Surface & Coatings Technology 186 (2004) 333 – 338 www.elsevier.com/locate/surfcoat
A repair process for fatigue damage using plasma nitriding Akgu¨n Alsaran *, Irfan Kaymaz, Ayhan Cßelik, Fatih Yetim, Mehmet Karakan Department of Mechanical Engineering, Ataturk University, 25240, Erzurum, Turkey Received 3 July 2003; accepted in revised form 23 December 2003 Available online 18 March 2004
Abstract The effect of plasma nitriding on the fatigue life of AISI 4140 steel having a certain fatigue loading has been investigated. Untreated and nitrided specimens were subjected to fatigue loading up to a certain level, and then nitrided for 0.5 h and subjected to fatigue testing again. It was observed that the process applied to untreated specimens improves the fatigue life of the specimens that were plasma nitrided for 0.5 h due to the subsurface crack nucleation. However, for already nitrided specimens, a significant improvement in the fatigue life has not been observed by applying the proposed process because the compound layer on the surface prevents the diffusion during the second nitriding. D 2004 Elsevier B.V. All rights reserved. Keywords: Plasma nitriding; Fatigue damage; Repair
1. Introduction Plasma nitriding is a wide-spread surface hardening process utilized to improve the fatigue strength, wear and corrosion resistance of both traditional and new materials. This process involves diffusional addition of nitrogen into the surface of materials. As a result, a compound layer and diffusion layer are formed on the surface of materials. The mechanical, tribological and corrosion properties are related to the structure formed on the surface [1– 5]. Fatigue cracks usually initiate from the materials surface. For materials hardened by plasma nitriding, the crack tends to initiate in subsurface rather than surface in high cycle fatigue. This is due to compressive residual stress formed in the diffusion layer, resulting in better resistance to cyclic slip. Thus, it is easier for cracks to initiate from internal discontinuous such as a non-metallic inclusion [6– 8]. The effect of plasma nitriding on fatigue behavior of various steels has been studied by several researches. In previous researches, it has been shown that the fatigue limit increases with increasing case depth and surface hardness, but the thickness of the compound layer does not influence the fatigue limit [9 – 11]. In addition, a few researchers
investigated the relation between the process parameters of nitriding and fatigue strength [12,13]. All the works have been interested in the increase of fatigue strength of plasma nitrided material. However, to our knowledge, it has not been investigated how plasma nitriding effects on the fatigue life of materials that have already subjected to fatigue up to certain cycles. On the other hand, the formation and growth of micro flows due to certain loading cycles can be delayed by means of the surface treatment processes, thus the fatigue damage can be repaired. Therefore, in this study, first untreated AISI 4140 steel has been subjected to fatigue up to a certain level of fatigue cycle. Then, the specimen has been plasma nitrided, and its fatigue life was examined in order to observe the improvement in the fatigue life of the material concerned. This process is denoted as Process 1 throughout this paper. In order to observe the same effect on the fatigue life of a nitrided material, the same procedure have been repeated for already nitrided AISI 4140 steel, and this process is denoted as Process 2 in this paper. In addition, the structural and mechanical properties were investigated by XRD, SEM, microhardness tester and rotating bending fatigue machine.
