Predictive modeling and experimental results for laser hardening of AISI 1536 steel with complex geometric features by a high power diode laser
Introduction
Many different types of components require heat treatment to produce desired surface qualities, such as wear resistance. One typical treatment is case hardening, where the surface is hardened to a specified depth. Industrial examples include bearing races, cylinder liners, gears, pistons, piston rings, spindles and valve seats [1]. Traditionally used heat treating processes include induction hardening, arc hardening, electron beam hardening, gas carburizing, nitriding and carbonitriding [2].
An alternative process to the traditional methods is laser hardening, also referred to as laser transformation hardening, which is a selective heat treating method. The laser beam has large power intensity that can be shaped to produce a desired power distribution, which rapidly heats the surface region of the workpiece. Because the power profile is shapeable, discrete patterns across a workpieces surface can be hardened [3]. The bulk of the material is unaffected and acts as a heat sink to cause rapid quenching, producing large cooling rates within the workpiece. A martensitic surface with a maximum case depth of 2 mm [4] can thus be produced; yielding a surface that can better withstand corrosive environments, high temperatures, and/or wear and have better fatigue properties. At the same time, the bulk of the material is unaffected so that it retains its original properties, such as ductility [2], [5].
When compared to the other processes, laser hardening causes little deformation of the part, so that post machining is practically eliminated. The energy input is more efficient, because only the portion of the part undergoing treatment is heated. The treatment can be carried out in air with no external quenching media. If the area to be hardened can be seen optically, then it can be hardened with a laser. A wider variety of materials can be hardened; even low carbon steels can be hardened because of the rapid heating and cooling rates caused by the laser. The resultant microstructure is often better than that produced by other methods, i.e., it is harder and more fatigue resistant [2].
Laser hardening is not without its own disadvantages. Lasers generally involve higher capital costs than other systems, but the lowered operating costs with high processing speed and elimination of coils, furnace and quenchant can help offset this. Also, if the laser tracks are overlapped, back tempering can occur so that a non-uniform hardness distribution is created. However, proper control of the laser power and overlapping tracks will mitigate this problem [2], [6].
During laser hardening of hypoeutectoid steels, two processes occur simultaneously. As the temperature rises above the eutectic temperature, A1, pearlite colonies transform to austenite. Carbon in the newly formed austenite, which is high in carbon content, then begins to diffuse to the low carbon ferrite. Given sufficient time, the carbon distribution will become homogeneous. Because the heating and cooling rates in laser hardening are very large, a homogenous distribution is not reached, but some of the ferrite may gain sufficient carbon to transform to martensite upon cooling. If the austenization temperature level, A3, is exceeded, then any remaining ferrite will transform to austenite. Because only a small portion of the workpiece is heated, the workpiece is quenched by conduction to the unheated region of the workpieces, causing high cooling rates and hardness values [7].
Widespread use of lasers in industry for heat treating has not occurred because of the unavailability of a laser that could produce a high enough energy flux into metals, as metals are notorious reflectors at the CO2 wavelength, which until recently were the only lasers that could produce a high enough energy density [2], [8]. To effectively harden a workpiece with a CO2 laser, a coating must first be applied to increase the absorptivity at the CO2 wavelength, which incurs extra cost to the process for application and removal. Newer high power diode and Nd:YAG lasers produce wavelengths in which metals show greater absorptivity so that they eliminate the use of coatings.
Early studies on laser hardening involved predicting the temperature within the workpiece and then gauging the microstructure that might occur. Carbon diffusion was ignored so that, if a critical temperature was reached, then the material at that layer was assumed to be transformed into martensite. Kou et al. [9] and Kou and Sun [10] developed a two-dimensional thermal model for cylindrically shaped parts, using a finite difference method with a varying mesh grid. Temperature-dependent material properties, including absorptivity, were considered within the model. The critical temperature predicted was 785 °C, which the predicted case depth was based upon, and thus a difference from this temperature would lead to deviations in the hardness predictions.
