Surface & Coatings Technology 201 (2007) 5273 – 5277 www.elsevier.com/locate/surfcoat
Fatigue life and mechanical behaviors of bearing steel by nitrogen plasma immersion ion implantation Hongxi Liu ⁎, Baoyin Tang, Langping Wang, Xiaofeng Wang, Bo Jiang State Key Laboratory of Advanced Welding Production Technology, Harbin Institute of Technology, Harbin 150001, China Available online 7 September 2006
Abstract Rolling contact fatigue (RCF) performance and mechanical characteristics of nitrogen plasma immersion ion implantation (PIII) on AISI52100 bearing steel surface has been investigated using a conventional three ball-on-rod rig. Testing investigations include optical microscopy (OM), friction and wear behavior, rolling contact fatigue life, and nano-indentation measurements. Results indicate that the rolling elements failed at the surface or near-surface layer. Moreover, an appreciable amount of surface wear was observed on the surface of the rolling elements after RCF tests. The maximum microhardness of treated samples is nearly twice as that of substrate. The friction coefficient of the treated samples is decreased from 0.90 to 0.15. The L10 life of the treated specimens increase by 99.6% and L50 life enhance by 236.3% at a Hertzian stress level of 5.1 GPa and 90% confidence level, respectively. OM morphology results indicate that the surface roughness, processing parameters, and inner defects play an important role in surface and near-surface initiated RCF. This improvement of fatigue and mechanical performance is attributed to form a combination of nitrides phase structure and residual compressive stress during the nitrogen-PIII process. © 2006 Elsevier B.V. All rights reserved. PACS: 61.72.Ww; 62.20.Mk; 81.40.Pq; 68.35.Gy; 68.47.De Keywords: Plasma immersion ion implantation; Rolling contact fatigue; Friction and wear; Mechanical performance; Bearing steel
1. Introduction Nitrogen ion implantation has most extensively been used to enhance the hardness, wear, fatigue and corrosion properties of metal and alloy materials surface [1–4]. The implantation causes changes in surface composition and chemical bond structure, leading to the formation of new metastable compounds and alloy layers. Bombardment may also form lattice imperfections due to radiation damage and cause structural changes within the near surface region of the solid [5,6]. Along with other surface hardening processes, such as plasma nitriding, PVD (physical vapor deposition), and CVD (chemical vapor deposition), etc., ion implantation has unique advantages. In addition to, an increase of compressive residual stress can be observed in most samples after ion implantation. Compressive residual stress on the surface is beneficial for preventing fatigue failure [7,8].
Since its invention in 1987 by Conrad and coworkers at the University of Wisconsin, PIII has been known to be an innovative technique for material surface modifications [9–11]. Although the PIII technique shows some similarities to ion-assisted coating and ion beam mixing, PIII has the following advantages: (1) non-lineof-sight process; (2) easy handling; (3) multiple targets; (4) stable modified layer (with no debonding and macroscopic defects); (5) low cost treatment for materials of complicated shape etc. Therefore, PIII technique has become more and more popular even in industrial applications. At present, PIII has been developed based on two approaches, one being to circumvent the line-of-sight restrictions of conventional ion implantation, the other one to combine ion implantation with thermally enhanced diffusion and thus bridge the gap between low temperature ion implantation and plasma diffusion treatment at elevated temperatures . Table 1 Chemical composition of the AISI52100 steel
⁎ Corresponding author. Tel.: +86 45186418728; fax: +86 45186416186. E-mail address: [email protected]
(H. Liu). 0257-8972/$ - see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.surfcoat.2006.07.183
H. Liu et al. / Surface & Coatings Technology 201 (2007) 5273–5277
Table 2 The detail parameters of PIII Sample no.
