearance of graphite is positive to the resistance of crack propagation. The unit containing amount of fine crystal combined with the quantity of grain boundary increase is also significantly beneficial for the resistance of crack propagation, as described in Ref. [21] that the fatigue cracks' sensitivity depends on the matrix strength and the grain boundary can prevent crack propagation. Reasonably, the fatigue cracks propagate in a zig-zag manner which is difficult to cross or along the fine and reinforced crystal particle as the zig-zag route. In comparison with linear propagation, the crack much more frequent deflection of prolonged route increases the absorbed energy and will contribute to greatly lower fatigue crack growth rate. In addition, owing to the residual stress incurred by accumulated plastic deformation which drives the propagation of cracks [22], the increase of plastic deformation resistance which corresponds to enhanced yield strength of materials can markedly postpone the propagation behavior. After laser remelting, refinement of the interlamellar spacing increases the yield strength of the steel according to Hall– Petch type relationship [23] as
σs = σ0 + K y·d−1/2
(3)
where ss is the yield stress, s0 represents the lattice friction stress required to move individual dislocations, ky is a constant value which depends on the contribution of measured grain boundary to yield strength, and d is the average diameter of grains. According to the formula, the yield strength of materials increases in inverse to grains' dimension. Indeed, the tensile test shows that the tensile strength and fracture strength of treated one are 195 MPa and 145 MPa which increase by 44.4% and 61.1%, respectively, compared to those of original sample. Due to grain refinement, the unit acts as a phase to contribute to high strength of sample's surface, that is, the crack propagation is evidently hindered by unit. With respect to the hindering effect of unit during the fatigue wear test, four types of cracks are revealed for discussion and marked a, b, c and d, respectively, as presented in Fig. 11. When the crack 'a' arrives the unit, the crack cannot be driven to further propagate as inadequate force and is subsequently stopped at the edge of unit. Additionally, when the crack ‘c’ reaches the region of unit, the crack is divided into two branches as response because the refinement grain
Fig. 11. The principle of the unit hinders the propagation of crack.
in unit makes the crack difficult to through the unit. The branches however propagate in different directions in which the left one connects the crack 'b' to form the semi-closed crack in the matrix region and the right one connects the crack 'd' to afford sufficient energy to deeply propagate in the unit. Consequently, the unit makes cracks propagation complex which contributes to delay the cracks' propagation compared with the linear manner in matrix. In conclusion, the initiation and propagation of cracks from the substrate region are significantly suppressed by bionic unit, which is considered to be the reason of FWR enhancement by LSM. 4.3. Behaviors of fatigue failure According to Fig. 7, a few and slight damage is observed on unit region while the large scale spalling and delamination are detected on the matrix region of both treated and untreated samples. During the fatigue wear processing, the cracks are driven by the accumulation force to form circular crack resulting in failure. Since the coarse grain size and the flake graphite are distributed in the matrix, the cracks are relatively easy to gather to longer cracks leading to the subsequently serious damages. By the propagation of cracks, the damage is resulted from the renegade metallic from the matrix along the crack when the cracks are associated to incompletely closed crack. This type of defect is defined as type I and the depth is about 4000 nm, as shown in Fig. 12a in detail. The phenomenon has been reported that the first directed towards the depth with an inclination and a branch begins to rise towards the contact surface inclined by about 70° to the initial direction, causing material removal [18]. Differently, as the existence of refined grain size in unit, the numerous grain boundaries make the propagation of cracks harder and more frequent deflection which is prone to form entirely closed crack. The followed occurrence of defect is resulted from the completely closed crack which is called
22
Z. Chen et al. / Optics & Laser Technology 72 (2015) 15–24
Fig. 12. The different types of delamination on surface: (a) type I of delamination on untreated sample and (b) type II of delamination on unit region.
