Contact fracture mechanism of electroplated Ni–P coating upon stainless steel substrate

Materials Science & Engineering A 563 (2013) 184–192

Contents lists available at SciVerse ScienceDirect

Materials Science & Engineering A journal homepage: www.elsevier.com/locate/msea

Contact fracture mechanism of electroplated Ni–P coating upon stainless steel substrate Akio Yonezu a,n, Michihiro Niwa a, Jiping Ye b, Xi Chen c,d,1 a

Department of Precision Mechanics, Chuo University, 1-13-27 Kasuga, Bunkyo, Tokyo 112-8551, Japan Cutting-edge Technology Department, Nissan ARC, LTD 1 Natsushima-cho, Yokosuka, Kanagawa 237-0061, Japan c Department of Earth and Environmental Engineering, Columbia University, 500W 120th Street, New York, NY 10027, USA d International Center for Applied Mechanics, SV Lab, School of Aerospace, Xi’an Jiaotong University, Xi’an 710049, China b

a r t i c l e i n f o

a b s t r a c t

Article history: Received 17 September 2012 Received in revised form 7 November 2012 Accepted 15 November 2012 Available online 23 November 2012

The contact fracture property and mechanism of electroplated Ni–P coating on stainless steel substrate were investigated using ball indentation testing, through a comprehensive experimental and numerical approach. First, the elastoplastic properties of both coating and substrate were evaluated using microindentation tests. Next, ball indentation test with large contact force was performed, such that the brittle coating on ductile substrate suffers from cracks, including ring crack (propagates circumferentially) and radial cracks (propagates radially), owing to the coating bending effect. The fracture nucleation process was investigated using the acoustic emission technique (AET). In addition, finite element method (FEM) with cohesive zone model (CZM) was carried out to compute stress field and simulate crack initiation around the impression during the test. By using the comprehensive experimental/computational framework, the nucleation process (mechanism) of such a complicate crack system was clarified. The present technique and fracture mechanism may be applicable to the analysis of structural integrity of other brittle coatings. & 2012 Elsevier B.V. All rights reserved.

Keywords: Electroplated Ni–P coating Contact fracture Ball indentation

1. Introduction Hard thin films or surface coatings on ductile metallic substrates are often used for contact and slide wear protection. Therefore, the characteristics and mechanism of their contact fracture are critical for ensuring their mechanical performances. Many hard coatings are deposited using the electroplate technique, which may achieve massive production with low cost and large area/thick deposition (even when the substrate geometry is complicated) [1–5]. Among the electroplated hard coatings, Ni–P material possesses high hardness, high strength and other superior mechanical properties, providing excellent performance (such as wear and corrosion resistance) for metallic ductile substrate/components [2,6–9]. By adding P (phosphorus) and conducting post-heat treatment, it is known that chemical combination of NiP3 phase precipitates and grows, leading to the enhanced hardness [9,10]. Characterizing the contact fracture properties of Ni–P coatings with different microstructures is the most important issue regarding structural integrity and application. Conventional evaluation n

Corresponding author. Tel.: þ81 3 3817 1829; fax: þ 81 3 3817 1820. E-mail addresses: [email protected] (A. Yonezu), [email protected] (X. Chen). 1 Tel.: þ1 212 854 3787; fax: þ1 212 854 7081. 0921-5093/$ - see front matter & 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.msea.2012.11.054

methods, such as sliding contact/scratch test and pin on disk, are commonly employed to probe the wear resistance, including coating strength and interface toughness [2], such that deposition technique may be further improved. However, these mechanical tests are difficult to elucidate the detailed fracture process (e.g., how and when does the coating fracture occur?). A quantitative mechanistic study is needed so as to guide the improvement of wear resistance. Indentation method is a convenient way to probe mechanical properties, and it has several advantages over the conventional tests. This technique does not require extensive sample preparation (the coating remains adhered to the substrate), and detection of the cracking process is possible using acoustic emission techniques [11,12]. However, indentation/contact loading sometimes shows complicate fracture morphology, such as radial crack, ring crack and lateral crack [13,14]. These differences are dependent on coating thickness, geometry of indenter, elastoplastic properties of both coating and substrate [15–17], which dictates various types of stress field and the maximum value that are responsible for coating fracture. In addition, the crack nucleation and propagation may interact with each other; for example, the nucleation of a crack affects the stress distribution, which may lead to the formation and propagation of other cracks. One of the critical challenges lies in an understanding of the process how several different types of cracks occurs. It may require reliable method to monitor the crack

