JOURNAL OF THEORETICAL AND APPLIED MECHANICS 44, 3, pp. 713-730, Warsaw 2006 FATIGUE FRACTURE OF NITRIDED AND CARBONITRIDED LAYERS Aleksander Nakonieczny Institute of Precision Mechanics, Warsaw, Poland e-mail: nakon@imp.edu.pl The phenomenon of fatigue fracture of nitrided and carbonitrided layers is described in the paper. Fatigue bechaviour is described by analysing properties of the substrate and surface layer. The effect of microhard- ness of the substrate on fatigue properties of nitrided layers is obtained. Smith’s diagrams for carbonitrided layersarepresented.Agraphicalpat- tern for prediction of the place of failure initiation is suggested. Future directions of the examination with tribology and corrosion effects taken into account are marked in the paper. Key words, fatigue fracture, nitriding, corbonitriding, hardness, fracture pattern 1. Introduction The level of applied technology is the decisive factor for product competitive- ness in the world market, but definitely not the only one. How competitive a product is depends on the entire process of manufacturing, service and recyc- ling possibilities. Among the most critical phases in this cycle are: • the phase of calculation and design • the phase of attaining material existence through various technological processes • the service phase. Added to the above is the final phase connected with utilization of used and worn products for newmanufacturing processes. 714 A. Nakonieczny All the above mentioned phases form the product’s ”lifetime”. The func- tionality of surface layers means the ability to fulfill requirements placed on the product in given service conditions. Utilization of products, which, in the majority of cases are machine components or assemblies, takes place in con- ditions where the product may be exposed to mechanical loading, thermal and/or chemical environment, and other hazards. Amongmechanical loadings, themost significant are thosewhichvarywith time, i.e. fatigue aswell as phenomena occurringduring the process of friction, measured in most cases by the amount of wear and coefficient of friction. Durability of products is decidedby surface layerswhich, dependingon the type of technological process applied, may have a thickness of orders of 1µm to several mm. The condition of the decisive surface layer is critical to wear resistance during the process of friction and to corrosion resistance. In the case of mechanical loading (especially fatigue resistance) as well as the action of corrosion (hydrogen corrosion), the critical role is played by the substrate, its condition and properties as well as the atomic relationship between the surface layer and the substrate. For this very reason, while considering life expectancy of machine com- ponents and assemblies, one should take into account the system: substrate– surface layer. Atributes to such a system are: thickness of the surface layer – ratio of this thickness to the entire cross-section, ratio of surface hardness to core hardness and the state of residual stresses, usually compressive within the surface layer, relative to the state of stresses in the substrate, which are usually tensile. An incorrectly applied surface layer may cause formation of a structural flaw in the transition zone of the layer and may lead to crack initiation, especially by the fatigue mechanism (Nakonieczny, 1984). The volumeheat treatment and other processes belonging to surface treat- ment do not function independently, but are an important part of the general domain of surface engineering. The surface engineering encompasses technology behind formation of sur- face layers, but also defines service properties of products, surface layer inve- stigation methodology as well as design considerations for the substrate-layer system according to given service conditions. Service properties of products, and hence the functionality of the surface treatment, may be assessed by defining the fatigue limit, wear resistance or corrosion resistance. Such evaluations are usually performed on specimens in laboratory conditions. However, most valuable information is to be gained from actual service trials. The costs of such investigations are, unfortunately, high and difficult to bear by smaller andmedium size companies. Fatigue fracture of nitrided... 715 2. The concept of structural notch Fatigue resistance of machine components is a function of their design, ma- terial and technological parameters as well as the type of loading in service conditions. When discussing the problem of fatigue resistance, one should di- scuss in detail the effect of these parameters, and in the case of loads, define fatigue characteristics, e.g. Smith’s plots for different types of loading such as bending, tension and torque. These problems have been sufficiently dealt with in technical literature. In the process of searching for methods of increasing fatigue resistance, there occur some constant elements, the implementation of which has a fa- vorable effect on it. To these belong enhancing treatments such as thermal, thermo-chemical as well as surface work hardening. In order to raise fatigue resistance, it is not sufficient to apply a chosen enhancement treatment, but rather of utmost importance is appropriate selec- tion of the initial volume heat treatment prior to successive surface, thermal and work hardening treatment. The problem of enhancing fatigue resistance of machine components by technological methods does not consist of appli- cation of one chosen treatment, but of a cycle of successive treatments. The appropriate selection of these treatments affects the structural noth formed in the process of enhancement, which has decisive influence on fatigue resistance (Nakonieczny, 1984). A structural noth occurs in all locations where, as a result of heat tre- atment (e.g. induction hardening), thermo-chemical treatments (carburizing, nitriding, etc.) orworkhardening