Z. Chen et al. / Optics & Laser Technology 72 (2015) 15–24
Fig. 13. The schematic of the occurrence of non-uniform deformation: (a) before and (b) after rolling contact.
delamination of type II, as Fig. 12b explicitly shows. A comparison of two types of delamination is that type II not only takes more time and energy to form the closed crack, but is also smaller and shallower (about 1500 nm in depth) in dimension than type I in RCF failure. The treated region is proportional to the enhancement degree of FWR in theory, and the experimental results however do not agree with the assumption. According to the previous research works that the RCF results from the plastic flow [24] and the maximum shear stress occurs at sub-surface [25], the fatigue micro-crack would be considered to be initial at the sub-surface. As known, the plastic deformation takes place with the exhaustion of material ductility and the degree of plastic deformation is also diverse with the difference of material ductility. Fig. 13 demonstrates the status of parent material and treated portion before (Fig. 13a) and after (Fig. 13b) rolling contact. Since the surface is modified with the reinforced unit, the assumption of non-uniform deformation between hard phase (unit) and soft phase (matrix) is reasonable in RCF condition, and the maximum non-uniform deformation does generally occur at the position where the maximum shear stress emerges beneath surface. By the occurrence of non-uniform deformation, the huge residual stress would take place at the interface where the micro-crack preferentially occurs after thousands of repetitive rolling contact. Once the crack forms, it will be driven to grow at interior of sample until a critical length and then toward the surface resulting in the surface damages of metallic debris removal (mass loss). Therefore, it is necessary to decrease the degree of the non-uniform deformation for reducing the possibility of crack initiation. As increase of unit yield stress, the degree of non-uniform deformation would be aggravated under the same magnitude of shear stress and hence relatively large fraction of soft phase is designed to reduce or even eliminate the phenomenon of non-uniform deformation. In this paper, different fractions of soft phase come true by adjusting the spaces of units to attain the balance of plastic deformation. Additionally, formation of soft zone caused by tempering effect and low compressive residual stresses in the overlapping tracks are reported to produce a long narrow crack along the edge of the overlapping tracks and cause premature failure with the adjustment of the unit track space [7,26]. The discussion can well explain the result that the ML of No. 2 is higher than that of No. 3, although No. 2 is with larger treated percentage. By combination of the discussion and result of ML, it can be concluded that when the s is more than 2 mm, the percentage of remelting area (reinforced region) is the dominant factor on the FWR improvement, while the development of damage becomes the prominent factor for so2 mm. Hence, the influence of the space is considered to be reasonable on RCF and the optimum space is 2 mm according to the wear results. What is more, Young et al. stated that during plastic deformation, the strength of the ductile phase is enhanced via plastic constraint by the surrounding reinforced hard phase [27] and the deformation of hard phase can be released due to
23
the excellent ductility of soft phase. This structure in this case that the soft substrate regions are mingled by the hard phase with relatively higher deformation resistance is similar to and accords the principle of bionic approach. Halling [28] suggested that a stick area exists in the contact area in which the particles would be broken away from surface by stick force. After LSM, because of the increased interaction of refined crystal, the stick force is suggested to be released or exhausted more or less. Thus, the fatigue life is considered to be prolonged. What is more, the traction coefficient is subsequently reduced after laser treatment no matter under the lubricated condition or the dry condition, which has been indicated by Franklin et al. [2]. According to Fletcher [25] who manifests that the maximum orthogonal shear stress increases with the increase of traction coefficient, the conclusion therefore can be stated that the way of LSM could significantly reduce the shear stress of materials correspondingly to improve the FWR. In fact, the optimum space is obtained just under the laboratorial condition. In actual application, the shear stress is far more than that in laboratory which interprets that the space is not commonly used in various applied loadings, because of the reason that gray cast iron presented a non-linear load–displacement ratio even at small strain levels, implying that Hooke's law cannot be applied to these materials [29]. The space of laser track optimization in this paper only gives effective surface morphology guidance for processing related parts and suggests that the effect factor should be taken into consideration in engineering.
