- Email: [email protected]
Experimental and numerical analysis of failures on a die insert for high pressure die casting
Engineering Failure Analysis 95 (2019) 171–180
Contents lists available at ScienceDirect
Engineering Failure Analysis journal homepage: www.elsevier.com/locate/engfailanal
Experimental and numerical analysis of failures on a die insert for high pressure die casting R. Markežiča, I. Nagličb, N. Molea, R. Šturma, a b
T
⁎
Faculty of Mechanical Engineering, University of Ljubljana, Aškerčeva 6, 1000 Ljubljana, Slovenia Faculty of Natural Sciences and Engineering, University of Ljubljana, Aškerčeva 12, 1000 Ljubljana, Slovenia
A R T IC LE I N F O
ABS TRA CT
Keywords: Die failures Nitriding Finite element analysis Microanalysis X-ray analysis
A comprehensive study of failures on a used insert of a die for high pressure die casting of aluminium alloy AlSi9Cu3Fe was conducted. The analysed insert was made of hot work tool steel Dievar, heat treated and plasma nitrided according to specifications. Before the start of analysis, the insert was subjected to 170,000 die casting cycles. X-ray diffraction, light microscopy, scanning electron microscopy and hardness measurement methods were used for experimental analysis of soldering, corrosion, erosion and thermal fatigue failure modes on the insert. A special procedure, which combines use of commercial software MAGMA5.3 and open source software CalculiX, was developed for numerical calculations of temperature fields on the insert. Experimental and numerical results showed strong dependence between die surface temperatures and die failure modes. Hardness drop, nitride diffusion layer removal and microstructural changes were observed at more thermally affected areas. Surface crack with a thin oxide and thicker soldered layer was identified and analysed.
1. Introduction High pressure die casting (HPDC), designated in this article as die casting, is a technology that enables fast and high quality production of geometrically complex parts with high mechanical and dimensional requirements. Due to high absolute values and rapid fluctuations of temperature, pressure and melt velocity, die casting dies are exposed to cyclic thermo-mechanical loading, which gradually leads to the occurrence of wear and damage. Typical maximum melt velocities during die filling are between 30 m/s and 100 m/s [1]. The additional pressure applied during the solidification phase is of magnitude from 50 MPa to 80 MPa [1,2]. One of the most important parameters in a die casting process is the temperature of the die. A lot of studies have been conducted to monitor the temperature at the die surface during a real die casting cycle [2–4]. Norwood et al. [2] performed temperature measurements during casting of aluminium alloy Al8Si3Cu with use of thermocouples and temperature-sensitive paints. The maximum and minimum measured temperatures were in the range from 400 °C to 450 °C and from 150 °C to 200 °C, respectively. Dargusch et al. [3] and Long et al. [4] measured temperatures in the die surface layer during the casting process of aluminium alloy AlSi9Cu3Fe. From the temperature measurements the heat transfer coefficient was determined and then used in a numerical simulation for surface temperature calculation. The minimum and maximum calculated temperatures were 240 °C and 500 °C, respectively. Die wear and damage mechanisms are generally divided in four groups: (1) soldering, (2) corrosion, (3) erosion, (4) thermal fatigue and cracking. ⁎
Corresponding author. E-mail addresses: [email protected] (R. Markežič), [email protected] (R. Šturm).
https://doi.org/10.1016/j.engfailanal.2018.09.010 Received 22 May 2018; Received in revised form 30 August 2018; Accepted 13 September 2018 Available online 15 September 2018 1350-6307/ © 2018 Elsevier Ltd. All rights reserved.
Engineering Failure Analysis 95 (2019) 171–180
R. Markežič et al.
Soldering (or die sticking) is seen as the adhesion of cast material to the die surface during the die filling and solidification phases. Soldering can be subdivided into two categories [5–8]: (1) metallurgical soldering, characterised by a high number of loading cycles and high temperature of the die surfaces [5,9–14] (2) mechanical soldering, characterised by a low number of loading cycles and high melt pressures [5]. A soldered layer is usually the product of the simultaneous action of metallurgical and mechanical soldering. Corrosion of dies in die casting is characterised by loss of die material from the die surface and it is a metallurgical process [14]. There is often a simultaneous occurrence of corrosion and soldering on a particular region of a die surface [14]. Erosion is also characterised by loss of die material from the die surface, but it is a mechanical process [15]. An increase in the occurrence of erosion with an increase of die surface temperature was identified in several studies, [16–18]. Baker et al. [19] observed the appearance of cracks on previously eroded die surfaces. The main cause for thermal cracking of die casting dies are high temperature and pressure gradients [20–22,24]. The most influential parameters on thermal cracking are: (a) maximum heating temperature and heating/cooling rate, (b) oxidation, (c) hardness and microstructure of the die material. Research has shown, that crack initiation and propagation are increased by higher heating temperatures and higher heating/cooling rates [22,24–28]. In some studies it has been observed, that crack growth is increased by surface oxidation [29–31]. From thermal fatigue tests it has been found out, that crack density and crack depth decrease with increasing surface hardness [26–28,32,33]. Thermal fatigue also affects the microstructure of the die material. It causes thermal softening of the die surface, which is seen as surface hardness drop [26,29,30,34,35]. Thermal softening and hardness drop are accelerated by higher heating temperatures [25,27,28,31]. To reduce the occurrence of damage and wear mechanisms on die casting dies, it is important to choose an appropriate material, heat treatment and optionally surface treatment of the die [29]. One of the most often used surface treatments on die casting dies is nitridation. Joshi et al. [12] found out, that nitridation decreases the rate of metallurgical soldering on die surfaces. However, soldering on nitrided surfaces is present due to mechanical soldering after the formation of thermal cracks and erosion of the nitrided layer [6]. Persson [27] discovered that nitrided surfaces have better resistance to thermo-mechanical fatigue than untreated surfaces do. An increase in the thermal stability of the microstructure of nitrided surfaces has also been observed. The final crack depth on a nitrided surface is dependent on the nitridation depth, because crack growth stops at the interface between diffusion zone and substrate material [36]. In this paper, a comprehensive analysis of a used insert of a die casting tool, made of hot work tool steel Dievar [37], is presented. The die casting tool was used for casting of aluminium alloy AlSi9Cu3Fe. The focus is on determining the influence of temperature on the occurrence of individual wear and damage mechanisms. 2. Experimental analysis 2.1. Materials and samples preparation A three-dimensional model of the analysed insert is presented in Fig. 1a. The chemical composition of hot work tool steel Dievar, used for production of the analysed insert, is presented in Table 1. Before use, the insert was heat treated and surface treated according to specifications. Heat treatment was composed of heating the insert to the austenitising temperature of 1020 °C for 30 min, followed by quenching in warm oil and final double tempering for 2 h at 550 °C. After heat treatment, the insert was additionally surface plasma nitrided for 10 h at 480 °C to reach the final nitridation depth of 150 μm and final surface hardness of 1100 HV. The die casting tool with the analysed insert was used for die casting of aluminium alloy AlSi9Cu3Fe. The die casting cycle consisted of spraying the insert for 2 s, blowing the insert for 5 s, dosing the melt at a temperature of 670 °C, filling of the die, and solidification of the casting for 10 s. The duration of the entire die casting cycle was 62 s. The insert was exposed to 170,000 die casting cycles and then removed from the die casting tool for analysis. During die casting process the yellow coloured surface was in contact with melt and hence thermo-mechanically loaded. This surface will be in this paper designated as the active surface. During die casting process
Fig. 1. Insert presentation and samples preparation: (a) three-dimensional model of the analysed insert of dimensions Ø35x44.73 mm, (b) samples cutting scheme. 172