treatments (burnishing, shot-peening)ofma- chine components, the layer formed in these processes has different physical- chemical properties than that of the core, due to a large gradient of property changes. The value of the structural flaw coefficient β depends on the type of material and parameters of technological processeswhich cause this structural noth to form. In other words, it depends on the heat treatment and surface hardening. Industrially utilized thermal and surface work hardening treatments cause enhancement of fatigue resistance. Based on the research carried out at the Institute of Precision Mechanics (IMP), it can be accepted that as a result of implementation of such treatments, the fatigue limit (σ −1) rises on average by 15 upto 30%. The enhancement of fatigue resistance is obtained by structural changes, strengthening and favorable distribution of residual stresses which are formed as a result of thermal and surface treatments. 716 A. Nakonieczny Due to physical processes taking place within the material during the ap- plication of surface, thermal and work hardening treatments, changes in the microstructure andmechanical properties arise between the surface layer and the core of thematerial. The gradient of changes of physical-chemical proper- ties depends on the selected technological process and its parameters. Nume- rous instances of fatigue cracks have been noted, whose origins were traced to the transition zone between the hardened surface layer and the substrate. Figure 1 shows, as an example, a fatigue fracturewith the origin located under the hardened layer at the point where stresses are mounted. Themounting of stresses occurs as a result of residual stresses created during heat treatment and external stresses. Fig. 1. Location of initiation of fatigue cracking on nitrided 40HM (4140) grade steel –×100 3. The role of substrate For full evaluation of fatigue properties of thermally or mechanical treated materials, it is important to understand themechanism of failure. Understan- ding thismechanism is possible by determination of the condition andmutual relationship between the substrate and the surface layer. Solution to this pro- blem becomes possible through determination of the strength condition of the system: substrate–surface layer as a function of external loading. The investigations were conducted on structural steels 40HM (4140) and 38HMJ (Nitralloy 135M). The steels were hardened and tempered prior to Fatigue fracture of nitrided... 717 nitriding at two temperatures: 550◦C and 620◦C (Nakonieczny, 1984; Babul et al., 1996; Nakonieczny and Tacikowski, 1994). Nitridingwas carried outusing two types of atmosphere, i.e.NH3-NH3(diss) and NH3-N2. In the nitriding processes, the atmosphere gas composition was varied, as were the time of nitriding (4 and 16h) and nitriding potential –KN (from 1.65 to 4.8). Fatigue resistance tests were carried out on the PUNN machine (ma- nufactured by Schenck), applying rotational bending stresses with a notch (α = 1.02). The specimen diameter was Φ = 5.88± 0.02mm. The results of the fatigue resistance tests, metallurgical evaluation and process parameters are put together in Table 1. Figure 2 showsmicrohardness traverses in thenitridedcase for 40HMgrade steel, while Fig.3 shows the same for the 38HMJ grade. The results of tests show that the tempering temperature has an effect on the properties of the nitrided case. The effect of the tempering temperature on the basic properties of the nitrided case depends on the steel grade. A higher increase of hardness (by about 50%) as well as of case depth is observed on the low alloy chromium steel 40HM. Higher valus of the fatigue strength limit can be observed for higher hard- ness of substrate. The results of investigation shown inTable 1 that for 4140 steel and hard- ness 402HV0.5 and 396HV0.5 the fatigue streght limit have values 820MPa and 840MPa, respectively. For Nitralloy 135M we cannot observe the same phenomenon. For this material, we have another parameters of the surface layer, esspecially hardness and depth at which residal stress changes the sign. 4. Fatigue characteristics after carbonitriding Inmostmodernmethods of manufacturing it is recommended that the design stage consider different manufacturing technologies. In connection with that there is an urgent need to determine material characteristics, especially of materials after application ofmodern enhancing treatments. There also exists theneed todevelop calculationmethods of fatigue resistance after thermal and thermo-chemical treatment. This, however, is the next stage of the activity, possible to carry out only when the basic fatigue properties of steel following a heat treatment will be known. 7 1 8 A . N a k o n ie c z n y Table 1.Technological parameters and test results Steel Tempering Nitriding Core HV0.5 Hardness Fatigue Residual stresses grade temperature time hardness Max. on On limit At surface Depth at which [◦C] [h] HV0.5 cross- surface σ −1 [MPa] stress changes section [MPa] sign [mm] 40HM 550 4 402 677 757 820 600 0.32 (4140) 550 16 396 642 757 840 650 0.52 620 4 343 715 826 735 600 0.37 690 16 343 343 642 745 900 0.55 38HMJ 550 4 356 1030 1373 805 900 0.25 (Nitralloy 550 16 343 1030 1227 785 600 0.48 135M) 620 4 318 1030 1273 766 450 0.30 620 16 296 1030 1304 810 800 0.45 Fatigue fracture of nitrided... 719 Fig. 2. Microhardness traverses across the nitrided case on 40HM (4140) grade steel; 1 – tempering temperature 550◦C, time 4h; 2 – tempering temperature 550◦C, time 16h; 3 – tempering temperature 620◦C, time 4h; 4 – tempering temperature 620◦C, time 16h Fig. 3. Microhardness traverses across the nitrided case on 38HMJ (Nitralloy 135M) grade steel; 1 – tempering temperature 550◦C, time 4h; 2 – tempering temperature 550◦C, time 16h; 3 – tempering temperature 620◦C, time 4h; 4 – tempering temperature 620◦C, time 16h The carbonitriding treatment is used for components exposed to lighter loads and subjected to wear as well as bending (Tacikowski andNakonieczny, 1992; Nakonieczny, 1991a,b; Winderlich, 1990; Kogaev et al., 1985). For those components, which are subjected during service to contact fatigue, the case depths are designed deeper. For the present series of tests, a case depth of 0.7mmwas selected. 