5. Conclusion The following conclusions can be drawn from this investigation: After the laser processing, the treated region with an average hardness of 680 HV0.2 is created, which is increased by 162%, in contrast to that of the substrate with the average microhardness of 260 HV0.2. The remelting zone is mainly characterized of compact and refinement grains, such as retain austenite, martensite, Fe3C, etc. and there is no graphite in melt pool. Additionally, the tensile properties of treated sample are significantly enhanced such that tensile strength and fracture strength of the treated one are 195 MPa and 145 MPa which are to increase by 44.4% and 61.1%, respectively, compared to those of the original sample. For the result of Finite Element, the stress arising at the both sides of contact region is relatively higher than that at the center part. The worn surface profile agrees well with simulations that serious damages occur at both the sides of contact region while a number of cracks take place at the center of contact. Additionally, the roughness of worn track is about 300 nm at the duration of steady-state wear condition. According to the MLDP, the optimum unit space is 2 mm, which attains 40.2%. Depending on the worn surface, the defects on the matrix of treated samples are lesser than that on original sample, but it is more serious than that on unit region. However, it cannot suggest that the FWR is in direct proportion to the treated area. When the s is more than 2 mm, the percentage of remelting area (reinforced region) is the dominant factor, while the development of damage becomes the prominent factor for s o2 mm. For fatigue failures, there are two types of delamination in the present work: the one is characterized by the part edge to be cliff and the rest edge to be inclined from the bottom to surface and another one is all the edges to be cliff. Subsequently, the mechanisms of FWR improvement according to the results are pointed out that (i) for microscopic behavior, the bionic units not only delay the initiation of microcracks, but also significantly obstruct the propagation of cracks; (ii) for macroscopic behavior,
24
Z. Chen et al. / Optics & Laser Technology 72 (2015) 15–24
the hard phase resists the deformation and the soft phase releases the deformation.
Acknowledgment This article was supported by Project 985 – High Performance Materials of Jilin University, the Project 985 – Bionic Engineering Science and Technology Innovation and National Natural Science Foundation of China (No. 51275200). Moreover, the authors would like to express their thanks to the conscientious editor Marc Sentis and the anonymous reviewers on the manuscript revision.
References [1] Datsyshyn O, Panasyuk V. Pitting of the rolling bodies contact surface. Wear 2001;25:1347–55. [2] Franklin F, Weeda GJ, Kapoor A, Hiensch E. Rolling contact fatigue and wear behaviour of the infrastar two-material rail. Wear 2005;258:1048–54. [3] Šamec B, Potrč I, Šraml M. Low cycle fatigue of nodular cast iron used for railway brake discs. Eng Fail Anal 2011;18:1424–34. [4] Polonsky I, Chang T, Keer L, Sproul W. A study of rolling-contact fatigue of bearing steel coated with physical vapor deposition TiN films: coating response to cyclic contact stress and physical mechanisms underlying coating effect on the fatigue life. Wear 1998;215:191–204. [5] Colombo DA, Echeverría MD, Laino S, Dommarco RC, Massone JM. Rolling contact fatigue resistance of PVD CrN and TiN coated austempered ductile iron. Wear 2013;308:35–45. [6] Grum J, Šturm R. Comparison of measured and calculated thickness of martensite and ledeburite shells around graphite nodules in the hardened layer of nodular iron after laser surface remelting. Appl Surf Sci 2002;187:116–23. [7] Iino Y, Shimoda K. Effect of overlap pass tempering on hardness and fatigue behaviour in laser heat treatment of carbon steel. J Mater Sci Lett 1987;6:1193–4. [8] Tong X, Zhou H, Zhang ZH, Sun N, Shan HY, Ren LQ. Effects of surface shape on thermal fatigue resistance of biomimetic treated cast iron. Mater Sci Eng A 2007;467:97–103. [9] Zhou H, Guo QC, Lin PY, Zhang W, Zhang XL, Ren LQ. Influence of H13 steel unit on wear behavior of vermicular cast iron. Appl Surf Sci 2008;255:3394–9. [10] Wang CW, Zhou H, Zhang ZH, Jing ZJ, Cong DL, Meng C, Ren LQ. Tensile property of low carbon steel with gridding units. Appl Surf Sci 2013;273:128– 34.