2. Experimental details * Corresponding author. Tel.: +90-442-231-37-51; fax: +90-442-23609-57. E-mail address:
[email protected] (A. Alsaran). 0257-8972/$ - see front matter D 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.surfcoat.2003.12.017
AISI 4140 steel has been used in the study and its chemical composition is tabulated in Table 1. The speci-
334
A. Alsaran et al. / Surface & Coatings Technology 186 (2004) 333–338
Table 1 Chemical composition of AISI 4140 low alloy steel (%) Element
C
Mn
Si
Cr
Ni
Mo
V
S
Cu
P
Wt.%
0.36
0.80
0.005
0.014
0.30
0.85
0.075
0.07
0.143
0.034
mens were normalized at 850 jC for 30 min, and then cooled in air. For the plasma nitriding process, the specimens were placed into the plasma nitriding chamber after cleaning with alcohol, which was evacuated to 2.5 Pa. The equipment used in the experiment was designed by Alsaran and Celik [14]. Prior to the plasma nitriding, the specimens were subjected to cleaning by hydrogen sputtering for 15 min under a voltage of 500 V and a pressure of 5 102 Pa to remove surface contaminates. The plasma nitriding process of AISI 4140 steel was performed under process parameters including time of 0.5 and 1 h, a temperature of 500 jC and a gas mixture of 30% N2 + 70% H2. The reason of choosing the process time of 0.5 and 1 h is to compare at least one level of stresses for S – N curve of both untreated and nitrided specimen. In order to define the formation of phases in the compound layer, a Rigaku XRD diffractometer was used. The X-ray analysis was operated at 30 kV, 30 mA with CuKa radiation. After the plasma nitriding, the thin compound layer formed on the surface was removed by polishing prior to metallographic examination and hardness testing. The nitrided case depth and the hardness distribution were measured by using a Buhler 1600-4980T instrument at a constant load of 50 g and a loading time of 15 s. The surface hardness at 25 Am depth was chosen for comparison so that any possible effects from a compound layer would be negligible. The case depth is defined as the depth at which the hardness is 10% HV above the core hardness [15]. The fatigue strength was determined using a rotating bending fatigue machine. The geometry of the specimens used for fatigue testing is shown in Fig. 1. Rotating bending fatigue tests were performed at 5000 rev./min in laboratory air atmosphere and carried out until the complete failure of specimens. The machine is provided with a digital counter that shows the number of load cycles endured by the test specimen. The fatigue test machine stopped automatically as soon as the specimen failure occurred. Twenty-six specimens were used to determine the S – N curve for plasma
Fig. 1. Rotating bending fatigue test specimen, dimensions in millimeters.
nitrided and untreated specimens; i.e. 15 specimens for the fatigue life, (three specimens at each of five levels of stress amplitude) and 11 specimens for the fatigue limit region. The staircase method was employed to determine fatigue limit. In each test, the number of cycles to fatigue failure was noted on semi-log (S, LogN) graphs. The fracture surface was examined by SEM. Since the first aim of the study is to find out the change in the fatigue life of the sample that have been nitrided after having fatigue loading up to certain level, first untreated samples have been subjected to fatigue loading up to 30 and 60% of its fracture cycles. Then, they have been nitrided for 0.5 h. After this process, they have been subjected to further fatigue loading to determine their fatigue life (Process 1). As the second aim of this study, the same procedure has been repeated for already nitrided AISI 4140 steels for 0.5 h (0.5 + 0.5 h) (Process 2), i.e. 0.5 h pre-nitrided and 0.5 h post-nitrided.
3. Result and discussion Fig. 2 shows XRD results of the nitrided AISI 4140 steel for process time of 0.5 h and 0.5 + 0.5 h. As seen in the figure, polyphase (cV and e) iron nitride was produced on the surface of AISI 4140. With increasing treatment time, the intensity of e and cV nitride phases increases but the intensity of a-Fe phase decreases. Fig. 3 depicts the SEM micrograph of nitrided AISI 4140 steel after etching in 2% nital solution. The continuous compound layer was produced on surface. The thicknesses of compound layer for 0.5 h and 0.5 + 0.5 h are 4.5 – 5.5 Am and 6 – 7 Am, respectively. 3.1. Results and discussion of Process 1 The fatigue fracture cycles according to the stress levels for Process 1 are given in Table 2, for which an example is given for 710 MPa as follows: Firstly, the S – N curve for the untreated specimen was determined, secondly the specimens were subjected to fatigue loading up to 30% (25 000 cycles) and 60% (50 000 cycles) of the fatigue fracture cycle of the untreated specimen (83 000 cycles) which is the average value of the fatigue fracture cycle of the untreated specimen at 710 MPa. Finally, the specimens were nitrided for 0.5 h and subjected to the fatigue loading again. The values of fracture cycles of the specimen subjected to fatigue loading up to 30% are given in Table 2, which is indicated as Process C. A significant improvement in fatigue life has been observed when the Processes A and C are compared.
A. Alsaran et al. / Surface & Coatings Technology 186 (2004) 333–338
335
Fig. 2. XRD results of plasma nitrided specimens for process time of 0.5 and 0.5 + 0.5 h.