720 A. Nakonieczny The optimum microstructure of carbonitrided components is fine acicular martensite with a small amount of retained austenite and not containing co- arse carbide precipitations (Tacikowski and Nakonieczny, 1992; Nakonieczny, 1991a,b; Winderlich, 1990; Kogaev et al., 1985). Specimens prepared to meet the above conditions were subjected to rota- tional bending and one point bending fatigue tests. Such types of loadingwere selected based on the premise that bending is the most common method of loading during service as well as the fact that during bending one can observe the most favorable and strongest effect of surface strengthening. Simplified Smith’s curves were plotted to determine the fatigue resistance for at least three methods of bending (the Wöhler curve) as well as static strength and yield strength for a given type of loading andmaterials. The fa- tigue tests were carried out with the following coefficients of cycle asymmetry: R = −1; R = 0.1 and R = 0.3. The coefficients of 0.1 and 0.3 were selected to ensure possibility of running the tests only within the range of one sided bending stresses. The fatigue characteristic was developed for a material in the quenched and tempered condition, as a reference, and for materials with a diffusion case, heat treated to the same condition as the only heat treated version. Once these values were known, the surface coefficient of strengthening was determined from the equation m= σww −1 σ −1 (4.1) where σww −1 – fatigue limit of the enhanced specimen σ −1 – fatigue limit of the reference specimen. Bending tests were carried out on the Amsler machine. In order to ob- tain the bending effect on this machine, a prototype additional element was designedwhich, througha lever applies loading of the tested section of the spe- cimen under a constant bendingmoment (Fig.4). The frequency was 150Hz. The tests were carried out to NG =10 7 cycles. The material used in these tests was the 18HGT grade, normalized, with a fine grained ferrito-pearlitic microstructure. Carbonitriding of specimens made of 18HGT grade steel was carried out at temperature 860◦C in an endothermic atmosphere, enrichedwith ammonia and natural gas. Metallurgical evaluationswere carried out on 18HGTsteel in the quenched and tempered only and carbonitrided conditions. Fatigue fracture of nitrided... 721 Fig. 4. Schematic of equipment for fatigue testing In the normalized only condition, the specimens showed ferrito-pearlitic microstructure with very fine-grained pearlite (Tacikowski and Nakonieczny, 1992; Olszański, 1977; Olszański et al., 1979; Wyszkowski, 1974). Themicrostructure of specimenswithdiffusioncaseswasdeterminedbased on photomicrographs andmicrohardness measurements. Specimensmade of 18HGT steel, after carbonitriding and quenching with tempering exhibit in the subsurface zone a microstructure of tempered mar- tensite (Fig.5). In order to determine the bending resistance, a static bending test was carried out. For the heat treated (normalized) only steel, it was not possible to obtain a static bending strength value because of ductility of the material. Only the yield strength was determined, and for the specimens it amounted to 822.8MPa. The static bending test carried out on carbonitrided specimens is shown in Fig.6. In this case it was not possible to determine the yield strength, and only the elastic limit in the point Awas established as 2289.8MPa. An analysis of results of the static tests delivers newdata. The tensile plot for the carbonitrided material is characteristic for brittle materials. There is no necking and no elongation of the specimen. Similar behaviour was noted when the bending strength test was performed. In no case it was possible to determine the yield strength. 722 A. Nakonieczny Fig. 5. Microstructure of the carbonitrided case on a specimenmade of carbonitrided 18HGT grade steel. Etched by Nital – ×500 Fig. 6. Plot of the static bending test of a specimen of carbonitrided 18HGT grade Based on fatigue tests for rotational bending, which were performed on specimensmade of carbonitrided andheat treated (normalized) onlymaterial, it was possible to determine the coefficient of surface strengthening, i.e. the ratio of m=σ −1 (with diffusion case) to σ−1 (with no case). For 18HGT steel after carbonitriding this coefficient was 2.48. With the aid of results obtained in static and fatigue tests for the case of two and one side bending, simplified Smith’s curves were plotted for the heat treated only material and for the carbonitrided material (Fig.7). To plot the Fatigue fracture of nitrided... 723 chart, values of unlimited fatigue resistancewere used fromtheWöhler curves. The upper limit of the chart for the heat treated only material is the yield strength obtained from the bending test. Fig. 7. Simplified Smith plot for 18HGT grade steel after normalizing (1) and carbonitriding (2) TheSmithcurve formaterials after thermo-chemical treatmentdiffers from that for the reference heat treated only material because its upper limit is determined by the bending yield strength Rg (Fig.7). A designer of the com- ponent, basing his design on the presently available tables containing data of the ultimate properties of the steel after hardening and fatigue properties for alternating stresses, creates a design, which consumes big amounts of thema- terial. As a result of using values of σ −1 taken from catalogues, the strength of the assembly is compromised. The values of the fatigue resistance σgR (where −1