[11] Chen ZK, Lu SC, Song XB, Zhang HF, Yang WS, Zhou H. Effects of bionic units on the fatigue wear of gray cast iron surface with different shapes and distributions. J Opt Laser Technol 2015;66:166–74. [12] Zhang XC, Xu BS, Tu ST, Xuan FZ, Wang HD, Wu YX. Fatigue resistance and failure mechanisms of plasma-sprayed CrC–NiCr cermet coatings in rolling contact. Int J Fatigue 2009;31:906–15. [13] Grum J, Šturm R. Influence of laser surface melt-hardening conditions on residual stresses in thin plates. Surf Coat Technol 1998;100–101:455–8. [14] Grum J, Šturm R. Influence of laser remelting process parameters on residual stresses in nodular cast iron 500-7. Mater Manuf Process 2000;15:815–27. [15] Donzella G, Faccoli M, Ghidini A, Mazzù A, Roberti R. The competitive role of wear and RCF in a rail steel. Eng Fract Mech 2005;72:287–308. [16] García A, Cadenas M, Fernández MR, Noriega A. Tribological effects of the geometrical properties of plasma spray coatings partially melted by laser. Wear 2013;305:1–7. [17] Zakharov SM, Goryacheva IG. Rolling contact fatigue defects in freight car wheels. Wear 2005;58:1142–7. [18] Taylor D, Hughes M, Allen D. Notch fatigue behaviour in cast irons explained using a fracture mechanics approach. Int J Fatigue 1996;18:439–45. [19] Bertolino G, Perez-Ipiña JE. Geometrical effects on lamellar graycast iron fracture toughness. J Mater Process Technol 2006;179:202–6. [20] Benyounis KY, Fakronb OMA, Abboudb JH, Olabic AG, Hashmic MJS. Surface melting of nodular cast iron by Nd–YAG laser and TIG. J Mater Process Technol 2005;170:127–32. [21] Ochi Y, Masaki K, Matsumura T, Sekino T. Effect of shot-peening treatment on high cycle fatigue property of ductile cast iron. Int J Fatigue 2001;23:441–8. [22] Ringsberg J, Loo-Morrey M, Josefson B, Kapoor A, Beynon JH. Prediction of fatigue crack initiation for rolling contact fatigue. Int J Fatigue 2000;22:205–15. [23] Hugo RC, Kung H, Weertman JR, Mitra R, Knapp JA, Follstaedt DM. In situ TEM tensile testing of DC magnetron sputtered and pulsed laser deposited Ni thin films. Acta Mater 2003;51:1937–43. [24] Garnham JE, Davis CL. The role of deformed rail microstructure on rolling contact fatigue initiation. Wear 2008;265:1363–72. [25] Fletcher DI. The influence of lubrication on the fatigue of pearlitic rail steel [Ph.D. thesis]. UK: Department of Mechanical Engineering, University of Sheffield; 1999. [26] Van Brussel B, Hegge H, De Hosson JTM, Delhez R, De Keijser TH, Van der Pers N. Development of residual stress and surface cracks in laser treated low carbon steel. Scr Metall Mater 1991;25:779–84. [27] Young CH, Bhadeshia H. Strength of mixtures of bainite and martensite. Mater Sci Technol 1994;10:209–14. [28] Halling J. Principles of tribology. New York: The Macmillan; 1975. p. 174–201. [29] Haenny L, Zambelli G. The stiffness and modulus of elasticity of graycast irons. J Mater Sci Lett 1983;2:239–42.