A. Yonezu et al. / Materials Science & Engineering A 563 (2013) 184–192

2. Materials

propagation and fracture processes in-situ during indentation test, as well as to effectively analyze the stress field upon indentation loading and during crack propagation. In this study, ball indentation test was carried out to simulate contact fracture of electroplated Ni–P coating on stainless steel substrate. Acoustic emission technique was utilized to monitor the timing of coating cracks. Stress field upon indentation was computed by finite element method (FEM), where the cohesive element was used to simulate crack nucleation. The comprehensive experimental/numerical approach helps to clarify stress criterion of complicate coating crack system.

Table 1 shows the electrolysis condition of Ni–P alloy electroplated onto SUS304 in the plating bath. After electroplating, heat treatment of 350 oC was performed for 1 h in vacuum. According to Refs. [9,10], post-heat treatment is crucial for the mechanical properties of the coating. Initially, the element of P (phosphorus) is a solution in the matrix of Ni, and then NiP3 (having high hardness) gradually precipitates during post-heat treatment; when the post-heat treatment temperature is between 300 and 350 1C, the maximum hardness (HV¼800–1000) is achieved [9,10] for wear protection, whereas the fracture toughness exhibits the lowest value [9]. Therefore, the investigation of crack morphology and cracking resistance due to contact loading is a critical issue for the Ni–P electroplated coating (especially when the heat treatment temperature is about 350 1C). Fig. 1(a) shows the cross section near the interface between coating and substrate. The coating thickness is about 180 mm observed by scanning electron microscope. The coating is well bonded with the steel substrate; indeed, there was no

Table 1 Condition of Ni–P electrolyte plating. Composition

g/l

Plating

NiSO4 6 H2O NiCl2 6 H2O H3PO3 H3PO4

150 150 2.5–16 20–130

Temperature pH Current density Anode

185

50 75 1C1 0.5–1.5 150 mA/cm2 Ni-plate

Ni-P layer

Ni-P

SUS304

SUS304 Ni-P layer

Ni-P

SUS304

SUS304

100μm 5μm

5μm

Indentation force F, mN

150 Experiment FEM

100 Ni-P

SUS304

50

0 0

0.5

1

1.5

Interface

200

300 0

5

Hardness H,GPa

10

Surface

Ni-P

0

100

Interface

200 SUS304

Ni-P

100

Depth from the surface, μm

Surface

0

SUS304

Depth from the surface, μm

Penetration depthh, μm

300 0

100

200

300

Young’s modulus E, GPa

Fig. 1. Experimental procedure and results of micro indentation test to evaluate elastoplastic properties of Ni–P coating and stainless steel substrate. (a) Micrograph of cross-section of Ni–P/SUS304, (b) Micro indentation method for cross-section of Ni–P coating and SUS304, and their impressions. (c) Their indentation curves combined with computational ones using finite element method. Changes in hardness (d) and Young’s modulus (e) in cross-section of Ni–P/SUS304 material.