Engineering Failure Analysis 95 (2019) 171–180
R. Markežič et al.
Table 1 Dievar chemical composition [37]. Elements
C
Si
Mn
Cr
Mo
V
Wt%
0.35
0.2
0.5
5.0
2.3
0.6
the red coloured surface was in contact with other elements of the die casting tool. Two samples were cut out of the insert for light microscopy and scanning electron microscopy. One sample was cut in such a way that it contains a part of the active surface and represents the loaded part of the insert after 170,000 casting cycles, as shown in Fig. 1b. This sample will be designated as SL1 (sample-loaded 1). Another sample was cut from the opposite side of the insert and does not contain a part of the active surface, as shown in Fig. 1b. It represents the initial state of the insert at the beginning of it's operational life. This sample will be designated as SU1 (sample-unloaded 1). Samples for light microscopy and scanning electron microscopy were first ground and polished with diamond paste. Final polishing was performed with 0.05 μm silica. Samples for light microscopy analyses were additionally etched with Vilellas reagent. After finishing the analysis of samples SL1 and SU1, three additional samples were cut out from the insert, as shown in Fig. 1b: (1) sample representing the base material, designated as SB2 (sample-base material 2), (2) sample representing the loaded part of the insert, designated as SL2 (sample-loaded 2), (3) sample representing the initial state at the beginning of inserts operational life, designated as SU2 (sample-unloaded 2). Samples SB2, SL2 and SU2 were used for X-ray diffraction analysis. 2.2. Methods Samples were experimentally analysed with four different methods: X-ray diffraction (XRD), light microscopy, scanning electron microscopy (SEM) and hardness measurements. A scanning electron microscope JEOL JSM-7600F equipped with energy-dispersive Xray spectrometer X-Max 20 SDD-EDS detector was used for characterisation of the microstructure. Images and analyses were made at an acceleration voltage of 10 kV. X-ray diffraction (XRD) patterns were acquired with use of X-ray diffractometer PANalytical X'Pert PRO. Non-monochromated X-rays (0.15418 nm) produced by the Cu target tube were used in these experiments. Hardness measurements were conducted on Ernst Leitz Orthoplan equipment with additional microhardness pneumatic indenter with use of the Vickers method. Analyses were made on different positions of surface layers of samples SL1 and SU1, depicted in Fig. 2a and Fig. 2b, respectively. 3. Numerical analysis The goal of the numerical analysis was to obtain time-temperature profiles of the whole die casting cycle in the die surface layer. For this purpose, a special procedure was applied, which includes the use of the commercial software MAGMA (release 5.3) and the open source software CalculiX. The reason for using this procedure is in the computation methods used by each software. MAGMA is a commercial software which allows calculation of different influential parameters through the whole die casting cycle. The temperature computations in MAGMA are based on the Cartesian cut cell method, which is a variation of the finite volume method (FVM). CalculiX is an open source software which also allows temperature computations, but in this case the calculations are based on the finite element method (FEM). The steps of the applied procedure for temperature computations were the following: (1) casting material and die temperature field computation with the use of MAGMA, (2) temperature field computation of the analysed die insert with the use of CalculiX and applied film boundary conditions by known castings and other tool elements temperatures from step 1. For the die preparation phase, the temperature-dependent film coefficients for spraying, blowing and contact of die with air from the MAGMA library were applied. For the die filling and solidification phase, the casting-die interface temperature-dependent film coefficient for casting of aluminium alloy AlSi9Cu3Fe, reported by Long et al. [4], was applied. For the boundary conditions at the interface between the analysed die insert and other die elements, a temperature independent film coefficient of value 2000 W/m2 K
Fig. 2. Presentation of the analysed positions on samples: (a) sample SL1, (b) sample SU1. 173
Engineering Failure Analysis 95 (2019) 171–180
R. Markežič et al.
Fig. 3. Comparison of cross-sections of meshes used in simulations with: (a) MAGMA (FV mesh), (b) CalculiX (FE mesh).
was used. In real production the die is preheated to a certain temperature with the use of the internal tempering system before the start of the die casting process. This condition was taken into account in computations with MAGMA by setting the initial temperature condition of the die to 150 °C. After the start of the die casting process in real production, it takes several casting cycles to achieve stable cyclic temperature conditions on the die. During computations in MAGMA, five heating cycles and three production cycles were simulated to reach stable cyclic temperature conditions on the die. During heating cycle simulations no die filling is computed, so it is assumed that the whole die is filled at the same moment. While during production cycle simulations, the entire die filling is computed step by step. The temperature fields from the last production cycle were used for the definition of boundary conditions in simulations with CalculiX. In Fig. 3a and b cross-sections of meshes used in the simulations are depicted. From comparison of Fig. 3a and Fig. 3b we can conclude, that with the FE mesh in CalculiX, geometrical features, like radiuses, are better described, than with the FV mesh in MAGMA. On the FE mesh for CalculiX, two specific regions were constructed: (1) five layers of linear hexahedral elements up to a depth of 200 μm into the die surface layer, due to expected high temperature gradients, (2) linear tetrahedral mesh on the remaining region of the insert. 4. Results 4.1. Temperature computations Four temperature profiles of the whole die casting cycle at different positions of the analysed insert are shown in Fig. 4: (1) sample SU1, position P5, (2) sample SL1, position P1, (3) sample SL1, position P2, (4) sample SL1, position P3. For the temperature profile calculation, the procedure described in the chapter 3 which combines the use of the MAGMA and CalculiX software was applied. The presented temperature profiles are observed in the surface layer at a depth of 30 μm below the die surface. Temperatures, heating rates and cooling rates during a die casting cycle at positions P1, P2 and P3 of sample SL1 are summarised in Table 2. As expected, minimum and maximum temperatures are achieved during spraying and solidification phases, respectively. Maximum heating rates are achieved during die filling phase, and maximum cooling rates are achieved during die spraying phase. The highest temperature gradients are present at position P1 of sample SL1. At position P5 of sample SU1 the temperature is almost constant during the whole
Fig. 4. Comparison of temperature profiles of a whole die casting cycle 30 μm below the die surface at different positions of the analysed insert. 174
Engineering Failure Analysis 95 (2019) 171–180
R. Markežič et al.
Table 2 Temperatures, heating rates and cooling rates during a die casting cycle at positions P1, P2 and P3 of sample SL1, 30 μm below the die surface.
Maximum temperature (°C) Minimum temperature (°C) Maximum heating rate (°C/s) Maximum cooling rate (°C/s)
SL1 P1
SL1 P2
SL1 P3
515 75 16,000 780
390 115 11,500 450
340 125 11,600 390
die casting cycle. It fluctuates 5 °C to 10 °C around 200 °C. 4.2. Hardness measurements Four hardness profiles of the surface layer at different positions of the analysed insert are shown in Fig. 5: (1) mean hardness profile of all positions on sample SU1, designated as reference hardness profile, (2) sample SL1, position P1, (3) sample SL1, position P2, (4) sample SL1, position P3. Maximum hardnesses in the surface layer at positions P1, P2 and P3 of sample SL1 are summarised in Table 3. For hardness indentations up to a depth of 30 μm into the surface layer and for hardness indentations at depths greater than 30 μm into the surface layer, 25 gf and 50 gf loads were used, respectively. The hardness of the base material at depths greater than 150 μm from the die surface is approximately 600 HV. At all analysed positions, except at position P1 of sample SL1, an increase in hardness in the surface layer was identified due to nitridation. From a comparison between the reference hardness profile and the hardness profile at position P3 of sample SL1 we can conclude that both profiles are very similar. Maximum measured hardness on sample SU1 is between 860 HV and 970 HV. Maximum measured hardness at position P3 of sample SL1 is in the range of maximum measured hardnesses on sample SU1. At position P2 of sample SL1 the maximum measured hardness is lower than the maximum measured hardness at position P1 of sample SL1. The hardness profile at position P1 of sample SL1 does not have a specific increase of hardness in the surface layer. At this position hardness is approximately constant at all depths below the surface. 4.3. XRD analysis In Fig. 6, three XRD patterns of different samples are shown: (1) sample SU2 (sample-unloaded 2), (2) sample SL2 (sample-loaded 2), (3) sample SB2 (sample-base material 2). The XRD pattern of sample SU2 shows that besides martensite, iron nitrides (FeN, Fe2N, Fe3N and Fe4N), cementite (Fe3C) and oxide (Fe3O4) are present. In comparison with SU2 the sample SL2 does not contain iron nitrides and on the other hand, contains more Fe3C and Fe3O4. The XRD pattern of sample SB2, which represents the base material, indicates that in addition to martensite it also contains vanadium carbides (VC). 4.4. Light microscopy and SEM observations Light microscopy and backscattered electron (BE) images of the base material at a depth of 2 mm below the active surface are presented in Fig. 7a and Fig. 7b, respectively. The microstructure of the base material is mostly composed of tempered martensite with precipitated VC and some borides (designated with B). Tempered martensite is in the backscattered electron image, Fig. 7b,
Fig. 5. Comparison of hardness profiles in the surface layer at different positions of the analysed insert. 175
Engineering Failure Analysis 95 (2019) 171–180
R. Markežič et al.
Table 3 Maximum hardnesses in the surface layer at positions P1, P2 and P3 of sample SL1.