Ashby and Easterling [7] developed both a thermal and a kinetic model for laser hardening, modeling both hypo- and hyper-eutectoid steels. A simplified heating model was utilized, from which a “kinetic force” based upon maximum temperature and cycle duration was determined. In order to simplify the problem, temperature-dependent material properties were not used. The kinetic model took into account initial microstructure's transformation from ferrite and pearlite to austenite, followed by its quenching to martensite. Laser processing diagrams were created to show how process variables affect hardness and case depth. Tests were conducted using a low (0.1% C) and medium (0.6% C) carbon steels with laser powers between 0.5 and 2.5 kW, using Gaussian and “top hat” energy profiles. In order to improve the absorptivity, carbon black dispersed in alcohol with a binder was painted on the surface. The data presented showed a good agreement between the experimental and predicted results; however, the simplifying assumptions cast doubt upon the agreement. For instance, an absorptivity value of 0.7 was used, but no basis as to how this value was determined was discussed. No temperature measurements were presented either, which are needed for thermal model validation.
Similar kinetic models to those of Ashby and Easterling [7] were developed by Safonov et al. [11], [12]. In both cases, slight changes were made to the model proposed by Ashby and Easterling. In the earlier paper, comparisons were made between their model and an analytical solution, with a good agreement found. In the later paper, the kinetic model was again refined to include the dependence of the diffusion coefficient on temperature and carbon concentration and solved as a multifront Stefan problem. The model was compared with experimental results, which showed a case depth larger than predicted value by 20–30%.
Patwa and Shin [13] coupled the kinetic model concepts of Ashby and Easterling [7] with the thermal model for a rotating, cylindrical body undergoing laser heating developed by Rozzi et al. [14] for heat treating of AISI 5150H by a CO2 laser. Both the thermal and kinetic models were experimentally validated with both showing good accuracy. A case depth of 0.5 mm with a hardness of 62 HRC was achieved with initial microstructures of ferrite/pearlite and tempered martensite.
While a one-dimensional diffusion model can provide the case depth achieved during laser hardening, it cannot provide any information about the resulting microstructure. It is also unable to handle complex geometries, where the changing geometry may cause large thermal gradients in multiple directions. Inoue et al. [15], [16] and Ohmura et al. [17] conducted research on multidimensional diffusion during laser hardening. The first step where pearlite transforms to austenite is modeled using Fick's 2nd law of diffusion in two dimensions with appropriate boundary conditions and solved using the alternating direct implicit method. The homogenization of austenite was then modeled in one dimension, similar to Ashby and Easterling [7]. The rationalization given is that the heat affected microstructure varies most significantly in the z-direction and thus this approach is not truly a multi-dimensional model.
Jacot et al. [18], [19], [20] presented a true two-dimensional diffusion model for the austenization of pearlite for hypoeutectoid steels. In the earliest paper, only the homogenization of austenite is considered, as the pearlite to austenite transformation is considered to be instantaneous. In the subsequent two papers, they developed the model further so that the initial microstructure of ferrite, pearlite and ferrite-pearlite interface cells could be considered. As the workpiece is heated above the eutectic (A1) temperature, austenite begins nucleating within the interface and pearlite cells. A kinetics law, corresponding to the interlamellar spacing of the pearlite and thermal history, is used to govern the growth of austenite from pearlite. At the same time, diffusion occurs within the austenite and ferrite, which is governed by Fick's law. The model was developed to determine the amount of diffusion time required for homogenization of austenite for a 0.49% C steel. Only an isothermal temperature profile was considered, such as what occurs in furnace hardening, whereas in laser hardening the temperature profile varies over the width and depth of the laser track.
The current study focuses on developing a predictive model for laser hardening of complex parts by coupling the 2D diffusion model with a 3D transient thermal model. The goal of this study is to harden the surface on a steel alloy crankshaft with a complex geometric feature as deeply as possible without surface melting. A high power diode laser used eliminates the use of coating, thus making the process more industrially viable. The previously validated thermal model is again validated for AISI 1536 to predict the heat flow within the workpiece, from which a thermal history is acquired. A numerical two-dimensional kinetic model is developed and coupled with the thermal model to predict the phase transformations occurring during laser hardening based upon the thermal history. Afterwards, microstructural analysis is performed to characterize the resultant microstructure and measure its microhardness. Finally, laser hardening capability of complex geometry is established through both analytical modeling and experimental studies.