The improvement obtained after applying Process C is even higher than the Process B that designates the fatigue fracture cycles of 0.5 h nitrided specimen. The reason can be explained by the fact that the crack which initiated at the surface was repaired by the nitriding process leading to the crack initiation at the subsurface. In other words, the formation of the nitride precipitates can hinder dislocation motion, and, therefore the slip bands penetrate through the subsurface. When the fatigue lives of Process B and Process C are compared, the fatigue life of Process C is slightly better than that of Process B. This situation can be explained by the behavior of the internal stresses occurred in front of the cracks in the specimens. At the beginning, the residual stresses in the specimen decrease with the partial fatigue. By
releasing the load, after the load is released, and then the specimen is nitrided, the residual stresses are recharged. Therefore a new residual stress situation occurs, which leads to a slight increase of the fatigue life of Process C. For the specimen subjected to 60% (50 000 cycles) of the fatigue fracture cycle (D), the fatigue limit is better than that of the untreated specimen for the same reason explained above. However, its value is lower than that of 0.5 h pre-nitrided specimen (B), which can be explained that the crack was in this case in the period of the growth by the non-occurrence of the fish-eye failure. The specimens having fish-eyes on the fracture surface are indicated in Tables 2 and 3. Also, the S– N curves obtained from rotating bending fatigue tests are given in Fig. 4. While the
Fig. 3. SEM micrograph of plasma nitrided specimen for process time of 0.5 h (a) and 0.5 + 0.5 h (b).
336
A. Alsaran et al. / Surface & Coatings Technology 186 (2004) 333–338
Table 2 The results of fatigue life for Process 1
Table 3 The results of fatigue life for Process 2
Stress levels
Processes
Stress levels
Processes
MPa
A
B
C
D
MPa
B
E
F
G
550
1 698 000 1 457 000 1 358 000
– – –
– – –
– – –
710
1 572 000* 1 310 000* 1 312 000*
– – –
2 150 000 2 328 000* 2 000 000*
5 106 V 5 106 V 5 106 V
590
857 000 598 000 772 000
– – –
– – –
– – –
750
627 000* 579 000* 455 000*
1 695 000* 1 759 000* 1 852 000*
1 390 000* 1 400 000* 1 150 000*
1 100 000* 5 106 V 5 106 V
630
272 000 538 000 427 000
– – –
– – –
5 106 V 5 106 V 5 106 V
790
389 000* 208 000* 285 000*
757 000* 785 000* 846 000*
728 000* 657 000* 576 000*
357 000* 398 000* 367 000*
670
140 000 237 000 176 000
– – –
5 106 V 5 106 V 5 106 V
1 101 000 5 106 V 5 106 V
830
230 000* 220 000* 178 000
385 000* 348 000* 328 000*
274 000 287 000* 205 000
178 000 223 000 155 000
710
73 000 65 000 113 000
1 572 000* 1 310 000* 1 312 000*
1 758 000* 1 650 000* 5 106 V
470 000 657 000 525 000
870
98 000* 86 000 73 000
203 000* 186 000* 197 000*
119 000* 95 000* 85 000
70 000 74 000 62 000
Process A: The fatigue fracture cycle of untreated specimen. Process B: The fatigue fracture cycle of nitrided specimen for 0.5 h. Process C: The fatigue fracture cycle of the specimen that was nitrided for 0.5 h after it was subjected to fatigue loading up to 30% of the fatigue life of the untreated specimen. Process D: The fatigue fracture cycle of the specimen that was nitrided for 0.5 h after it was subjected to fatigue loading up to 60% of the fatigue life of the untreated specimen. * The fish eye was observed in the fracture section of fatigue specimen.
fatigue strength of untreated specimen was 520 MPa, the fatigue strengths of the pre-nitrided specimens for 0.5 h was 690 MPa. 3.2. Results and discussion of Process 2 The fatigue fracture cycles at different stress levels for Process 2 are given in Table 3. The fatigue life of the specimen that was post-nitrided for 0.5 h after it was subjected to fatigue loading up to 30% of the fatigue life of the pre-nitrided specimen for 0.5 h (F) decreased according to the fatigue life of the specimen that were plasma nitrided for 1 h (E). However, an improvement in the fatigue fracture cycle was observed when compared to the specimen pre-nitrided for 0.5 h (B). For the specimen that was postnitrided for 0.5 h after it was subjected to the fatigue loading up to 60% of the fatigue life of the pre-nitrided specimen for 0.5 h (G), Process 2 has a negative effect on the fatigue fracture cycle at high stress levels while it improves the fatigue fracture cycles at low stress levels since the crack finds enough time to initiate at the subsurface. In addition, the S –N curves of Process 2 are given in Fig. 5. The fatigue strength of the plasma nitrided specimens for 1 h was 735 MPa. Fig. 6 shows the crack initiation and final fracture region of the fatigue specimens. When the fatigue fracture
910
– – –
78 000 67 000 92 000*
– – –
– – –
*The fish eye was observed in the fracture section of fatigue specimen. Process B: The fatigue fracture cycle of nitrided specimen for 0.5 h. Process E: The fatigue fracture cycle of nitrided specimen for 1 h. Process F: The fatigue fracture cycle of the specimen that was post-nitrided for 0.5 h after it was subjected to fatigue loading up to 30% of the fatigue life of the pre-nitrided specimen for 0.5 h. Process G: The fatigue fracture cycle of the specimen that was post-nitrided for 0.5 h after it was subjected to fatigue loading up to 60% of the fatigue life of the pre-nitrided specimen for 0.5 h.