Bias voltage (kV) Implanted time (h)
In recent years, the number of literatures regarding PIII has increased [13–15]. However, most reports treated implantation for the surface mechanical and chemical properties of work pieces. Papers rarely study the effect of fatigue life of industrial components after being treated by the PIII technique. In this paper, we report the results of PIII of nitrogen ion into AISI52100 stainless steel, which is a material used for bearing. The aim was to study the effect of nitrogen-PIII on the fatigue and mechanical behaviors of the AISI52100 bearing steel. 2. Experimental procedures Samples of AISI52100 stainless steel in quenched and tempered state (HRC 61–65) was used in this study. The principal chemical composition of this steel is listed in Table 1. Two groups of samples were prepared, the flat coupons were machined to a diameter of 15 mm and thickness of 3 mm; the fatigue samples, AISI52100 cylindrical rods, with a length of 200 mm and a diameter of 11 mm. Before deposition, all coupons were polished to a surface roughness Ra about 0.04 μm, followed by an ultrasonic clean in acetone, then kept in an electric-dryer to prevent the
surface from pollution again, and then put into the vacuum chamber. Nitrogen ion implantation was carried out in our multi-purpose plasma immersion ion implantation facility . The vacuum chamber was evacuated to a base vacuum of 5.0 × 10− 3 Pa, and then Ar + sputtering was introduced into the chamber to remove undesirable oxide and contamination layers. Nitrogen plasma was produced using a radio-frequency glow discharge plasma source. The flow rate is 50 sccm, pulse frequency 80 Hz, and working gas pressure 6.0 × 10− 1 Pa. Implanted pulse width is 60 μs, main arc current 120 A. Experimental details of the samples are shown in Table 2. After PIII, mechanical properties and fatigue life were evaluated and compared with the base material properties. The friction coefficient of all specimens was measured using a pin-ondisk wear tester, the radius of the sliding track was 3 mm, and the upper ball is made of silicon carbide (SiC) ceramic, 4 mm in diameter at ambient environment. The specimen was rotated at a constant sliding speed at 300 rpm, and the contact load was 0.3 N. No lubricant was used in the wear test. The microhardness and elastic modulus were evaluated using the nano-indentation system UMIS-2000 (CSIRO). A trigonal diamond indenter (Berkovich-type) with total inclined angle of 142.3° was used for the measurements. The tip radius of the indenter was approximately 50 nm, the load and depth resolution of the machine were 1 μN and 0.03 nm, respectively. Rolling contact fatigue (RCF) tests were performed on a three ball-on-rod RCF testing machine. The contact was lubricated by
Fig. 1. Weibull plots of the fatigue lives of RCF test rods treated by nitrogen-PIII at various substrate bias voltages, as well as the untreated rods. Data obtained using standard loading balls and 90% confidence level.
H. Liu et al. / Surface & Coatings Technology 201 (2007) 5273–5277 Table 3 Summary of the RCF test results Sample no. N0 N1 N2 N3
5.46 5.84 10.9 8.48
8.98 14.8 30.2 23.1
9.15 26.2 37.2 28.1
2.9 1.8 1.3 1.5
mineral oil, pure rolling conditions (not shown on the schematic graph in this paper). The loaded 11.1 mm diameter balls were made of AISI52100 steel. The normal load pressing each of the three balls against the rod specimen was set at 9 kg, corresponding to maximum Hertz contact pressure of 5.1 GPa. In such a load range, bulk plastic deformation did not occur in the specimens according to contact stress analysis using the classical Hertzian theory. The specimen was rotated at 2800 rpm. At this speed, the test rod experienced 3.95 × 105 contact loading cycles in 1 h of testing . The RCF testing machine ran continuously until a macroscopic fatigue spall was formed on one of the rolling elements. The formation of fatigue spall will cause an increase in the vibration level. A highly sensitive sensor is mounted on the lever arm detecting the surface deterioration as a result of surface fatigue. The test stops when the vibration level has exceeded the preset critical value. Before and after the test, the rod specimen, loading balls and retainer were thoroughly cleaned with acetone to wash away any debris that might have been produced during the test [18,19]. Finally, the characteristics of the fatigue specimens damage image were observed in an optical microscope. 3. Results and discussion The Weibull plots of untreated substrate rod and nitrogenimplanted samples are shown in Fig. 1. The figures depicted the percentage of specimens failed (ordinate) vs. the number of stress cycles to failure (abscissa). Results of the Weibull analysis are summarized in Table 3, which show the L10, L50 lives for the two groups of specimens at various pulse bias voltages, other parameters such as mean life and Weibull slope are also shown. Comparing the results for the treated and untreated samples reveal