186

A. Yonezu et al. / Materials Science & Engineering A 563 (2013) 184–192

3. Experimental method

delamination/exfoliation of coating when the indentation test was carried out, as discussed later. Furthermore, it is reported that pin-on-disk test showed no delamination of the coating [9]. Thus, adhesion strength is sufficient in practice, and the present study focuses on only the cracking of the coating. To evaluate mechanical properties (elastoplastic properties) of both Ni–P coating and steel substrate, micro-indentation tests were performed using commercial testing machine (dynamic ultra-micro-hardness tester: DUH-201W, Shimadzu Co.). Fig. 1(b) shows the experimental method. To examine the cross-section, the specimen was potted in a resin compound, and then polished in order to make the specimen surface smooth. As shown in Fig. 1(b), micro-indentation test with Berkovich indenter was performed with the maximum force of 100 mN. Several tests against the cross section were conducted to investigate the changes in mechanical properties across the interface of coating/substrate. Here, the location of indentation test was carefully chosen (i.e., space interval, surface roughness etc), to consider the location which is completely apart from the plastic deformation area developed by other impression measurements. During each test, the relationship between indentation force F and penetration depth h (F–h curve) was continuously recorded. The representative F–h curves were plotted by solid lines in Fig. 1(c). Based on the F–h curves, the micro-hardness (the maximum force Fmax divided by the contact area at Fmax) was estimated in Fig. 1(d), and the elastic properties (Young’s modulus) were estimated by Oliver–Pharr method [18,19] in Fig. 1(e). Here, the Poisson’s ratio is referred to be 0.37 obtained by the surface acoustic wave (SAW) technique [10]. In Fig. 1(d), the average hardness of Ni–P coating is about 7.5 GPa (the range is from 6 to 8 GPa), corresponding to Vickers hardness (HV¼ 800–900) measured by a previous study [10]. The hardness of steel substrate is about 3.5 GPa, which is quite lower than that of the coating, suggesting that the present Ni–P coating plays a big role as contact resistance. The Young’s modulus of coating/substrate shows similar values, i.e., Ni–P coating is 217 GPa, and SUS304 is 175 GPa. The estimated value (of the coating) agreed well with the results of SAW technique [10]. The plastic properties (stress–strain relationship) of coating and substrate were estimated by using the reverse analysis (single indentation method based on dimensionless analysis [20,21]). Here, the plastic properties were assumed with powerlaw constitutive equation, involving yield stress sY and work hardening exponent n. The estimated results were shown in Table 2. To verify the estimation, finite element simulation (employing these estimated properties) of the micro-indentation test was carried out and the resulting indentation curve (F–h) was shown in Fig. 1(c). The simulated results (as indicated by grey circle) agreed well with experimental data for both coating/ substrate. This suggested that the estimated elastoplastic properties were robust, and can be employed for stress analysis and crack nucleation subjected deep ball indentation (see Section 5.1). Noted that, according to Ref. [9], there might be internal residual stress of Ni–P coating. By using X-ray diffraction, the stress was measured to be almost zero for the present coating (with heat treatment temperature of 350 1C) [9]. Thus, we consider the internal stress-free coating.

Fig. 2 shows the experimental set up. Ball indentation tests were performed using an electro-hydraulic testing machine (EHF-1, Shimadzu Co.) equipped with a ball indenter and two eddy current sensors (EX-201 Keyence Co., displacement resolution: 0.4 mm). The diameter of indenter ball d is 10 mm. The software for controlling indentation test can continuously measure the relationship between indentation force (F) and indentation depth (h) at high resolution under various testing conditions. The indentation force F gradually increases with the rate of dF/dt ¼1 N/s up to the maximum indentation force Fmax (¼ 1000 or 2000 N), and sustains the constant value in 50 s, and gradually decreases with the same dF/dt until the force of zero. During the test, acoustic emission (AE) signals were monitored by using four small AE sensors mounted on the side surfaces of a specimen, as shown in Fig. 2. The AE system consists of resonant-type sensors (Type PICO, PAC), 40-dB preamplifiers (9913, NF Circuit Block), A/D converter (Gage Applied Inc.) and a personal computer. Here, AE signals with the higher amplitude than threshold value we set could be recorded by the hit-count method. The recorded digital data (waveform in time domain) was converted to FFT spectrum and Wavelet transformation to characterize the detected AE waveform. The detailed procedure is described later.

4. Experimental results Fig. 3 shows the history of indentation force for two different tests, with Fmax ¼1000 and 2000 N. The cumulative AE count is

4-AE sensors Specimen

Top view

2-Displacement sensors Load cell Ball indenter φ 10mm

4– Small AE sensors

Specimen Side view

Fig. 2. Experimental setup for ball indentation test with acoustic emission system.