Section snippets
Experimental facilities
A Nuvonyx ISL-4000L 4.0 kW direct diode laser mounted on a Panasonic VR-016 welding robot provided the optical energy that is focused upon the workpiece during laser hardening, which was held in a three-jaw chuck mounted to the YA-1GJB11 positioner (Fig. 1). A total of seven degrees of freedom were available with the setup, the top speed of the robot was 120 m/min and the maximum rotary speed of the positioner was 25 rpm. An air knife blew across the laser output window, which kept debris away
Cylindrical thermal model
Rozzi et al. [14], [24] developed a thermal model to predict temperatures throughout a workpiece during laser-assisted machining (LAM) of an opaque homogeneous ceramic workpiece, as shown in Fig. 4. During LAM, a low-pressure gas assist jet is used to protect the laser's focusing optics and impinges on the workpiece surface in the area of jet interaction of Fig. 4. However, the air knife used to protect the diode laser's output window did not impinge upon the surface. Eq. (1) shows the
Laser hardening of simple geometry parts
The experimental matrix of Table 7 was designed to explore the maximum case depth possible using the setup described above while tracing a spiral path along the surface of a cylindrical workpiece. Additionally, the surface melting and overlapping laser tracks were avoided to minimize part distortion and any tempering, respectively. Tests LHD-1, LHD-2 and LHD-3 were repeated twice to show repeatability of results, the first test of the sequence is referred to as LHD-#-1 and the second is
Conclusions
A numerical two-dimensional kinetic model was developed and coupled to the three-dimensional thermal model to predict the hardening that occurs in steel parts during laser heating. The model was able to properly predict the hardness of an isothermally heated workpiece, as validated by furnace tests. The absorptivity of oxidized steel was determined to be 0.67 and the emissivity found to be 0.90, through flame and laser heating. The diode laser, which provides a larger spot size than the CO2
Acknowledgement
This research was conducted at the Center for Laser Based Manufacturing and funded by the Indiana 21st Century Research and Technology Fund. Authors also wish to thank Mr. Benxin Wu in helping the prismatic thermal modeling.
References (28)
- et al.
J. Mater. Process. Technol.
(2004) - et al.
Surf. Coat. Technol.
(2002) - et al.
Acta Metall.
(1984) - et al.
Acta Metall.
(1997) - et al.
Acta Mater.
(1998) - et al.
Acta Metall.
(1999) - et al.
Int. J. Heat Mass Transfer
(2000) - et al.
Int. J. Mach. Tools Manuf.
(2006) Met. Sci. Heat Treat.
(1997)Surf. Eng.: Jt. Publ. Inst. Met. Wolfson Inst. Surf. Eng.
(2002)
J. Opt., A Pure Appl. Opt.
Mater. Sci.
Adv. Mater. Process.
Metall. Trans., A, Phys. Metall. Mater. Sci.
Cited by (71)
Theoretical study of laminar plasma quenching of rail steel using the fluid-solid-thermal coupling method
2022, International Journal of Thermal SciencesCitation Excerpt :The degree of austenite homogenization is affected by carbon diffusion. The higher the temperature, the more intense the carbon diffusion, and the easier the formation of uniform austenite [41]. At this time, the calculated austenite volume fraction at the node is approximately equal to 1.
Metallurgical characterization of laser hardened, mechanically textured 2.5 Ni-Cr-Mo low alloy steel and optimization using RSM
2021, Optics and Laser TechnologyCustomized laser beam intensity distribution for the laser surface treatment of geometrically convoluted components
2019, Journal of Materials Processing TechnologyCitation Excerpt :The software that controls the servomotors of the scanner uses this information to calculate the combined rotation of the two galvanometric mirrors. The most suitable ELPDD for a given process can be estimated by thermal analysis as it is done in the study of Yanez et al. (2002), for complex geometries or by Skvarenina and Shin (2006) where the evolution of the micro-structure is calculated. Fig. 2 shows two examples of flat geometrical patterns, made of several paths, and the resulting ELPDD, by iteratively scanning them with a prescribed speed for each path.
Laser-assisted selective fusing of thermal sprayed Ni-based self-fluxing alloys by using high-power diode lasers
2018, Optics and Laser TechnologyHigh throughput hybrid laser assisted machining of sintered reaction bonded silicon nitride
2018, Journal of Materials Processing Technology