that the RCF life of nitrogen-treated AISI52100 steel rods' surface are better than that of the untreated rod. The fatigue life improvement for both L10 and L50 in nitrogen-PIII condition, a 6.96%, 99.6%, 55.3% difference in the L10 life and 64.8%, 236.3%, 157.2% difference in the L50 life corresponding to −15 kV, −25 kV, −35 kV, respectively. For nitrogen-implanted specimens, the number of the tests performed for each pulse bias voltage was insufficient to estimate the Weibull distribution parameters with a high degree of confidence, taking into account the considerable scatter of fatigue lives, which is quite typical of RCF testing. Nonetheless, the fatigue life data obtained can be used for a qualitative analysis of nitrogen-implanted effect on the RCF life. From the results described above, its main reason is that the improvement in the RCF life of the nitrogen-treated specimens were attributed to an implanted wear resistance due, in part, to the increase in surface hardness and residual compressive stress shown to occur when AISI52100 material is implanted with nitrogen. A further contributing factor to the improvement of the wear resistance may arise due to a reduction in the friction coefficient. The microhardness and elastic modulus of treated samples vs. pulse bias voltage and implantation time are exhibited in Fig. 2. The improvement in microhardness after nitrogen-PIII treatment is evident, the higher the substrate bias voltage at a given condition, the higher the hardness measured. From Fig. 2(A), with the increase of pulse bias voltage, the microhardness of the treated specimen reaches the maximum value of 14.4 GPa at −35 kV; the maximum elastic modulus is 237 GPa. Compared with the substrate (7.4 GPa and 215 GPa), the maximum microhardness and elastic modulus value is increased by 94.6% and 10.2%, respectively. From Fig. 2(B), the maximum microhardness value of 12.5 GPa at 4 h, elastic modulus 0f 234 GPa at 3 h, increased by 68.9% and 8.84%, respectively. This indicates that implanting nitrogen into AISI52100 steel formed a hard surface layer. A higher nitrogen implantation energy (i.e., −35 kV) can also produce a higher microhardness than that of lower energy. The main reason is that a structural hardening, due to the solid solution, creating compressive residual stresses, and nitrides, locking dislocations due to plastic deformation, enough nitrogen atoms to form Fe-nitrides and Cr-nitrides in the case of the PIII process. According to literature [20,21], a thin layer of iron nitrides of Fe2N, Fe3N, Fe4N and chromium nitrides of CrN or Cr2N phases
Fig. 2. Microhardness and elastic modulus of all samples under different conditions: (A) different pulse bias voltages; (B) different implantation times.
H. Liu et al. / Surface & Coatings Technology 201 (2007) 5273–5277
Fig. 3. Friction coefficient of treated and untreated samples as a function of sliding cycles at various conditions: (A) different substrate bias voltages; (B) different implantation times.
can be formed at the surface region of the coupons after the nitrogen-PIII treatment. All these phases are beneficial for improving the microhardness of the treated sample. Friction curves of the nitrogen-PIII treatment samples in different processing parameters are depicted in Fig. 3. Variations in the friction coefficient with the number of sliding cycles for various samples are shown in Fig. 3(A) as implantation time was 4 h and Fig. 3(B) as pulse bias voltage was −25 kV. By comparison with the substrate, the friction coefficient of all nitrogenPIII treated samples decreased obviously. For sample N1, the friction coefficient rapidly increased to a range 0.80–0.85 and was near to 0.90 which represented the coefficient of self-friction of AISI52100 steel. For samples N2 and N3, at the beginning of the
test, the value is lower due to the existence of a small amount of adsorbate on the surface. As this layer is broken during the test, the friction coefficients begin to increase gradually after about 450 cycles. And it can be found that samples N2 and N3 hold nearly the same stable stage (about 450 cycles), but the final friction coefficient of sample N3 is less than that of sample N2. In Fig. 3(B), the sliding cycles in which the friction coefficient holds stable stage are different, sample N2 is the highest, followed by N5, then N4, and finally N6 decreases sequentially. With increasing rotating cycles, the friction coefficient of all samples eventually reaches the same value (about 0.75). This is because the implanted layer is damaged and the underlying substrate is exposed. But the friction coefficient of all treated samples remains
Fig. 4. Optical micrographs of the rolling tracks of treated and untreated specimens at various implantation times (magnification ×50): (A) N0: 16.7 h test; (B) N4: 40.6 h test; (C) N2: 111 h test; (D) N6: 54.7 h test.