Maximum hardness (HV)
SL1 P1
SL1 P2
SL1 P3
600
815
960
Fig. 6. XRD patterns of samples SB2, SU2 and SL2.
Fig. 7. Microstructure images of SL1 sample at depth 2 mm bellow the active surface: (a) light microscopy image after etching with Vilella's reagent for 5 s, (b) BE image.
visible as dark and bright grey areas of irregular shapes, which represent differently oriented martensite laths. VC are seen as small dark particles while borides are seen as larger bright particles. In Figs. 8a-8d, light microscopy images of surface layers at four different positions of the analysed insert are presented: (a) sample SU1, position P5, (b) sample SL1, position P1, (c) sample SL1, position P2, (d) sample SL1, position P3. In Fig. 8a, c, d a dark area just below the surface of approximate thickness between 100 μm and 150 μm is seen, marked as NL. This area represents the diffusion 176
Engineering Failure Analysis 95 (2019) 171–180
R. Markežič et al.
Fig. 8. Comparison of diffusion layer of surface nitridation at different positions of SL1 and SU1 samples after etching with Vilella's reagent for 2 s: (a) SU1, P5, (b) SL1, P1, (c) SL1, P2, (d) SL1, P3.
layer of the surface nitridation. Fig. 8b indicates that at position P1 of sample SL1 no layer of nitridation is present. In Fig. 9, the BE image of surface layer at position P4 of sample SL1 is shown. From analysis of the image we can notice that in the surface layer of sample SL1, small bright particles of irregular shape are present. Mycroanalysis have shown that these are particles of Fe3C phase. XRD analysis confirmed that Fe3 C phase was already present after the initial heat treatment at the beginning of inserts operational life (sample SU2), but its content increased during the die casting process due to thermal loading of the insert (sample SL2). A light microscopy image of a surface crack, identified at position P4 of sample SL1, is shown in Fig. 10. The length of the crack is 60 μm. Two layers are present on the surface. A very thin layer, marked as OL (oxide layer), was observed on the walls of the crack which propagates also out of the crack. A thicker layer, marked as SL (soldered layer), was observed on the die surface outside of the crack area. Microanalysis reveals that the thin layer is mainly composed of iron and oxygen while the thicker layer is mostly composed of aluminium with small amounts of oxygen and iron. Small bright particles of irregular shape are similar to those seen in Fig. 9 and represent precipitates of Fe3C.
Fig. 9. BE image of surface layer at position P4 of sample SL1. 177
Engineering Failure Analysis 95 (2019) 171–180
R. Markežič et al.
Fig. 10. Light microscopy image of a surface crack on sample SL1 at position P4 after etching with Vilella's reagent for 5 s.
5. Discussion Many differences can be observed from the comparison between experimental results of samples that represent the initial state at the beginning of inserts operational life (samples SU1 and SU2), and samples that represent the loaded part of the insert after 170,000 casting cycles (samples SL1 and SL2). From the comparison of calculated temperature profiles in Fig. 4 and measured hardness profiles in the surface layer on Fig. 5, an increase in surface hardness drop with the increasing of maximum temperature in the surface layer is detected. The reason for the surface hardness drop is in the occurrence of three wear and damage mechanisms: (1) erosion, (2) corrosion, (3) thermal softening and microstructural changes of the die material. Erosion causes the removal of the thin hard nitrided surface layer, which is seen as surface hardness drop. Erosion rate is temperature dependent and increases with increasing die surface temperature [15–17]. Similar to erosion, corrosion also causes loss of material from the die surface, which appears as surface hardness drop. Corrosion rate is also temperature dependent and increases with increasing die surface temperature [14]. The appearance of erosion and corrosion was confirmed by comparison of light microscopy observations in Figs. 8a-8d. In Fig. 8a, c and d a dark surface layer is seen, which represents the diffusion layer of nitridation. In Fig. 8b, which represents the light microscopy image of the surface layer at position P1 of sample SL1, where the maximum surface hardness drop was observed, the dark diffusion layer of nitridation is not present anymore. From analysis of XRD results we can find out that no iron nitride phases were detected on sample SL2, unlike on sample SU2, which additionally confirms the occurrence of erosion and corrosion. Possible reasons for surface hardness drop are also thermal softening and microstructural changes in die material as a result of thermal fatigue [26,29,30,34,35]. Thermal softening is temperature dependent and is accelerated by increasing maximum temperature of thermal cycling [25,27,28]. Due to high temperatures, microstructural changes can occur. XRD analysis indicates an increase of Fe3C content in sample SL2 compared with sample SU2. Precipitation and growth of Fe3C was additionally confirmed with microanalysis and light microscopy analysis, as shown in Figs. 9 and 10, respectively. The precipitation of Fe3C is a consequence of martensite decomposition at high temperatures. Hardness drop and removal of diffusion layer of nitridation are caused by the simultaneous occurrence of erosion, corrosion and thermal softening. High surface temperatures accelerated the occurrence of erosion, corrosion, thermal softening and consequently hardness drop at certain positions. In Fig. 10 a surface crack at position P4 of sample SL1 is shown. The formation of the surface crack is a consequence of high tensile stresses and strains that form in the surface layer during the die preparation phase. High stresses and strains are promoted by high thermal gradients due to fast heating and cooling of the material in the surface layer. Another factor that affects stresses and strains in the surface layer is the geometry of the die. Cracks more often form on small radiuses or sharp changes in geometry of the die. Crack formation is also affected by microstructural changes and thermal softening of the material. Precipitation of Fe3C, which was observed near the crack area as shown in Figs. 9 and 10, affects microstructure and mechanical properties of the material and consequently crack formation and propagation. With the use of light microscopy analysis it has been observed that crack growth stopped at the interface between diffusion zone of nitridation and substrate material, like also observed by Pellizzari et al. [36]. The thin layer on the surface, inside and outside the crack area represents the surface oxide. In some studies it was found out that surface oxidation increases crack growth [29–31]. An increase in the amount of surface iron oxide Fe3O4 on the sample SL2, relative to sample SU2, was confirmed also with XRD analysis. The thicker layer on the surface, outside the crack area, is a consequence of mechanical and metallurgical soldering. From the comparison of experimental and numerical results we can observe an increase in the occurrence of wear and damage mechanisms with the increasing of temperature in the surface layer. Thermal fatigue and thermal cracking are further increased by high heating and cooling rates in the surface layer, which cause high thermal gradients, stresses and strains. The negative effects of wear and damage mechanisms would be reduced by lowering maximum temperatures and temperature fluctuations during a die casting process in the die surface layer. A simultaneous occurrence of different wear and damage mechanisms at specific surface areas was also present, which prevents the study of effects and characteristics of a single mechanism during a die casting process. For this purpose, many laboratory tests have been developed, which allow isolation and analysis of a single wear and damage mechanism at a time [1,9,27,28,35,36].
178
Engineering Failure Analysis 95 (2019) 171–180
R. Markežič et al.
6. Conclusions An experimental analysis of an insert of a die casting tool, used for casting of aluminium alloy AlSi9Cu3Fe, was performed. XRD, light microscopy, SEM and hardness measurement methods were used during experimental analysis. A numerical computation of temperature fields on the insert for a whole casting cycle was performed with a specially developed procedure, which combines the use of commercial software MAGMA and open source software CalculiX. The following conclusions can be drawn:
• Four wear and damage mechanisms were identified on the analysed insert: (1) soldering, (2) corrosion, (3) erosion, (4) thermal fatigue and cracking. • An increase in surface hardness drop rate with the increase of maximum temperature in the surface layer during a die casting cycle • • • • •
was detected. Reasons for surface hardness drop are in the occurrence of erosion, corrosion, thermal softening and microstructural changes of the die material. Removal of the diffusion layer of nitridation was observed at surface areas with maximum surface layer temperatures. XRD analysis shows that surfaces not being exposed to the melt contains nitrides while those being exposed to the melt during the casting process do not. This result confirms the occurrence of erosion and corrosion. XRD analysis shows an increase in the content of Fe3C at surfaces that were in contact with the melt, compared with those that were not. This confirms the occurrence of microstructural changes in the die material during die casting process. A surface crack was identified with use of light microscopy analysis at the surface in contact with the melt. Crack growth stopped at the interface between diffusion zone of nitridation and substrate material. A thin layer of oxide inside and outside the crack and a thicker soldering layer outside the crack were observed. An increase in the occurrence of wear and damage mechanisms with increasing temperature in the surface layer was detected. The negative effects of wear and damage mechanisms would be reduced by lowering maximum temperatures and temperature fluctuations during a die casting process.
