- Email: [email protected]
The effect of induction heating power on the microstructural and physical properties of investment cast ASTM-F75 CoCrMo alloy
j m a t e r r e s t e c h n o l . 2 0 1 9;8(5):4417–4424
Available online at www.sciencedirect.com
www.jmrt.com.br
Original article
The effect of induction heating power on the microstructural and physical properties of investment cast ASTM-F75 CoCrMo alloy Thomas J. Fleming a,∗ , Alan Kavanagh a , Greg Duggan a , Brian O’Mahony a , Mackenzie Higgens b a b
DePuy Synthes, Loughbeg, Ringaskiddy, Co. Cork P43 ED82, Ireland DePuy Orthopaedics, Inc., 325 Paramount Drive, Raynham, 02767 MA, USA
a r t i c l e
i n f o
a b s t r a c t
Article history:
Investment cast ASTM F-75 Cobalt Chromium Molybdenum (CoCrMo) alloy is commonly
Received 21 September 2018
used for orthopaedic implants due to its biocompatibility and wear characteristics. Apart
Accepted 24 July 2019
from the values listed in the standard, the relationship between individual processing
Available online 13 August 2019
parameters and the resulting physio-mechanical properties are not well documented. This
Keywords:
ical properties of investment cast CoCrMo alloy. Furnace loads ranging from 30 kg to 45 kg
Investment casting
of alloy were inductively melted using 75 kW and 125 kW of power and cast into tensile test
study characterises the effect of induction heating power on the microstructural and phys-
Cobalt chrome molybdenum
bars. The bars were ground to create samples conforming to ASTM E8M and tested per ISO
ASTM F-75
6892-1. It was found the mean ultimate tensile strength, yield strength and % elongation
Orthopaedic devices
increased when the material was melted at 125 kW. There may be scope to improve the mean
Induction heating
mechanical properties of F75 by using higher induction melting power whilst reducing cycle
ISO 5832-4
time and cost. © 2019 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
1.
Introduction
Investment casting, or ‘lost wax’ casting as it is also known, is based on one of the oldest metal-forming process developed over 5000 years ago by the Egyptians for the production of gold jewellery. In the investment casting process a ceramic slurry is applied, or ‘invested’, around a disposable pattern, usually
∗
wax, and hardens to form a disposable casting mould. The wax pattern is ‘lost’ when it is melted from the ceramic mould, which later disintegrates further in the process to recover the casting. The most common metals used for total joint replacements are Cobalt-based alloys, Titanium-based alloys and 316 L series stainless steels. This is due in large part to their mechanical stability and biocompatibility in service. Cobalt base alloys were first patented in 1907 by Elwood Haynes and were referred to commercially as the Stellite® alloys. They were originally developed to produce fine cutlery, but soon found widespread application as tool materials for cutting and
Corresponding author. E-mail: tfl[email protected] (T.J. Fleming). https://doi.org/10.1016/j.jmrt.2019.07.052 2238-7854/© 2019 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http:// creativecommons.org/licenses/by-nc-nd/4.0/).
