CC BY
110
Surface pattern analysis of 35CrMo4 steel under high cycle fatigue test
S. Uvarov1, E. Mikhailov1, O. Plekhov1, and T. Palin-Luc2
1 Institute of Continuous Media Mechanics RAS, Perm, 614013, Russia 2 ENSAM LAMEFIP (EA 2727), Talence, F-33405, France
The present work is devoted to the investigation of high cycle fatigue (more than 105 cycles). Smooth specimens made of 35CrMo4 quenched and tempered steel were loaded at 56 Hz in fully reversed plane bending. The pattern investigation was made with a high-resolution 3D optical profilometer after the tests in specimen areas loaded at different stress amplitudes (stress gradient). Persistent slip bands (extrusions mainly) appear for stress amplitude higher than approximately 0.8 times of the endurance limit. At location loaded at stress amplitude close to the endurance limit a density of slip bands sharply rises and more coarse structures (clusters) appear. A strong connection between the endurance limit and the surface density of persistent slip bands was found.
1. Introduction
The evolution of material structure under cyclic loading has been the object of intensive studies during the last century. It has been shown that fatigue crack initiation in metals is accompanied by the appearance of specific dislocation patterns such as veins-channels structures, persistent slip bands-matrix structures, labyrinth, shell structures. Under cyclic loading the tendency of metals to form ordered structures is even more pronounced and the result is the formation of localized bands of intense deformation with well defined periodicity and amplitude known as persistent slip bands [1]. An adequate description of these processes should be based on more than a single relevant spatial or time scale, and each scale corresponds to phenomena that can be treated discretely or collectively. The creation of this description requires both a detailed theoretical investigation of the nonlinear laws of the defect kinetics and the development of new experimental techniques. One powerful way to obtain the nonlinear kinetic equations of the defect density based on a statistical physics approach was proposed in [2]. To progress in the development of this approach and to determine additional constants in the constitutive equations we
Table 1
Chemical composition of 35CrMo4 steel (in wt %)
C Mn Si S P Ni Cr Mo
0.37 0.79 0.30 0.010 0.019 < 0.17 1.00 0.18
strongly need the detailed investigation of the processes of plastic deformation, damage and failure responses in a wide range of loading conditions. Usually in high cycle fatigue no visible irreversible strain can be detected at the macroscopic level. Despite of this, even if the macroscopic stress state is homogeneous, heavy local distortions have been observed microscopically [3]. The localization of the plastic strain creates a heterogeneous distribution of heat sources on the surface of a specimen. It makes quite interest the investigation of the infrared radiation of the surface using infrared thermographics technique [4, 5].
In this paper we try both to find the connection between the stress amplitude and the specimen surface patterns and to estimate by infrared thermography the correlation between damage localization and surface patterns.
2. Experimental condition
The material under investigation was 35CrMo4 steel. Its mechanical properties and chemical composition are gi-
Table 2
Mechanical properties of quenched and tempered 35CrMo4 steel
Young’s Yield Maximum Fracture Fatigue Fatigue
modulus, stress tensile elongation, limit* at limit* at
GPa Re0.2> strength, % 107 cycles, 105 cycles,
MPa MPa MPa MPa
200 950 1068 11.5 525±37 660±21
* for smooth specimen in fully reversed plane bending
© S. Uvarov, E. Mikhailov, O. Plekhov, and T. Palin-Luc, 2004
Е
Е
Fig. 1. Specimen geometry (arrows indicate bending moment)
ven in Tables 1 and 2, respectively. The mean grain size of this steel is around 10 ^m. The specimens (Fig. 1) were machined from round heat treated bars according to the following procedure: 30 minutes in oven at 850 °C, then quenched in oil, then 1 hour at 550 °C, then leave in air for cooling. Before experiments all the specimens were polished by hand using emery paper and diamond powder up to grade 1 jwm.
The specimens were loaded in fully reversed sinusoidal plane bending on a resonance fatigue testing machine (test frequency = 56 Hz). The nominal stress amplitude was varied from 200 to 650 MPa. The tests were automatically stopped if the resonance frequency decreased more than around 10 %. The depth of the corresponding fatigue macrocrack detected is around 1 mm. During the tests, an infrared camera (CEDIP Jade III MWR) was used to record the evolution of the temperature field of the specimen surface. The camera could distinguish temperature peculiarities with thermal resolution up to 0.1 mK. The surface pattern of the specimen was investigated with a 3D surface profilometer NewView 5000 (lateral resolution up to 0.5 jwm, height resolution 0.1 nm).
3. Surface of the fatigue specimen
Since the flat surface of the specimen is loaded — due to bending — with a stress gradient, the correlation between surface topology and stress amplitude can be investigated. With the profilometer the microshear density on the specimen surface was estimated by scanning an area of 2.5x0.5 mm2 on the flat specimen surface, i.e. where the macroscopic stress gradient exists.