Table 2 Mechanical properties of electroplated Ni–P coating and SUS304 substrate. Material

Young’s modulus E GPa

Poisson’s ratio n

Hardness H GPa

Yield stress sy GPa

Work hardening exponent n

Ni–P coating SUS304 substrate

217 175

0.37* 0.3

7.51 3.49

2.40 0.22

0.05 0.47

n

Morikawa et al. The Iron and Steel Institute of Japan (ISIJ), vol.82, pp.935–940, 1996

A. Yonezu et al. / Materials Science & Engineering A 563 (2013) 184–192

Force-time

2000

AE count (Fmax=2kN)

1500 100 1000

500

Cumulative AE count

200

2500

Indentation force F, N

187

AE count (Fmax=1kN)

0

0 0

1000

2000

3000

4000

Test time, s Fig. 3. Changes in indentation force and cumulative AE count with test time.

200 μm also plotted in the right axis. Due to the electric noise from the equipment and its surrounding, the threshold value for the detection of AE needs to be changed for each test2. Thus, for each test the cumulative AE count is slightly different, but the tendency is similar. Fig. 4 show the micrographs of specimen surface after the tests. In the test of Fmax ¼1000 N (Fig. 4(a)), despite some friction/ wear tracks on the surface, but there is no clear coating crack. On the other hand, the test of Fmax ¼2000 N (Fig. 4(b)) shows complicated coating crack morphology: one type is circumferential crack, namely ‘‘ring crack’’, and the other type initiates from the ring crack and propagate radially, namely ‘‘radial crack’’. Again, there is no delamination (coating spalling) thanks to the strong adhesive strength of the coating/substrate system. The present study therefore focuses on the formation mechanisms of ring crack and radial crack, which may provide useful insights for the mechanical/material design of coating/substrate system. Fig. 5 shows representative AE waveforms during the tests. To characterize the waveforms, we conducted FFT (fast Fourier transformation) and WT (wavelet transformation). The results of FFT spectrum and relative intensity of WT were also shown in the figures. The detected AEs were roughly classified into two types, by the wave amplitude and frequency: Fig. 5(a) shows small AE with relatively low frequency (less than 250 kHz), and Fig. 5(b) shows large AE having higher frequency (up to 600 kHz). For the test with Fmax ¼1000 N, all AEs can be regarded as the type of Fig. 5(a), while the test with Fmax ¼2000 N contains both AE types as shown in Fig. 5(a) and (b). Therefore, AE of Fig. 5(a) may possibly come from friction between indenter and specimen surface, since there was no crack in the Fmax ¼1000 N test (see Fig. 4(a)). Consequently, among AEs in the Fmax ¼2000 N test, the small AE were neglected, and the larger AEs were plotted as triangles on the F–h curve in Fig. 6. It is found that the first AE was detected at about F¼1200 N and several AEs were subsequently monitored up to about F ¼1500 N. Furthermore, the unloading process was found to start AE generation from F ¼1600 N during unloading. Therefore, it is expected that the coating cracks occur during both loading and unloading. To explore the fracture mode of coating cracking (i.e., crack opening vector), the first portion of the detected AE waveforms were investigated. As a representative case, AEs at F¼1200 N (in loading) and at F¼ 1000 N (in unloading) were shown in Fig. 7(a) and (b). They reveal that all AE signals (detected all

2 Due to experimental difficulty, the threshold value needed to be changed. Thus, the comparison of the number of cumulative AE count is not focused in this study.

Fig. 4. Micrographs of impressions tested with Fmax ¼1000 N (a) and 2000 N (b).

channels) shows same polarity, indicating that the fracture mode is mode-I type [21], i.e., the normal tensile stress is responsible for the present cracking. Based on these experimental evidences, the stress field is investigated by finite element method in the next section, so as to further clarify the contact cracking behavior.

5. Discussion 5.1. Crack initiation By taking advantage of symmetry, an axisymmetric model is established for FEM analysis with the commercial code MARC and MENTAT [22]. Fig. 8 shows FEM model for ball indentation onto the coating and substrate. The material parameters are taken from Table 2, for both coating and substrate. A rigid ball indenter with radius d ¼10 mm is employed as a close analog to that used in experiment. To verify both the present FEM model and measured material property, the computed indentation curve is plotted as symbols (grey circle) in Fig. 6, showing reasonable agreement with the experimental curve.