H. Liu et al. / Surface & Coatings Technology 201 (2007) 5273–5277
smaller than that of the substrate throughout the test. The main reason is the nitrogen diffusion into the modification layer and certain nitride phases formed during the PIII process, these hardened phases can decrease the surface friction coefficient and prolong the wear life. If the cycles in which the friction coefficient begins a steep rise is defined as the wear life, then wear life of the −35 kV treated sample is the longest in Fig. 3(A) and 4 h treated coupon is the longest in Fig. 3(B). This is because high bias voltage and long treatment time corresponding to high ion energy, deeper nitrogen diffusion layers and more nitride phases can be produced, therefore, the friction and wear behavior of these samples are better than that of others. But for N5, the friction coefficient less than that of sample N4, the main reason is that more treatment time can produce more rough surfaces, so sample N6 has a relatively bad wear effect. The results of the friction test are also in agreement with the above microhardness and elastic modulus data. Optical microscopy (Olympus BX60) is employed to examine the PIII-treated rods surfaces. Fig. 4 shows a series of surface rolling contact fatigue damage contours with the variation of implantation time, after wear particles and lubricant deposits are removed. Fig. 4(A), (B), (C) and (D) illustrates the fatigue spall images after 16.7, 40.6, 111, and 54.7 h of fatigue testing, respectively. The profile of the fatigue spall looks like an ellipse. Fatigue crack propagation originates from the inner inclusions or surface defects in the modification region and nearly always perpendicular to the rolling direction, then grow both parallel to and perpendicular to the rolling direction. The geometry characteristics of secondary surface cracks are also clearly found (denoted here by the arrow). The fatigue damage modes, may be surface wear that was associated with asperity contact or inclusions in the presence of microslip/sliding within the contact region. Nitrogen implanting improves the fatigue behavior of bearing steel by restricting fatigue crack initiation. So implanting may be effective, on one hand, by inhibiting dislocation movement at the surface (due to the induced compressive stress during fatigue); on the other hand, by stopping the development of slip bands on the surface (which nucleate cracks). The mechanism of fatigue crack nucleation, growth and propagation will be further studied in another study. 4. Conclusions The mechanical and fatigue properties of the AISI52100 bearing steel, can be substantially improved by nitrogen plasma immersion ion implantation technique. The RCF life of all treated samples is prolonged, and the maximum value of L10 life and L50
life increased by 99.6% and 236.3%, respectively. Surface wear is the main fatigue damage mode. The maximum microhardness and the effective elastic modulus value of the treated specimens increased by 94.6% and 10.2%, respectively. The microhardness increase is mainly attributed to the constraint of plastic deformation of the substrate by hard nitrides modification layer and high compressive residual stress. The nitrogen-PIII also decreased the friction coefficient from 0.90 to about 0.15, which are depending on the pulse bias voltage and implantation time. Therefore, nitrogen-PIII is a potential method for improving the mechanical performance and RCF live of AISI52100 bearing steel. Acknowledgements The authors would like to acknowledge the financial support from the National Defense Preparation Research Foundation for this research work. We would like to thank the Harbin Bearing Group Company for supplying the AISI52100 bearing materials for the rolling contact fatigue tests. References  Hara Yoshihito, Y. Tetsuji, A. Kingo, et al., Surf. Coat. Technol. 156 (2002) 166.  S. Mukherjee, P.M. Raole, P.I. John, Surf. Coat. Technol. 157 (2002) 111.  X.B. Tian, C.B. Wei, S.Q. Yang, et al., Surf. Coat. Technol. 198 (2005) 454.  M. Ueda, M.M. Silva, Surf. Coat. Technol. 169/170 (2003) 408.  A.K. Goel, N.D. Sharma, R.K. Mohindra, et al., Indian J. Phys. 67 (A1) (1993) 75.  H.K. Sanghera, J.L. Sullivan, S.O. Saied, Appl. Surf. Sci. 141 (1999) 57.  Manyuan Li, Emile J. Knystautas, Surf. Coat. Technol. 138 (2001) 220.  Z.W. Deng, R. Souda, Diamond Relat. Mater. 11 (2002) 1676.  Koumei Baba, Ruriko Hatada, Mater. Chem. Phys. 54 (1998) 135.  J.R. Conrad, J.L. Radtke, R.A. Dodd, et al., J. Appl. Phys. 62 (1987) 4591.  Koumei Baba, Ruriko Hatada, Surf. Coat. Technol. 158/159 (2002) 741.  W. Wanga, J.H. Booske, C. Baum, et al., Surf. Coat. Technol. 111 (1999) 97.  A. Mitsuo, S. Uchida, T. Aizawa, Surf. Coat. Technol. 186 (2004) 196.  Z.M. Zeng, R.Y. Fu, X.B. Tian, et al., Surf. Coat. Technol. 186 (2004) 260.  S. Mukherjee, M.J. Maitz, M.T. Pham, et al., Surf. Coat. Technol. 196 (2005) 312.  S.Y. Wang, P.K. Chu, B.Y. Tang, et al., Nucl. Instrum. Methods Phys. Res., B Beam Interact. Mater. Atoms 127/128 (1997) 1000.  I.A. Polonsky, L.M. Keer, T.P. Chang, et al., Wear 215 (1998) 191.  D.G. Lover, in: J.J.C. Hoo (Ed.), A Ball-Rod Rolling Contact Fatigue Tester, ASTM STP, 1982, p. 107.  Y.H. Chen, A.P. Igor, Y.W. Chung, et al., Surf. Coat. Technol. 154 (2002) 152.  A. Mitsuo, S. Uchid, T. Aizaw, Surf. Coat. Technol. 186 (2004) 196.  W.L. Li, Y. Sun, W.D. Fei, Appl. Surf. Sci. 187 (2002) 192.