As shown in Fig. 2, it has been found that the first fatigue slip bands appear in areas loaded with a stress amplitude equal to 430 ±40 MPa which is around 0.8 times of the endurance limit. Length of slip bands is 5-10 jwm. This ratio between the stress amplitude above which first damage initiates (at the mesoscopic scale) on the specimen surface and the endurance limit is close to the same ratio found for a spheroidal graphite cast iron and corresponding to the appearance of microcracks arrested at a phase boundary [6]. In [7], fatigue slip bands were pointed as the most frequent origin of fatigue crack nucleation. The first micro-
8 6 & 4
(D
Q
♦ Microshear density 1 1 1 1 1 1
■ Cluster density i i j 1 1 1 i Н4Ч i i — bf1 i i ! !♦) L ! I i
I I I I _L I I I I I I
l T I I _ 1 I г Si i i
I I i i i i i 1 J L
I I I I I I I I I I
i T I I 1 T г 1 1 1
ri__ ■ i 1 ■ u ! ■ ■ "I ■ ■ Г ——1 1 1
400 500 600
Stress amplitude, MPa
700
Fig. 2. Microshear and cluster densities versus stress amplitude at different areas (70x50 ^m) on the same specimen loaded at 620 MPa after 65 200 cycles. There are small (1-3 mm length) cracks on the surface
cracks are detectable within the bands later during cycling. Some authors assume that the value of the stress amplitude creating the first fatigue damage is the ultimate fatigue limit [8], below which no damage appears after unlimited number of cycles.
Furthermore from Fig. 2 one can conclude that there is a sharp change in microshear density on the specimen surface area where the stress amplitude is 540 MPa. Width of the jump is around 50 MPa. This magnitude is close to the endurance limit value (if the scatter is considered) for this material (525 ±37 MPa). In areas loaded with a higher stress amplitude the microshear density has a plateau. The slight decrease of the density with the stress increasing can be related to the coalescence of slip bands. This threshold value corresponds also to the creation of slip band clusters (Fig. 6). Initial state of the surface is presented in Fig 3. Surface
Fig. 3. Initial (after polishing) state of the surface: a — 3D profile of the surface (the higher, the lighter); field of view is 70x50 ^m; b — 2D profile along a slice
Fig. 4. Surface of specimen loaded at 513 MPa stress amplitude: a — 3D profile of the surface (the higher, the lighter); field of view is 70 x 50 |um; marks show extrusions on the surface; b — 2D profile along a slice
location loaded at the stress amplitude 513 MPa is presented in Fig. 4.
The emergence of large scale structures like clusters in the form of walls (Fig. 5) and cells means that deformation and damage accumulation process within a cluster is strongly
Fig. 6. Surface of the specimen loaded at stress level 660 MPa after 65200cycles; field of view is 280x100 ^m; total length of observed walls of slip band is around 1 mm (all is not shown in this picture)
correlated. These “collective modes” can lead to the appearance of dynamic stochasticity in dissipation (temperature) dynamics at scales 0.1-1 mm. Temperature difference between cluster zone and surrounding should be also reasonable and detectable by the infrared camera. Damage accumulation and fatigue crack initiation observed by the infrared camera is shown in Fig. 7. Infrared image investigation results and discussion are presented in [9] and [10].
4. Conclusions and prospects
Different types of damage localization at the mesoscopic scale were observed after fatigue tests on 35CrMo4 smooth specimens by using 3D surface profiler. A strong connection
Fig. 5. Surface of specimen loaded at 618 MPa stress amplitude: a — 3D profile of the surface (the higher, the lighter); field of view is 70 x 50 |um; lines show clusters of slip bands; b — 2D profile along a slice
Distance, mm
Fig. 7. Infrared image of dissipation localization due to damage localization on an area loaded at 660 MPa (a); image size is 3.7 x 2.8 mm; a slice along the black line (b)
between the endurance limit and the surface density of persistent slip bands was shown. The appearance of mesoscopic defect structures consisting of strong interacting defects can lead to the correlated thermal behavior of adjacent points. Other investigations have to be carried out, for instance with IR camera and SEM observations and the interrupted test technique, to best confirm these the first conclusions.
Acknowledgement
Authors acknowledge the French Ministry of Research and the Russian Foundation for Basic Research (RFBR grant No. 02-01-00736-a, 03-01-06370-Mac, 04-01-96009) for financial support.
References
[1] S.J. Basinski and Z.S. Basinski, Dislocations in Solids, F.K. Nabarro (Ed.), V. 4, North-Holland, Amsterdam, 1969.
[2] O.B. Naimark, Defect induced instabilities in condensed matter, JETP Letters, 67, No. 9 (1998) 751.
[3] S. Suresh, Fatigue of Materials, Cambridge University Press, Cambridge, 1991.
[4] M.P. Luong, Infrared thermographics scanning of fatigue in metals, Nuclear Engineering and Design, 158, (1995) 363.
[5] G. La Rosa and A. Risitano, Int. J. of Fatigue, 22, 1 (2000) 65.
[6] T. Palin-Luc, S. Lasserre, and J.-Y. Bernard, Fat. Fract. Eng. Mat. Struct., 21 (1998) 191.
[7] M. Klesnil and P. Lukas, Materials Science Monographs, 71: Fatigue of Metals, Elsevier, Amsterdam, 1992.
[8] L.R. Botvina, Damage Kinetics of Construction Materials, Nauka, Moscow, (1989) 110.
[9] O. Plekhov, S. Uvarov, T. Palin-Luc, and O. Naimark, Investigation of Fatigue Crack Initiation and Growth in 35CD4 Steel by Infrared Thermography, in Proc. (CD-Rom) Int. Conf. Fatigue Crack Paths. Parma, Italy, (2003).
[10] O. Plekhov, T. Palin-Luc, N. Saintier, S. Uvarov, and O. Naimark, Fatigue crack initiation and growth in 35CrMo4 steel investigated by infrared thermography, Fat. Fract. Eng. Mat. Struct. (submitted in 2004).