188

A. Yonezu et al. / Materials Science & Engineering A 563 (2013) 184–192

Output, V

Power spectrum 10-13

III 0.1

I

0

-0.1 0

20

40

60

20 II

10

0 0

80

500

Timeμs

1000

1500

Frequency kHz

Output, V

0.1

Power spectrum 10-13

Relative intensity

I

0

-0.1 0

20

40

60

80

III

30 II 20

10

0 0

500

1000

1500

Frequency kHz

Timeμs

Relative intensity

Fig. 5. Representative AE waveforms (I), FFT spectrum (II) and relative intensity of wavelet transform (III). (a) shows small AE with relatively low frequency and (b) shows large AE having higher frequency.

Experiment

0.05

Output V

FEM AE timing

2000 1500

Ch.2

Ch.3

Ch.4

-0.05 0.05

1000 500 0

Ch.1 0

Output V

Indentation force F, N

2500

0

10

20

30

40

50

0

-0.05 20

25

Penetration depth h, μm

Fig. 9 represents a contour map of the indentation stress field computed by FEM simulation when the first AE was detected (as discussed in Fig. 6). Fig. 9(a) shows the map for the radial stress component (srr), which is responsible for ring crack, and Fig. 9(b) is the circumferential component (syy) for radial crack. For srr, a large tensile stress (up to about 1.8 GPa) occurs outside the contact region, where the ring crack may be produced if such a tensile stress is sufficiently high. Indeed, such a prominent tensile stress is contributed by the large local bending curvature of the film, assisted by the extensive plastic deformation of the substrate [12,15,16]. When the ring crack appears inside the prominent tensile stress zone, Mode-I type AE should be produced (to be consistent with Fig.7). On the other hand, syy (in Fig. 9(b)) is relatively small (less than about 1.0 GPa which is

Time μs

30

Time μs

Output V

0.05

Ch.1

Ch.2

Ch.3

Ch.4

0

-0.05 0.05

Output V

Fig. 6. Indentation curve (indentation force F and penetration depth h: F–h curve) obtained by the test with maximum force of 2000 N. Triangle on the curve indicates the timing of AE generation, having large amplitude with higher frequency (which is incorporated in Fig. 5(b)). In addition, grey circle indicates the simulated indentation curve with finite element method (using the model of Fig. 8).

25

30

0

-0.05 20

25

Time μs

30

25

30

Time μs

Fig. 7. AE waveforms detected by four sensors, at indentation force F of 1200 N during loading (a) and at F¼ 1000 N during unloading.

A. Yonezu et al. / Materials Science & Engineering A 563 (2013) 184–192

θ

189

r

σrr, σθθ GPa z

2.2

interface

1.8 1.4 1.0

σrr

0.6 0.2 -0.2 -0.6 -1.0

interface

-1.4 -1.8 σθθ

180μm Fig. 9. Counter map of stress distribution when the first AE was detected (see Fig. 6). (a) normal stress along radial direction, srr, and (b) normal stress along circumferential direction, syy.

Ni-P

Crack zone

SUS304 2000N

4

1500N

Critical stress

Fig. 8. Two dimensional (axisymmetric) FEM model to compute the stress field around impression in Ni–P coating on SUS304 substrate; entire model (a) and zoom-in view (b).

σrr, GPa

180μm

2

1200N

0 F=500N

-2 quite lower than the srr component). Therefore, the first detected AE should be from ring crack, suggesting that the ring crack initiated first. The surface distribution of srr as a function of the indentation force is given in Fig. 10. As expected, the maximum tensile stress increases and shifts outwards with increase in indentation force (deeper penetration), which is associated with the increased coating bending curvature outside the contact zone. When the indentation force reaches the critical value (F ¼1200 N), the location of the maximum stress (r ¼390 mm) roughly coincides with the radius of the ring crack (as indicated by dashed vertical lines3) observed in experiment (see Fig. 4(b)). Thus, we can obtain the critical fracture strength of the coating, sC ¼ 1.8 GPa. Note that, this value of sC is seemed to be the intrinsic strength of the coating (since the present Ni–P coating does not have large internal stress [9], Section 2). From the above investigations, it is revealed that when a ball indenter makes deep contact with the surface of brittle coating, the steel substrate undergoes extensive plastic deformation, which bends the coating and leads to large tensile stress outside

3 The ring crack is not a precise circle, and the corresponding radius has some fluctuation; the range of ring crack radius (400–410 mm) is given by the two vertical dashed lines in Fig. 10.