From the results of this study, an increase in the occurrence of wear and damage mechanisms with increasing temperature in the surface layer was detected. Results show also an increase in surface hardness drop rate with increasing temperature in the surface layer. Adequate die surface hardness is one of the main die characteristics that helps to prevent initiation and propagation of die wear and damage mechanisms. By knowing the change of die surface hardness during the die casting process, the accuracy of die life predictions would be improved and new methods for die lifespan prolongation could appear. Further research will be focused on determining the effect of thermal loading conditions during the die casting process on die surface hardness and prediction of change in surface hardness with increasing number of casting cycles. Acknowledgements Funding: The authors acknowledge the financial support from the Slovenian Research Agency (research core funding No. P2–0270). Declarations of interest None. References [1] S. Gopal, A. Lakare, R. Shivpuri, Evaluation of thin film coatings for erosive-corrosive wear prevention in die casting dies, Surf. Eng. 15 (4) (1999) 297–300. [2] A.J. Norwood, P.M. Dickens, R.C. Soar, R.A. Harris, Surface temperature of tools during the high-pressure die casting of aluminium, Proc. Inst. Mech. Eng. B J. Eng. Manuf. 221 (12) (2007) 1659–1664. [3] M. Dargusch, A. Hamasaiid, G. Dour, T. Loulou, C. Davidson, D.H. StJohn, The accurate determination of heat transfer coefficient and its evolution with time during high pressure die casting of Al9%Si-3%Cu and Mg-9%Al-1%Zn Alloys, Adv. Eng. Mater. 9 (11) (2007) 995–999. [4] A. Long, D. Thornhill, C. Armstrong, D. Watson, Determination of the heat transfer coefficient at the metal–die interface for high pressure die cast AlSi9Cu3Fe, Appl. Therm. Eng. 31 (17) (2011) 3996–4006. [5] K. Domkin, J.H. Hattel, J. Thorborg, Modeling of high temperature- and diffusion-controlled die soldering in aluminum high pressure die casting, J. Mater. Process. Technol. 209 (8) (2009) 4051–4061. [6] S. Gulizia, M.Z. Jahedi, E.D. Doyle, Performance evaluation of PVD coatings for high pressure die casting, Surf. Coat. Technol. 140 (3) (2001) 200–205. [7] Z.W. Chen, Formation and progression of die soldering during high pressure die casting, Mater. Sci. Eng. A 397 (1–2) (2005) 356–369. [8] H. Zhu, J. Guo, J. Jia, Experimental study and theoretical analysis on die soldering in aluminum die casting, J. Mater. Process. Technol. 123 (2) (2002) 229–235. [9] M. Sundqvist, S. Hogmark, Effects of liquid aluminium on hot-work tool steel, Tribol. Int. 26 (2) (1993) 129–134. [10] P.A. Hogan, Die Solder Prediction and Reduction, PhD thesis Worcester Polytechnic Institute, 2008. [11] R. Shivpuri, M. Yu, K. Venkatesan, Y.L. Chu, A study of erosion in die casting dies by a multiple pin accelerated erosion test, J. Mater. Eng. Perform. 4 (2) (1995) 145–153. [12] V. Joshi, A. Srivastava, R. Shivpuri, Intermetallic formation and its relation to interface mass loss and tribology in die casting dies, Wear 256 (11−12) (2004) 1232–1235. [13] Q. Han, S. Viswanathan, Analysis of the mechanism of die soldering in aluminum die casting, Metall. Mater. Trans. A 34 (1) (2003) 139–146. [14] M. Yu, R. Shivpuri, R.A. Rapp, Effects of molten aluminum on H13 dies and coatings, J. Mater. Eng. Perform. 4 (2) (1995) 175–181. [15] Z.W. Chen, M.Z. Jahedi, Die erosion and its effect on soldering formation in high pressure die casting of aluminium alloys, Mater. Des. 20 (6) (1999) 303–309. [16] K. Venkatesan, R. Shivpuri, Experimental and numerical investigation of the effect of process parameters on the erosive wear of die casting dies, J. Mater. Eng. Perform. 4 (2) (1995) 166–174.
179
Engineering Failure Analysis 95 (2019) 171–180
R. Markežič et al.
[17] A. Mohammed, M.B. Marshall, R. Lewis, Development of a method for assessing erosive wear damage on dies used in aluminium casting, Wear 332–333 (2015) 1215–1224. [18] L.F. Hou, Y.H. Wei, Y.G. Li, B.S. Liu, H.Y. Du, C.L. Guo, Erosion process analysis of die-casting inserts for magnesium alloy components, Eng. Fail. Anal. 33 (2013) 457–564. [19] D.W.C. Baker, K.H. Jolliffe, D. Pearson, The resistance of materials to impact erosion damage, Philos. Trans. R. Soc. A Math. Phys. Eng. Sci. 260 (1110) (1966) 193–203. [20] A. Persson, S. Hogmark, J. Bergström, Temperature profiles and conditions for thermal fatigue cracking in brass die casting dies, J. Mater. Process. Technol. 152 (2) (2004) 228–236. [21] C. Rosbrook, Analysis of Thermal Fatigue and Heat Checking in Die-Casting Dies: A Finite Element Approach, PhD thesis Ohio State University, 1992. [22] F. Medjedoub, G. Dour, S. Le Roux, P. Lamesle, M. Salem, P. Hairy, F. Rézaï-Aria, Experimental conditions and environment effects on thermal fatigue damage accumulation and life of die-casting steel X38CrMoV5 (AISI H11), Int. J. Microstruct. Mater. Propert. 3 (2–3) (2008). [24] D. Klobčar, J. Tušek, Thermal stresses in aluminium alloy die casting dies, Comput. Mater. Sci. 43 (4) (2008) 1147–1154. [25] J. Sjöström, J. Bergström, Thermal fatigue testing of chromium martensitic hot-work tool steel after different austenitizing treatments, J. Mater. Process. Technol. 153 (2004) 1089–1096. [26] D. Klobčar, J. Tušek, B. Taljat, Thermal fatigue of materials for die-casting tooling, Mater. Sci. Eng. A 472 (1) (2008) 198–207. [27] A. Persson, Strain-based approach to crack growth and thermal fatigue life of hot work tool steels, Scand. J. Metall. 33 (2004) 53–64. [28] A. Persson, S. Hogmark, J. Bergström, Simulation and evaluation of thermal fatigue cracking of hot work tool steels, Int. J. Fatigue 26 (10) (2004) 1095–1107. [29] G. Roberts, G. Krauss, R. Kennedy, Tool Steels, 5th ed., ASM, Metals Park, 1998, p. 1998. [30] J. Sjöström and J. Bergström. Evaluation of the cyclic behaviour during high temperature fatigue of hot work tool steels. In Proceedings of 6th International Tooling Conference, pages 721–736. [31] D. Klobčar, L. Kosec, B. Kosec, J. Tušek, Thermo fatigue cracking of die casting dies, Eng. Fail. Anal. 20 (2012) 43–53. [32] M. Muhič, F. Kosel, A. Pukšič, D. Klobčar, A new approach to monitoring thermal fatigue cracks in die casting moulds, Int. J. Mat. Res. 102 (2011) 69–75. [33] D. Mellouli, N. Haddar, A. Köster, H.F. Ayedi, Hardness effect on thermal fatigue damage of hot-working tool steel, Eng. Fail. Anal. 45 (2014) 85–95. [34] M. Muhič, J. Tušek, F. Kosel, D. Klobčar, M. Pleterski, Thermal fatigue cracking of die-casting dies, Meta 49 (1) (2010) 9–12. [35] M. Fazarinc, T. Muhič, G. Kugler, M. Terčelj, Thermal fatigue properties of differently constructed functionally graded materials aimed for refurbishing of pressure-die-casting dies, Eng. Fail. Anal. 25 (2012) 238–249. [36] M. Pellizzari, A. Molinari, G. Straffelini, Thermal fatigue resistance of gas and plasma nitrided 41CrAlMo7 steel, Mater. Sci. Eng. A 352 (1–2) (2003) 186–194. [37] Bohler-Uddelholm, Dievar Data Sheets ®. Technical report, Uddeholm, (2013).