4418
j m a t e r r e s t e c h n o l . 2 0 1 9;8(5):4417–4424
high-speed machining due to their corrosion resistance and elevated temperature applications [1]. Several manufacturing methods, such as metal forming (e.g., rolling, forging), powder metallurgy (e.g., metal injection moulding), additive manufacturing and casting have been used successfully to produce Cobalt alloy articles [2]. The continued use of Cobalt alloys in biomedical applications is due to its long clinical heritage [3]. The first Cobalt Chromium (Co-Cr) alloy applied to orthopaedic surgery dates to 1937 [4]. This family of Co-Cr, as specified by ASTM F75 [5], is widely used in the manufacture of knee and hip orthopaedic implants and in dental appliances. ASTM F-75 is closely related to the Co-Cr-Mo-C and Co-Cr-W-Ni alloys named Stellite-21 and Stellite-25, respectively. These were developed as anti-corrosion alloys for use in gas turbine blades and vanes in the 1930 s. ASTM F-75 alloy was developed to provide outstanding corrosion resistance, wear behaviour [6–8] and biocompatibility—primarily due to the minimisation of the Nickel (Ni) content. Ni has been shown to be highly toxic in vivo [9,10]. To date, much of the research in the area of induction heating with respect to casting has been focused primarily on foundry productivity and energy consumption [11–13] with limited emphasis on casting quality or cast metallurgy. In Castings [14] John Campbell states “induction furnaces enjoy the great advantage of extremely rapid melting”, suggesting the benefits of high melt rates. However, Campbell also notes that the rate of induction stirring of the melt due to high induction power densities may exacerbate oxide entrainment, although he was aware of no supporting evidence at the time of writing. In recent years the novel Constrained Rapid Induction Melting Single Shot Up-Casting “CRIMSON” process was developed for light metal casting to reduce energy consumption and improve metal yield through a just in time, single shot melting process [15,16]. By rapidly melting the correct amount of metal for a single mould, the molten metal has less chance to react with the atmosphere to form an oxide film or to absorb hydrogen. Improvements in casting quality and defect occurrence are purported, however this is also attributed to the countergravity running system adopted for this process and not the induction melting process specifically. Swamy et al. [17] found superior mechanical properties and microstructure exhibited by test Al–Si–Mg Alloys castings when produced by means of rapid heating combined with vibration during solidification. While literature exists in this field, very little information has been gathered in an industrial setting, let alone the medical device field. The intent of this paper is to apply sound statistical techniques to quantify the relative effect of induction heating power on the physical and mechanical properties of cast CoCrMo alloy in an industrial environment. Thereby providing recommendations on best foundry practice to maximise mechanical performance.
2.
Materials and methods
The study was carried out in the investment casting process at DePuy Synthes with the objective to evaluate the broad
Table 1 – Chemical composition by weight % of ASTM F-75 Co-C alloy with BS7252 and Delores Stellite 21 Co-Cr alloys also listed. Element
ASTM F-75
BS7252,IV
Chromium, Cr Molybdenum, Mo Nickel, Ni Iron, Fe Carbon, C Silicon, Si Manganese, Mn Tungsten, W Phosphorus, P Sulphur, S Nitrogen, N Aluminium, Al Titanium, Ti Boron, B Cobalt, Co
27–30 % 5–7 % <0,5 % <0.75 % <0.35 % <1 % <1 % <0.2 % <0.02 % <0.01 % <0.25 % <0.1 % <0.1 % <0.01 % Balance
26.5-30.0% 4.5–7.0% 1.0 % max 1.0% max 0.35% max 1.0% max 1.0% max Balance
Stellite 21 26–29% 4.5–6.0% 2.0–3.0% 0.20–0.35% Balance
effect of induction heating power on the microstructural and physical properties of ASTM F-75. The chemical composition of the ASTM F-75 Co-Cr alloy is shown in Table 1 alongside similar Co-Cr alloys BS7252 Part 4 and Delores Stellite 21 for comparison. Tensile test bar patterns were produced in wax, Fig. 1, and assembled to a wax down sprue to form a casting tree. This tree was shelled using a proprietary ceramic system, primarily composed of zircon slurry and alumino-silicate stucco, and once dried, the wax casting tree was removed using an autoclave leaving a ceramic casting mould with a wall-thicknesses ˜ of 5–10 mm. Prior to casting, the ceramic shells were fired at temperatures up to 1100 ◦ C to sinter the ceramic matrix, burn out residual wax and pre-heat the moulds to accept the molten metal. To capture typically observed raw material variation, alloy for this study was drawn from 21 independent master heats provided by Arconic (USA). Three individual investment casts were performed per independent master heat (N = 63). Each runner system comprised 8 tensile specimens for redundancy. The study comprised of n = 256 individual data points. A small-scale induction furnace was used for this study, and two induction heating powers were used for this experiment, 125 kW and 75 kW. Given the limited access to production equipment in a live manufacturing environment and, as the intent of this paper is to study the broad effect of melt power, only two levels were chosen. It is not the intention of this work to perform response surface modelling of the entire power range, however this may be an avenue for future work. Of the n = 256, 170 were produced using 125 kW of induction heating power and the remaining 86 produced using 75 kW of induction heating power. Furnace charge sizes ranged from 30 kg to 45 kg. The charges were loaded into a refractory crucible and air-melted using a medium frequency coreless induction furnace (VIP 3000 Hz Power-Trak, Inductotherm UK). An inert fluid system was used to minimise atmospheric interactions with the melt. A type-R immersion thermocouple was used to target a melt ◦ C to 1600 ◦ C. Once the temperature ˜ ˜ temperature range of 1500
j m a t e r r e s t e c h n o l . 2 0 1 9;8(5):4417–4424
4419
Fig. 1 – Tensile test bar wax pattern, units in mm.