700N

-4 0

200

400

600

800

1000

Distance from the center of impression r, μm Fig. 10. Radial stress sr distribution as a function of distance r, with the increases in indentation force. The thick line corresponds to the critical indentation force of 1200 N, when first AE is detected (see Fig. 6).

the contact region. When the stress (srr) reaches the critical value of coating strength, ring crack initiates. Thus, the next question is how the radial crack initiates after the ring crack formation. 5.2. Subsequent crack initiation To investigate how radial crack forms, the existence of ring crack must be incorporated with stress analysis during indentation. Thus, we employed the cohesive zone model (CZM) in the FEM to compute the stress field in conjunction with ring crack formation. The CZM is applicable to both ductile and brittle materials [23–27]. For instance, Mode-I fracture of a brittle coating on a silicon substrate [25], hydrogen embrittlement in the steel [28,29] and the interfacial fracture of an IC-interconnect

190

A. Yonezu et al. / Materials Science & Engineering A 563 (2013) 184–192

(e.g., copper and low-K materials) [24] have been successfully investigated using the CZM. The CZM essentially models the fracture process zone in a plane ahead of the crack tip. The zone is assumed to be subjected to cohesive traction. The model usually describes the gradual degradation of the adhesion between two regions along the crack propagation plane. The mechanical response of the cohesive zone obeys a traction– separation law that yields the relationship between the separation distance v of the two material faces at an interface and the traction stress s acting between them. Although numerous traction–separation laws for the cohesive zone element have

Damage parameter: D 1.0 0.9 0.8 0.7 0.6 0.5 0.4

θ

r

z

been proposed (for example see Ref. [23] for an overview), this study employed an exponential law (called the Smith–Ferrante type [24]) due to its simplicity. This exponential law requires two independent materials parameters, i.e., the maximum stress smax and the crack growth resistance KC. Since the smax roughly corresponds to the critical stress (fracture strength), smax is set to be 1.8 GPa from Fig. 10. However, the other parameter KC is unknown. Thus, several values (0.5, 1.0 and 5.5 MPa m1/2) were employed to simulate ring crack formation. To compute the stress field involving ring crack formation, we introduced CZM element into FEM model (same with Fig. 8). Here, the CZM elements are implemented at the location where the ring crack forms (r ¼400 mm in Fig. 4(b) and Fig. 10). Fig. 11 shows the contour map of damage parameter D around the impression (when D becomes one, crack completely forms [22,24]). When Kc ¼ 5.5 MPa m1/2, as shown in Fig. 11, the ring crack (due to srr component) propagates from the surface to the interface. In fact, actual crack was found to propagate up to the interface from the cross sectional observation. Note that the choice of Kc does not affect the subsequent stress field after crack propagation. 5.3. Mechanism of cracking system

0.3 0.2 0.1 0

interface 50 μm

Fig. 11. Contour map of CZM damage for ring crack nucleation, when indentation force is about 1200 N.

Fig. 12 shows the snapshot of the normal stress syy distribution during the loading process. Fig. 12(a) and (b) shows the stress field at F ¼1500 N, whereas Fig. 12(c) and (d) shows the result at the maximum indentation force (2000 N). For comparison, the model with no cohesive zone element was also computed, as shown in Fig. 12(a) and (c). This model simulates the stress field due to indenter contact, and does not induce any crack formation. Thus, syy (in Fig. 12(a) and (c)) does not show large tensile stress, which is the same trend with Fig. 9(b). On the other hand, the model with CZM element in Fig. 12(b) and (d) exhibits different

No CZM element θ

z

r

CZM element Crack position

interface

interface

σθθ GPa 2.2 1.8 1.4

50μm

50μm

1.0 0.6 0.2 -0.2 -0.6

interface interface

-1.0 -1.4 -1.8 50μm

50μm

Fig. 12. Contour map of stress syy distribution around the impression during indentation loading: homogenous model with no CZM element ((a) and (c)) and model with CZM element ((b) and (d)). (a) and (b) shows the map at F ¼1500 N (F/Fmax ¼ 0.75), while (c) and (d) is at the maximum force (Fmax ¼2000 N).