180
Contents lists available at ScienceDirect
Engineering Failure Analysis journal homepage: www.elsevier.com/locate/engfailanal
Experimental and numerical analysis of failures on a die insert for high pressure die casting R. Markežiča, I. Nagličb, N. Molea, R. Šturma, a b
T
⁎
Faculty of Mechanical Engineering, University of Ljubljana, Aškerčeva 6, 1000 Ljubljana, Slovenia Faculty of Natural Sciences and Engineering, University of Ljubljana, Aškerčeva 12, 1000 Ljubljana, Slovenia
A R T IC LE I N F O
ABS TRA CT
Keywords: Die failures Nitriding Finite element analysis Microanalysis X-ray analysis
A comprehensive study of failures on a used insert of a die for high pressure die casting of aluminium alloy AlSi9Cu3Fe was conducted. The analysed insert was made of hot work tool steel Dievar, heat treated and plasma nitrided according to specifications. Before the start of analysis, the insert was subjected to 170,000 die casting cycles. X-ray diffraction, light microscopy, scanning electron microscopy and hardness measurement methods were used for experimental analysis of soldering, corrosion, erosion and thermal fatigue failure modes on the insert. A special procedure, which combines use of commercial software MAGMA5.3 and open source software CalculiX, was developed for numerical calculations of temperature fields on the insert. Experimental and numerical results showed strong dependence between die surface temperatures and die failure modes. Hardness drop, nitride diffusion layer removal and microstructural changes were observed at more thermally affected areas. Surface crack with a thin oxide and thicker soldered layer was identified and analysed.
1. Introduction High pressure die casting (HPDC), designated in this article as die casting, is a technology that enables fast and high quality production of geometrically complex parts with high mechanical and dimensional requirements. Due to high absolute values and rapid fluctuations of temperature, pressure and melt velocity, die casting dies are exposed to cyclic thermo-mechanical loading, which gradually leads to the occurrence of wear and damage. Typical maximum melt velocities during die filling are between 30 m/s and 100 m/s [1]. The additional pressure applied during the solidification phase is of magnitude from 50 MPa to 80 MPa [1,2]. One of the most important parameters in a die casting process is the temperature of the die. A lot of studies have been conducted to monitor the temperature at the die surface during a real die casting cycle [2–4]. Norwood et al. [2] performed temperature measurements during casting of aluminium alloy Al8Si3Cu with use of thermocouples and temperature-sensitive paints. The maximum and minimum measured temperatures were in the range from 400 °C to 450 °C and from 150 °C to 200 °C, respectively. Dargusch et al. [3] and Long et al. [4] measured temperatures in the die surface layer during the casting process of aluminium alloy AlSi9Cu3Fe. From the temperature measurements the heat transfer coefficient was determined and then used in a numerical simulation for surface temperature calculation. The minimum and maximum calculated temperatures were 240 °C and 500 °C, respectively. Die wear and damage mechanisms are generally divided in four groups: (1) soldering, (2) corrosion, (3) erosion, (4) thermal fatigue and cracking. ⁎
Corresponding author. E-mail addresses: [email protected] (R. Markežič), [email protected] (R. Šturm).
https://doi.org/10.1016/j.engfailanal.2018.09.010 Received 22 May 2018; Received in revised form 30 August 2018; Accepted 13 September 2018 Available online 15 September 2018 1350-6307/ © 2018 Elsevier Ltd. All rights reserved.
Engineering Failure Analysis 95 (2019) 171–180
R. Markežič et al.
Soldering (or die sticking) is seen as the adhesion of cast material to the die surface during the die filling and solidification phases. Soldering can be subdivided into two categories [5–8]: (1) metallurgical soldering, characterised by a high number of loading cycles and high temperature of the die surfaces [5,9–14] (2) mechanical soldering, characterised by a low number of loading cycles and high melt pressures [5]. A soldered layer is usually the product of the simultaneous action of metallurgical and mechanical soldering. Corrosion of dies in die casting is characterised by loss of die material from the die surface and it is a metallurgical process [14]. There is often a simultaneous occurrence of corrosion and soldering on a particular region of a die surface [14]. Erosion is also characterised by loss of die material from the die surface, but it is a mechanical process [15]. An increase in the occurrence of erosion with an increase of die surface temperature was identified in several studies, [16–18]. Baker et al. [19] observed the appearance of cracks on previously eroded die surfaces. The main cause for thermal cracking of die casting dies are high temperature and pressure gradients [20–22,24]. The most influential parameters on thermal cracking are: (a) maximum heating temperature and heating/cooling rate, (b) oxidation, (c) hardness and microstructure of the die material. Research has shown, that crack initiation and propagation are increased by higher heating temperatures and higher heating/cooling rates [22,24–28]. In some studies it has been observed, that crack growth is increased by surface oxidation [29–31]. From thermal fatigue tests it has been found out, that crack density and crack depth decrease with increasing surface hardness [26–28,32,33]. Thermal fatigue also affects the microstructure of the die material. It causes thermal softening of the die surface, which is seen as surface hardness drop [26,29,30,34,35]. Thermal softening and hardness drop are accelerated by higher heating temperatures [25,27,28,31]. To reduce the occurrence of damage and wear mechanisms on die casting dies, it is important to choose an appropriate material, heat treatment and optionally surface treatment of the die [29]. One of the most often used surface treatments on die casting dies is nitridation. Joshi et al. [12] found out, that nitridation decreases the rate of metallurgical soldering on die surfaces. However, soldering on nitrided surfaces is present due to mechanical soldering after the formation of thermal cracks and erosion of the nitrided layer [6]. Persson [27] discovered that nitrided surfaces have better resistance to thermo-mechanical fatigue than untreated surfaces do. An increase in the thermal stability of the microstructure of nitrided surfaces has also been observed. The final crack depth on a nitrided surface is dependent on the nitridation depth, because crack growth stops at the interface between diffusion zone and substrate material [36]. In this paper, a comprehensive analysis of a used insert of a die casting tool, made of hot work tool steel Dievar [37], is presented. The die casting tool was used for casting of aluminium alloy AlSi9Cu3Fe. The focus is on determining the influence of temperature on the occurrence of individual wear and damage mechanisms. 2. Experimental analysis 2.1. Materials and samples preparation A three-dimensional model of the analysed insert is presented in Fig. 1a. The chemical composition of hot work tool steel Dievar, used for production of the analysed insert, is presented in Table 1. Before use, the insert was heat treated and surface treated according to specifications. Heat treatment was composed of heating the insert to the austenitising temperature of 1020 °C for 30 min, followed by quenching in warm oil and final double tempering for 2 h at 550 °C. After heat treatment, the insert was additionally surface plasma nitrided for 10 h at 480 °C to reach the final nitridation depth of 150 μm and final surface hardness of 1100 HV. The die casting tool with the analysed insert was used for die casting of aluminium alloy AlSi9Cu3Fe. The die casting cycle consisted of spraying the insert for 2 s, blowing the insert for 5 s, dosing the melt at a temperature of 670 °C, filling of the die, and solidification of the casting for 10 s. The duration of the entire die casting cycle was 62 s. The insert was exposed to 170,000 die casting cycles and then removed from the die casting tool for analysis. During die casting process the yellow coloured surface was in contact with melt and hence thermo-mechanically loaded. This surface will be in this paper designated as the active surface. During die casting process
Fig. 1. Insert presentation and samples preparation: (a) three-dimensional model of the analysed insert of dimensions Ø35x44.73 mm, (b) samples cutting scheme. 172