surface is examined, the fish eye crack formation is observed after plasma nitriding (Fig. 6a). The fish eye crack has a circular shape and nuclei with non-metalic inclusion in its origin (Fig. 6b). The fish eye was not observed on the fatigue fracture surface at high stress levels since there was not enough fatigue loading cycles to form the fish eye (Fig. 6c).
Fig. 4. The S – N curves of Process 1.
A. Alsaran et al. / Surface & Coatings Technology 186 (2004) 333–338
337
Fig. 5. The S – N curves of Process 2.
Fig. 7. Hardness profiles of plasma nitrided AISI 4140 steel for 0.5, 0.5 + 0.5 and 1 h.
The hardness profiles of nitrided specimen are given in Fig. 7. The maximum hardness and case depth were obtained for a process time of 1 h. The surface hardness and case depth for the process time of 0.5 + 0.5 h is between 0.5 and 1 h. The reason of low hardness and case depth for nitriding process at 0.5 + 0.5 h is that the compound layer formed during first nitriding hinders N diffusion toward the core due to ceramic behavior. This
incident is the reason of decreasing fatigue life for prenitriding process time of 1 h.
4. Conclusions After the plasma nitriding process, the compound layer including e and cV iron nitrides formed on the surface of
Fig. 6. SEM micrographs of fatigue fracture surface of specimens; (a) Crack nucleation at the subsurface for Process E exposed to 790 MPa stress, resulting in 846 000 cycles of fatigue life, (b) a non-metallic inclusion in the nitrided case (c) crack nucleation at the surface for Process D exposed to 710 MPa stress, resulting in 470 000 cycles of fatigue life.
338
A. Alsaran et al. / Surface & Coatings Technology 186 (2004) 333–338
the materials. With the increasing process time, the intensity of iron nitrides increases. The surface hardness and the thickness of the diffusion layer after the two stages of the nitriding process (0.5 + 0.5 h) were lower than that of the nitriding process for 1 h. On the other hand, the fatigue life observed after the nitriding process applied to the specimens that are subjected to the fatigue loading up to 30 and 60% fatigue life of the untreated specimen increased. However, the fatigue life of the specimens treated by the Process 2 resulted in a lower fatigue life than that of the specimen nitrided for 1 h. As a conclusion, the fatigue life of machine parts that had not been subjected to any surface treatment may be increased if they are nitrided after a certain working period. It should be noted that the proposed process must be used with surface treatment methods that do not make a significant change in the dimension of the specimen.
References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15]
B. Edenhofer, Heat Treat. Met. 2 (1974) 59. E. Metin, O.T. Inal, J. Mater. Sci. 22 (1987) 2783. M.B. Karamis, Wear 147 (1991) 385. A.D. Rocha, T. Strohaecker, V. Tomala, T. Hirsch, Surf. Coat. Technol. 115 (1) (1999) 24. A. Alsaran, Mater. Chart 49 (2) (2002) 171. B.K. Jones, J.W. Martin, Met. Technol. (1978) 217. J.M. Cowling, Fatigue Cracking in the Nitrided Steel, Parthenon Press, 1996, p. 5. K. Hussain, A. Tauqir, A. Haq, A.Q. Khan, Int. J. Fatigue 21 (1999) 163. A. Cßelik, S. Karadeniz, Surf. Coat. Technol. 72 (1995) 169. K. Genel, M. Demirkol, M. Capa, Mater. Sci. Eng. A 279 (1 – 2) (2000) 207. A. Alsaran, M. Karakan, A. Cßelik, Mater. Chart 48 (4) (2002) 323. Y. Sun, T. Bell, Mater. Sci. Eng. A 140 (1991) 419. A. Alsaran, A. Cßelik, C. Cßelik, Surf. Coat. Technol. 160 (2002) 219. A. Alsaran, A. Celik, Mater Chart 47 (3 – 4) (2001) 207. O.T. Inal, C.V. Robino, Thin Solid Films 95 (1982) 195.