A. Yonezu et al. / Materials Science & Engineering A 563 (2013) 184–192

No CZM element

θ

191

CZM element

r

Crack position

z

interface

interface σθθ GPa 2.2 1.8 1.4

50μm

50μm

1.0 0.6 0.2 -0.2

interface

-0.6

interface

-1.0 -1.4 -1.8

50μm

50μm

Fig. 13. Contour map of stress syy distribution around the impression during indentation unloading: homogenous model with no CZM element ((a) and (c)) and model with CZM element ((b) and (d)). (a) and (b) shows the map at F¼ 1500 N (F/Fmax ¼ 0.75), while (c) and (d) is at the full unloading (F ¼0 N).

3

Interface Maximum stress position Maximum σθθ, GPa

2.8

Indentation force F 3

1500N 2000N

2.5

loading

1500N σθθ, GPa

2

0N

unloading

1.5

2.6 2.4 2.2 2 1.8 1.6 0

1

500

1000

1500

2000

2500

Indentation force F, N 0.5 0

Fig. 15. Changes in maximum stress syy (in Fig. 14) as a function of indentation force during unloading.

0

50

100

150

200

Depth from the surface, μm Fig. 14. Stress syy distribution along the ring crack path (the depth from coating surface to interface). Four lines with various symbols indicate the results of 1500 and 2000 N (loading process) and 1500 and 0 N (unloading process).

syy stress field, owing to the ring crack formation. In particular, possible area for large tensile syy develops near the interface. Fig. 13 shows a snapshot of syy distribution during unloading process, in a similar fashion with Fig. 12. In Fig. 13(a) and (c), without the CZM element model there is no large syy distribution, whereas with the CZM element, the significant larger syy

develops at the right side of crack path near interface, and its magnitude increases during the unloading process and reaches the maximum value upon full unloading. Therefore, it is found that ring crack formation significantly changes the subsequent stress field during the indentation, in particular the large syy upon unloading, and the radial crack is seemed to initiate at the right side of crack path. Fig. 14 shows the stress syy distribution along the crack path with respect to the distance from coating surface to interface. The four curves indicate the results of 1500 and 2000 N (loading) and 1500 and 0 N (unloading). While the overall syy magnitude increases during the unloading process, the position of the

192

A. Yonezu et al. / Materials Science & Engineering A 563 (2013) 184–192

maximum syy does not change (at about 175 mm below the surface). Finally, the change in the maximum syy value is investigated as a function of the indentation force (during unloading process) in Fig. 15. As expected, the maximum syy increases with decreasing indentation force. Referring to Fig. 6, the AE occurrence was seen at about F ¼1600 N under unloading. In Fig. 15, this force (F ¼1600 N) corresponds to syy ¼1.82 GPa, which is reasonably agreement with the critical stress srr for ring crack in Fig. 10. Therefore, it is found that radial crack is produced by syy, which develops near the interface and ring crack path, suggesting that ring crack formation (during loading) is crucial for subsequent radial crack nucleation during unloading.

6. Conclusion This study investigated the contact fracture property of electroplated Ni–P coating on stainless steel substrate, which is important for its application as contact/sliding member for wear resistance. Ball indentation test with large contact/indentation force produces two types of cracks in the coating, namely the ring crack and radial crack. To elucidate the fracture process, acoustic emission technique (AET) was employed to identify the timing of crack initiation during the test. In addition, finite element method (FEM) was carried out to compute the stress field around the impression during the test. The cohesive zone modeling (CZM) was embedded with FEM to simulate the crack interaction. It is found that the ring crack first initiates during loading process, due to tensile radial stress (owing to the coating bending effect). Subsequently, radial crack nucleates from ring crack path (i.e., near ring crack tip at interface) due to the large circumferential stress developed upon unloading. By using the comprehensive experimental/computational indentation framework (combined with AE and FEM), the mechanism of the complicate coating crack system is clarified, and the stress criterion for each cracking system is quantified. Based on these findings, further systematic study may suggest how to control or prevent the coating cracks. This will become useful guidance for material design in coating industry. The comprehensive experimental/computational framework is also applicable to other coating/substrate systems.