Engineering Failure Analysis 95 (2019) 171–180
R. Markežič et al.
Table 1 Dievar chemical composition [37]. Elements
C
Si
Mn
Cr
Mo
V
Wt%
0.35
0.2
0.5
5.0
2.3
0.6
the red coloured surface was in contact with other elements of the die casting tool. Two samples were cut out of the insert for light microscopy and scanning electron microscopy. One sample was cut in such a way that it contains a part of the active surface and represents the loaded part of the insert after 170,000 casting cycles, as shown in Fig. 1b. This sample will be designated as SL1 (sample-loaded 1). Another sample was cut from the opposite side of the insert and does not contain a part of the active surface, as shown in Fig. 1b. It represents the initial state of the insert at the beginning of it's operational life. This sample will be designated as SU1 (sample-unloaded 1). Samples for light microscopy and scanning electron microscopy were first ground and polished with diamond paste. Final polishing was performed with 0.05 μm silica. Samples for light microscopy analyses were additionally etched with Vilellas reagent. After finishing the analysis of samples SL1 and SU1, three additional samples were cut out from the insert, as shown in Fig. 1b: (1) sample representing the base material, designated as SB2 (sample-base material 2), (2) sample representing the loaded part of the insert, designated as SL2 (sample-loaded 2), (3) sample representing the initial state at the beginning of inserts operational life, designated as SU2 (sample-unloaded 2). Samples SB2, SL2 and SU2 were used for X-ray diffraction analysis. 2.2. Methods Samples were experimentally analysed with four different methods: X-ray diffraction (XRD), light microscopy, scanning electron microscopy (SEM) and hardness measurements. A scanning electron microscope JEOL JSM-7600F equipped with energy-dispersive Xray spectrometer X-Max 20 SDD-EDS detector was used for characterisation of the microstructure. Images and analyses were made at an acceleration voltage of 10 kV. X-ray diffraction (XRD) patterns were acquired with use of X-ray diffractometer PANalytical X'Pert PRO. Non-monochromated X-rays (0.15418 nm) produced by the Cu target tube were used in these experiments. Hardness measurements were conducted on Ernst Leitz Orthoplan equipment with additional microhardness pneumatic indenter with use of the Vickers method. Analyses were made on different positions of surface layers of samples SL1 and SU1, depicted in Fig. 2a and Fig. 2b, respectively. 3. Numerical analysis The goal of the numerical analysis was to obtain time-temperature profiles of the whole die casting cycle in the die surface layer. For this purpose, a special procedure was applied, which includes the use of the commercial software MAGMA (release 5.3) and the open source software CalculiX. The reason for using this procedure is in the computation methods used by each software. MAGMA is a commercial software which allows calculation of different influential parameters through the whole die casting cycle. The temperature computations in MAGMA are based on the Cartesian cut cell method, which is a variation of the finite volume method (FVM). CalculiX is an open source software which also allows temperature computations, but in this case the calculations are based on the finite element method (FEM). The steps of the applied procedure for temperature computations were the following: (1) casting material and die temperature field computation with the use of MAGMA, (2) temperature field computation of the analysed die insert with the use of CalculiX and applied film boundary conditions by known castings and other tool elements temperatures from step 1. For the die preparation phase, the temperature-dependent film coefficients for spraying, blowing and contact of die with air from the MAGMA library were applied. For the die filling and solidification phase, the casting-die interface temperature-dependent film coefficient for casting of aluminium alloy AlSi9Cu3Fe, reported by Long et al. [4], was applied. For the boundary conditions at the interface between the analysed die insert and other die elements, a temperature independent film coefficient of value 2000 W/m2 K
Fig. 2. Presentation of the analysed positions on samples: (a) sample SL1, (b) sample SU1. 173
Engineering Failure Analysis 95 (2019) 171–180
R. Markežič et al.
Fig. 3. Comparison of cross-sections of meshes used in simulations with: (a) MAGMA (FV mesh), (b) CalculiX (FE mesh).
was used. In real production the die is preheated to a certain temperature with the use of the internal tempering system before the start of the die casting process. This condition was taken into account in computations with MAGMA by setting the initial temperature condition of the die to 150 °C. After the start of the die casting process in real production, it takes several casting cycles to achieve stable cyclic temperature conditions on the die. During computations in MAGMA, five heating cycles and three production cycles were simulated to reach stable cyclic temperature conditions on the die. During heating cycle simulations no die filling is computed, so it is assumed that the whole die is filled at the same moment. While during production cycle simulations, the entire die filling is computed step by step. The temperature fields from the last production cycle were used for the definition of boundary conditions in simulations with CalculiX. In Fig. 3a and b cross-sections of meshes used in the simulations are depicted. From comparison of Fig. 3a and Fig. 3b we can conclude, that with the FE mesh in CalculiX, geometrical features, like radiuses, are better described, than with the FV mesh in MAGMA. On the FE mesh for CalculiX, two specific regions were constructed: (1) five layers of linear hexahedral elements up to a depth of 200 μm into the die surface layer, due to expected high temperature gradients, (2) linear tetrahedral mesh on the remaining region of the insert. 4. Results 4.1. Temperature computations Four temperature profiles of the whole die casting cycle at different positions of the analysed insert are shown in Fig. 4: (1) sample SU1, position P5, (2) sample SL1, position P1, (3) sample SL1, position P2, (4) sample SL1, position P3. For the temperature profile calculation, the procedure described in the chapter 3 which combines the use of the MAGMA and CalculiX software was applied. The presented temperature profiles are observed in the surface layer at a depth of 30 μm below the die surface. Temperatures, heating rates and cooling rates during a die casting cycle at positions P1, P2 and P3 of sample SL1 are summarised in Table 2. As expected, minimum and maximum temperatures are achieved during spraying and solidification phases, respectively. Maximum heating rates are achieved during die filling phase, and maximum cooling rates are achieved during die spraying phase. The highest temperature gradients are present at position P1 of sample SL1. At position P5 of sample SU1 the temperature is almost constant during the whole
Fig. 4. Comparison of temperature profiles of a whole die casting cycle 30 μm below the die surface at different positions of the analysed insert. 174
Engineering Failure Analysis 95 (2019) 171–180
R. Markežič et al.
Table 2 Temperatures, heating rates and cooling rates during a die casting cycle at positions P1, P2 and P3 of sample SL1, 30 μm below the die surface.
Maximum temperature (°C) Minimum temperature (°C) Maximum heating rate (°C/s) Maximum cooling rate (°C/s)
SL1 P1
SL1 P2
SL1 P3
515 75 16,000 780
390 115 11,500 450
340 125 11,600 390
die casting cycle. It fluctuates 5 °C to 10 °C around 200 °C. 4.2. Hardness measurements Four hardness profiles of the surface layer at different positions of the analysed insert are shown in Fig. 5: (1) mean hardness profile of all positions on sample SU1, designated as reference hardness profile, (2) sample SL1, position P1, (3) sample SL1, position P2, (4) sample SL1, position P3. Maximum hardnesses in the surface layer at positions P1, P2 and P3 of sample SL1 are summarised in Table 3. For hardness indentations up to a depth of 30 μm into the surface layer and for hardness indentations at depths greater than 30 μm into the surface layer, 25 gf and 50 gf loads were used, respectively. The hardness of the base material at depths greater than 150 μm from the die surface is approximately 600 HV. At all analysed positions, except at position P1 of sample SL1, an increase in hardness in the surface layer was identified due to nitridation. From a comparison between the reference hardness profile and the hardness profile at position P3 of sample SL1 we can conclude that both profiles are very similar. Maximum measured hardness on sample SU1 is between 860 HV and 970 HV. Maximum measured hardness at position P3 of sample SL1 is in the range of maximum measured hardnesses on sample SU1. At position P2 of sample SL1 the maximum measured hardness is lower than the maximum measured hardness at position P1 of sample SL1. The hardness profile at position P1 of sample SL1 does not have a specific increase of hardness in the surface layer. At this position hardness is approximately constant at all depths below the surface. 4.3. XRD analysis In Fig. 6, three XRD patterns of different samples are shown: (1) sample SU2 (sample-unloaded 2), (2) sample SL2 (sample-loaded 2), (3) sample SB2 (sample-base material 2). The XRD pattern of sample SU2 shows that besides martensite, iron nitrides (FeN, Fe2N, Fe3N and Fe4N), cementite (Fe3C) and oxide (Fe3O4) are present. In comparison with SU2 the sample SL2 does not contain iron nitrides and on the other hand, contains more Fe3C and Fe3O4. The XRD pattern of sample SB2, which represents the base material, indicates that in addition to martensite it also contains vanadium carbides (VC). 4.4. Light microscopy and SEM observations Light microscopy and backscattered electron (BE) images of the base material at a depth of 2 mm below the active surface are presented in Fig. 7a and Fig. 7b, respectively. The microstructure of the base material is mostly composed of tempered martensite with precipitated VC and some borides (designated with B). Tempered martensite is in the backscattered electron image, Fig. 7b,
Fig. 5. Comparison of hardness profiles in the surface layer at different positions of the analysed insert. 175
Engineering Failure Analysis 95 (2019) 171–180
R. Markežič et al.
Table 3 Maximum hardnesses in the surface layer at positions P1, P2 and P3 of sample SL1.
Maximum hardness (HV)
SL1 P1
SL1 P2
SL1 P3
600
815
960
Fig. 6. XRD patterns of samples SB2, SU2 and SL2.
Fig. 7. Microstructure images of SL1 sample at depth 2 mm bellow the active surface: (a) light microscopy image after etching with Vilella's reagent for 5 s, (b) BE image.
visible as dark and bright grey areas of irregular shapes, which represent differently oriented martensite laths. VC are seen as small dark particles while borides are seen as larger bright particles. In Figs. 8a-8d, light microscopy images of surface layers at four different positions of the analysed insert are presented: (a) sample SU1, position P5, (b) sample SL1, position P1, (c) sample SL1, position P2, (d) sample SL1, position P3. In Fig. 8a, c, d a dark area just below the surface of approximate thickness between 100 μm and 150 μm is seen, marked as NL. This area represents the diffusion 176
Engineering Failure Analysis 95 (2019) 171–180
R. Markežič et al.
Fig. 8. Comparison of diffusion layer of surface nitridation at different positions of SL1 and SU1 samples after etching with Vilella's reagent for 2 s: (a) SU1, P5, (b) SL1, P1, (c) SL1, P2, (d) SL1, P3.
layer of the surface nitridation. Fig. 8b indicates that at position P1 of sample SL1 no layer of nitridation is present. In Fig. 9, the BE image of surface layer at position P4 of sample SL1 is shown. From analysis of the image we can notice that in the surface layer of sample SL1, small bright particles of irregular shape are present. Mycroanalysis have shown that these are particles of Fe3C phase. XRD analysis confirmed that Fe3 C phase was already present after the initial heat treatment at the beginning of inserts operational life (sample SU2), but its content increased during the die casting process due to thermal loading of the insert (sample SL2). A light microscopy image of a surface crack, identified at position P4 of sample SL1, is shown in Fig. 10. The length of the crack is 60 μm. Two layers are present on the surface. A very thin layer, marked as OL (oxide layer), was observed on the walls of the crack which propagates also out of the crack. A thicker layer, marked as SL (soldered layer), was observed on the die surface outside of the crack area. Microanalysis reveals that the thin layer is mainly composed of iron and oxygen while the thicker layer is mostly composed of aluminium with small amounts of oxygen and iron. Small bright particles of irregular shape are similar to those seen in Fig. 9 and represent precipitates of Fe3C.
Fig. 9. BE image of surface layer at position P4 of sample SL1. 177
Engineering Failure Analysis 95 (2019) 171–180
R. Markežič et al.
Fig. 10. Light microscopy image of a surface crack on sample SL1 at position P4 after etching with Vilella's reagent for 5 s.
5. Discussion Many differences can be observed from the comparison between experimental results of samples that represent the initial state at the beginning of inserts operational life (samples SU1 and SU2), and samples that represent the loaded part of the insert after 170,000 casting cycles (samples SL1 and SL2). From the comparison of calculated temperature profiles in Fig. 4 and measured hardness profiles in the surface layer on Fig. 5, an increase in surface hardness drop with the increasing of maximum temperature in the surface layer is detected. The reason for the surface hardness drop is in the occurrence of three wear and damage mechanisms: (1) erosion, (2) corrosion, (3) thermal softening and microstructural changes of the die material. Erosion causes the removal of the thin hard nitrided surface layer, which is seen as surface hardness drop. Erosion rate is temperature dependent and increases with increasing die surface temperature [15–17]. Similar to erosion, corrosion also causes loss of material from the die surface, which appears as surface hardness drop. Corrosion rate is also temperature dependent and increases with increasing die surface temperature [14]. The appearance of erosion and corrosion was confirmed by comparison of light microscopy observations in Figs. 8a-8d. In Fig. 8a, c and d a dark surface layer is seen, which represents the diffusion layer of nitridation. In Fig. 8b, which represents the light microscopy image of the surface layer at position P1 of sample SL1, where the maximum surface hardness drop was observed, the dark diffusion layer of nitridation is not present anymore. From analysis of XRD results we can find out that no iron nitride phases were detected on sample SL2, unlike on sample SU2, which additionally confirms the occurrence of erosion and corrosion. Possible reasons for surface hardness drop are also thermal softening and microstructural changes in die material as a result of thermal fatigue [26,29,30,34,35]. Thermal softening is temperature dependent and is accelerated by increasing maximum temperature of thermal cycling [25,27,28]. Due to high temperatures, microstructural changes can occur. XRD analysis indicates an increase of Fe3C content in sample SL2 compared with sample SU2. Precipitation and growth of Fe3C was additionally confirmed with microanalysis and light microscopy analysis, as shown in Figs. 9 and 10, respectively. The precipitation of Fe3C is a consequence of martensite decomposition at high temperatures. Hardness drop and removal of diffusion layer of nitridation are caused by the simultaneous occurrence of erosion, corrosion and thermal softening. High surface temperatures accelerated the occurrence of erosion, corrosion, thermal softening and consequently hardness drop at certain positions. In Fig. 10 a surface crack at position P4 of sample SL1 is shown. The formation of the surface crack is a consequence of high tensile stresses and strains that form in the surface layer during the die preparation phase. High stresses and strains are promoted by high thermal gradients due to fast heating and cooling of the material in the surface layer. Another factor that affects stresses and strains in the surface layer is the geometry of the die. Cracks more often form on small radiuses or sharp changes in geometry of the die. Crack formation is also affected by microstructural changes and thermal softening of the material. Precipitation of Fe3C, which was observed near the crack area as shown in Figs. 9 and 10, affects microstructure and mechanical properties of the material and consequently crack formation and propagation. With the use of light microscopy analysis it has been observed that crack growth stopped at the interface between diffusion zone of nitridation and substrate material, like also observed by Pellizzari et al. [36]. The thin layer on the surface, inside and outside the crack area represents the surface oxide. In some studies it was found out that surface oxidation increases crack growth [29–31]. An increase in the amount of surface iron oxide Fe3O4 on the sample SL2, relative to sample SU2, was confirmed also with XRD analysis. The thicker layer on the surface, outside the crack area, is a consequence of mechanical and metallurgical soldering. From the comparison of experimental and numerical results we can observe an increase in the occurrence of wear and damage mechanisms with the increasing of temperature in the surface layer. Thermal fatigue and thermal cracking are further increased by high heating and cooling rates in the surface layer, which cause high thermal gradients, stresses and strains. The negative effects of wear and damage mechanisms would be reduced by lowering maximum temperatures and temperature fluctuations during a die casting process in the die surface layer. A simultaneous occurrence of different wear and damage mechanisms at specific surface areas was also present, which prevents the study of effects and characteristics of a single mechanism during a die casting process. For this purpose, many laboratory tests have been developed, which allow isolation and analysis of a single wear and damage mechanism at a time [1,9,27,28,35,36].
178
Engineering Failure Analysis 95 (2019) 171–180
R. Markežič et al.
6. Conclusions An experimental analysis of an insert of a die casting tool, used for casting of aluminium alloy AlSi9Cu3Fe, was performed. XRD, light microscopy, SEM and hardness measurement methods were used during experimental analysis. A numerical computation of temperature fields on the insert for a whole casting cycle was performed with a specially developed procedure, which combines the use of commercial software MAGMA and open source software CalculiX. The following conclusions can be drawn:
• Four wear and damage mechanisms were identified on the analysed insert: (1) soldering, (2) corrosion, (3) erosion, (4) thermal fatigue and cracking. • An increase in surface hardness drop rate with the increase of maximum temperature in the surface layer during a die casting cycle • • • • •
was detected. Reasons for surface hardness drop are in the occurrence of erosion, corrosion, thermal softening and microstructural changes of the die material. Removal of the diffusion layer of nitridation was observed at surface areas with maximum surface layer temperatures. XRD analysis shows that surfaces not being exposed to the melt contains nitrides while those being exposed to the melt during the casting process do not. This result confirms the occurrence of erosion and corrosion. XRD analysis shows an increase in the content of Fe3C at surfaces that were in contact with the melt, compared with those that were not. This confirms the occurrence of microstructural changes in the die material during die casting process. A surface crack was identified with use of light microscopy analysis at the surface in contact with the melt. Crack growth stopped at the interface between diffusion zone of nitridation and substrate material. A thin layer of oxide inside and outside the crack and a thicker soldering layer outside the crack were observed. An increase in the occurrence of wear and damage mechanisms with increasing temperature in the surface layer was detected. The negative effects of wear and damage mechanisms would be reduced by lowering maximum temperatures and temperature fluctuations during a die casting process.