Acknowledgment The work of A.Y. is supported in part of Grant-in-Aid for Young Scientist of (B) (No. 22760077) of the Ministry of Education,

Culture, Sports, Science and Technology, Japan, and Research Grant for Science and Technology of SUZUKI Foundation. The work of X.C. is supported by the National Natural Science Foundation of China (11172231), the World Class University program through the National Research Foundation of Korea (R32-2008-000-20042-0), and the National Science Foundation (CMMI-0643726).

References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29]

K.-H. Hou, M.-C. Jeng, M.-D. Ger, Wear 262 (2007) 833–844. H. Tian, N. Saka, E. Rabinowicz, Wear 142 (1991) 51–85. Z. Ping, J. Yuan, Y. He, X. Li, Acta Metall. Sinica (Eng. Lett.) 22 (2009) 225–232. Y.D. He, H.F. Fu, X.G. Li, W. Gao, Sci. Mater. 58 (2008) 504–507. P. Peeters, G. Hoorn, T. Daenen, A. Kurowski, G. Staikov, Electrochim. Acta 47 (2001) 161–169. S.-H. Kim, Mater. Lett. 61 (2007) 3589–3592. W. Sha, J.-S. Pan, J. Alloys Compd. 182 (1992) L1–L3. W. Sha, P. J.-S., J. Less-Common Met. 166 (1990) L11–L13. T. Yashiki, T. Nakayama, J. Kato, Iron Steel Inst. Jpn. 81 (1995) 48–53. Y. Morikawa, H. Cho, T. Nakayama, M. Takemoto, Iron Steel Inst. Jpn. 82 (1996). A. Yonezu, B. Xu, X. Chen, Mater. Sci. Eng. 507 (2009) 226–235. A. Yonezu, B. Xu, X. Chen, Thin Solid Films 518 (2010) 2082–2089. R.F. Cook, G.M. Pharr, J. Am. Ceram. Soc. 73 (1990) 787–817. X. Chen, J.W. Hutchinson, A.G. Evans, J. Am. Ceram. Soc. 88 (2005) 1233–1238. B.R. Lawn, Y. Deng, P. Miranda, P. Pajares, H. Chai, D.K. Kim, J. Mater. Res. 17 (2002) 3019–3035. H. Chai, B.R. Lawn, J. Mater. Res. 19 (2004) 1752–1761. A. Yonezu, H. Cho, T. Ogawa, M. Takemoto, Sci. Tech. Adv. Mater. 7 (2006) 97–103. W.C. Oliver, G.M. Pharr, J. Mater. Res. 7 (1992) 1564–1583. W.C. Oliver, G.M. Pharr, J. Mater. Res. 19 (2004) 3–20. N. Ogasawara, N. Chiba, X. Chen, Scr. Mater. 54 (2006) 65–70. A. Yonezu, X. Chen, Diamond Relat. Mater. 19 (2010) 40–49. Marc, Marc 2010, Theory and User’s Manual A 2010. N. Chandra, H. Li, C. Shet, H. Ghonem, Int. J. Solids Struct. 39 (2002) 2827–2855. B.A.E. Hal, R.H.J. Peerlings, M.G.D. Geers, O. Sluis, Microelectron. Reliab. 47 (2007) 1251–1261. Z. Xia, W.A. Curtin, B.W. Sheldon, Acta Mater. 52 (2004) 3507–3517. V. Tvergaard, J.W. Hutchinson, J. Mech. Phys. Solids 40 (1992) 1377–1397. G.I. Barenblatt, Adv. Appl. Math. 7 (1962) 56–129. V. Olden, C. Thaulow, R. Johnsen, E. Østby, T. Berstad, Eng. Fract. Mech. 75 (2008) 2333–2351. A. Yonezu, T. Hara, T. Kondo, H. Hirakata, K. Minoshima, Mater. Sci. Eng. A 531 (2012) 147–154.