From the results of this study, an increase in the occurrence of wear and damage mechanisms with increasing temperature in the surface layer was detected. Results show also an increase in surface hardness drop rate with increasing temperature in the surface layer. Adequate die surface hardness is one of the main die characteristics that helps to prevent initiation and propagation of die wear and damage mechanisms. By knowing the change of die surface hardness during the die casting process, the accuracy of die life predictions would be improved and new methods for die lifespan prolongation could appear. Further research will be focused on determining the effect of thermal loading conditions during the die casting process on die surface hardness and prediction of change in surface hardness with increasing number of casting cycles. Acknowledgements Funding: The authors acknowledge the financial support from the Slovenian Research Agency (research core funding No. P2–0270). Declarations of interest None. References [1] S. Gopal, A. Lakare, R. Shivpuri, Evaluation of thin film coatings for erosive-corrosive wear prevention in die casting dies, Surf. Eng. 15 (4) (1999) 297–300. [2] A.J. Norwood, P.M. Dickens, R.C. Soar, R.A. Harris, Surface temperature of tools during the high-pressure die casting of aluminium, Proc. Inst. Mech. Eng. B J. Eng. Manuf. 221 (12) (2007) 1659–1664. [3] M. Dargusch, A. Hamasaiid, G. Dour, T. Loulou, C. Davidson, D.H. StJohn, The accurate determination of heat transfer coefficient and its evolution with time during high pressure die casting of Al9%Si-3%Cu and Mg-9%Al-1%Zn Alloys, Adv. Eng. Mater. 9 (11) (2007) 995–999. [4] A. Long, D. Thornhill, C. Armstrong, D. Watson, Determination of the heat transfer coefficient at the metal–die interface for high pressure die cast AlSi9Cu3Fe, Appl. Therm. Eng. 31 (17) (2011) 3996–4006. [5] K. Domkin, J.H. Hattel, J. Thorborg, Modeling of high temperature- and diffusion-controlled die soldering in aluminum high pressure die casting, J. Mater. Process. Technol. 209 (8) (2009) 4051–4061. [6] S. Gulizia, M.Z. Jahedi, E.D. Doyle, Performance evaluation of PVD coatings for high pressure die casting, Surf. Coat. Technol. 140 (3) (2001) 200–205. [7] Z.W. Chen, Formation and progression of die soldering during high pressure die casting, Mater. Sci. Eng. A 397 (1–2) (2005) 356–369. [8] H. Zhu, J. Guo, J. Jia, Experimental study and theoretical analysis on die soldering in aluminum die casting, J. Mater. Process. Technol. 123 (2) (2002) 229–235. [9] M. Sundqvist, S. Hogmark, Effects of liquid aluminium on hot-work tool steel, Tribol. Int. 26 (2) (1993) 129–134. [10] P.A. Hogan, Die Solder Prediction and Reduction, PhD thesis Worcester Polytechnic Institute, 2008. [11] R. Shivpuri, M. Yu, K. Venkatesan, Y.L. Chu, A study of erosion in die casting dies by a multiple pin accelerated erosion test, J. Mater. Eng. Perform. 4 (2) (1995) 145–153. [12] V. Joshi, A. Srivastava, R. Shivpuri, Intermetallic formation and its relation to interface mass loss and tribology in die casting dies, Wear 256 (11−12) (2004) 1232–1235. [13] Q. Han, S. Viswanathan, Analysis of the mechanism of die soldering in aluminum die casting, Metall. Mater. Trans. A 34 (1) (2003) 139–146. [14] M. Yu, R. Shivpuri, R.A. Rapp, Effects of molten aluminum on H13 dies and coatings, J. Mater. Eng. Perform. 4 (2) (1995) 175–181. [15] Z.W. Chen, M.Z. Jahedi, Die erosion and its effect on soldering formation in high pressure die casting of aluminium alloys, Mater. Des. 20 (6) (1999) 303–309. [16] K. Venkatesan, R. Shivpuri, Experimental and numerical investigation of the effect of process parameters on the erosive wear of die casting dies, J. Mater. Eng. Perform. 4 (2) (1995) 166–174.
179
Engineering Failure Analysis 95 (2019) 171–180
R. Markežič et al.
[17] A. Mohammed, M.B. Marshall, R. Lewis, Development of a method for assessing erosive wear damage on dies used in aluminium casting, Wear 332–333 (2015) 1215–1224. [18] L.F. Hou, Y.H. Wei, Y.G. Li, B.S. Liu, H.Y. Du, C.L. Guo, Erosion process analysis of die-casting inserts for magnesium alloy components, Eng. Fail. Anal. 33 (2013) 457–564. [19] D.W.C. Baker, K.H. Jolliffe, D. Pearson, The resistance of materials to impact erosion damage, Philos. Trans. R. Soc. A Math. Phys. Eng. Sci. 260 (1110) (1966) 193–203. [20] A. Persson, S. Hogmark, J. Bergström, Temperature profiles and conditions for thermal fatigue cracking in brass die casting dies, J. Mater. Process. Technol. 152 (2) (2004) 228–236. [21] C. Rosbrook, Analysis of Thermal Fatigue and Heat Checking in Die-Casting Dies: A Finite Element Approach, PhD thesis Ohio State University, 1992. [22] F. Medjedoub, G. Dour, S. Le Roux, P. Lamesle, M. Salem, P. Hairy, F. Rézaï-Aria, Experimental conditions and environment effects on thermal fatigue damage accumulation and life of die-casting steel X38CrMoV5 (AISI H11), Int. J. Microstruct. Mater. Propert. 3 (2–3) (2008). [24] D. Klobčar, J. Tušek, Thermal stresses in aluminium alloy die casting dies, Comput. Mater. Sci. 43 (4) (2008) 1147–1154. [25] J. Sjöström, J. Bergström, Thermal fatigue testing of chromium martensitic hot-work tool steel after different austenitizing treatments, J. Mater. Process. Technol. 153 (2004) 1089–1096. [26] D. Klobčar, J. Tušek, B. Taljat, Thermal fatigue of materials for die-casting tooling, Mater. Sci. Eng. A 472 (1) (2008) 198–207. [27] A. Persson, Strain-based approach to crack growth and thermal fatigue life of hot work tool steels, Scand. J. Metall. 33 (2004) 53–64. [28] A. Persson, S. Hogmark, J. Bergström, Simulation and evaluation of thermal fatigue cracking of hot work tool steels, Int. J. Fatigue 26 (10) (2004) 1095–1107. [29] G. Roberts, G. Krauss, R. Kennedy, Tool Steels, 5th ed., ASM, Metals Park, 1998, p. 1998. [30] J. Sjöström and J. Bergström. Evaluation of the cyclic behaviour during high temperature fatigue of hot work tool steels. In Proceedings of 6th International Tooling Conference, pages 721–736. [31] D. Klobčar, L. Kosec, B. Kosec, J. Tušek, Thermo fatigue cracking of die casting dies, Eng. Fail. Anal. 20 (2012) 43–53. [32] M. Muhič, F. Kosel, A. Pukšič, D. Klobčar, A new approach to monitoring thermal fatigue cracks in die casting moulds, Int. J. Mat. Res. 102 (2011) 69–75. [33] D. Mellouli, N. Haddar, A. Köster, H.F. Ayedi, Hardness effect on thermal fatigue damage of hot-working tool steel, Eng. Fail. Anal. 45 (2014) 85–95. [34] M. Muhič, J. Tušek, F. Kosel, D. Klobčar, M. Pleterski, Thermal fatigue cracking of die-casting dies, Meta 49 (1) (2010) 9–12. [35] M. Fazarinc, T. Muhič, G. Kugler, M. Terčelj, Thermal fatigue properties of differently constructed functionally graded materials aimed for refurbishing of pressure-die-casting dies, Eng. Fail. Anal. 25 (2012) 238–249. [36] M. Pellizzari, A. Molinari, G. Straffelini, Thermal fatigue resistance of gas and plasma nitrided 41CrAlMo7 steel, Mater. Sci. Eng. A 352 (1–2) (2003) 186–194. [37] Bohler-Uddelholm, Dievar Data Sheets ®. Technical report, Uddeholm, (2013).
180