range was achieved, the material was cast into the moulds. By supplying 75 kW and 125 kW of power, the charges reached casting temperature in approximately 20 and 10 min, respectively. This corresponds to melting rates of 3.00–4.50 kg/min & 1.50–2.25 kg/min, respectively. The resultant cast test bars were machine ground to possess a nominal gauge length of 32 mm and a nominal diameter of 6.3 mm with a 1% taper applied to the gauge as per ASTM E8M Fig. 8, specimen 3. The samples were then tested with respect to ISO 5832-4 using strain rates per method A of ISO 6892-1 of 0.00025 s−1 through yield at 0.2% offset and 0.00007 s−1 after yield until fracture. ¨ ISO 5832-4 provides the following direction Should any of the test pieces not meet the specified requirements, or should they break outside the gauge limits, two further test pieces representative of the same batch shall be tested in the same manner. The alloy shall be deemed to comply only if both addi¨ as tional test pieces meet the specified requirementshowever the intention of this experiment is not to qualify alloy or to perform manufacturing batch release, all test results have been used for data analysis. Post tensile testing several of the untested tensile specimens were subject to metallurgical evaluation. Samples were selected to represent populations of the lowest and highest observed mechanical properties. Specimens were mounted both in the transverse (radial) and longitudinal (axial) directions. The mounts were planar ground and polished using successively finer grits of silicon carbide paper and a mirror finish achieved using colloidal silica. The mounts were subject to optical examination at 100× magnification to assess porosity. Subsequently the mounts were electrolytically etched using an aqueous solution ammonium persulphate and 6 V DC. The etched mounts were re-examined to assess the resultant microstructures.
3.
Results and discussion
3.1.
Mechanical properties ISO-5832-4
3.1.1.
Ultimate tensile strength
It was found that there was a statistical difference in the means and standard deviations of Ultimate Tensile Strength (p = 0.000 and p = 0.000, respectively). The material cast using 125 kW demonstrated values of 811.5 ± 18.8 MPa
Fig. 2 – Ultimate tensile strength of ASTM F-75 processed using 75 kW and 125 kW of induction melting power and tensile tested as per ISO 5832-4.
whereas the material cast using 75 kW demonstrated values of 791.0 ± 36.2 MPa, as shown in Fig. 2. The values obtained by using 75 kW to melt the ASTM F-75 material followed a 3-Parameter Weibull distribution (p = 0.050). Using this distribution predicts 5314 parts per million below the 665 MPa lower bound defined in ISO 5832-4. The values obtained by using 125 kW to melt the ASTM F-75 material followed a 3-Parameter Weibull distribution (p = 0.453). Using this distribution predicts 0 parts per million below the 665 MPa lower bound defined in ISO 5832-4.
3.1.2.
Yield strength
It was found that there was a statistical difference in the means and standard deviations of Yield Strength (p = 0.000 and p = 0.000, respectively). The material cast using 125 kW demonstrated values of 518.9 ± 12.8 MPa whereas the material cast using 75 kW demonstrated values of 495.0 ± 25.2 MPa, Fig. 3. The values obtained by using 75 kW to melt the ASTM F-75 material followed a 3-Parameter Weibull distribution (p = 0.286). Using this distribution predicts 43,182 parts per million below the 450 MPa lower bound defined in ISO 5832-4. The values obtained by using 125 kW to melt the ASTM F-75 material followed a 3-Parameter Weibull distribution
4420
j m a t e r r e s t e c h n o l . 2 0 1 9;8(5):4417–4424
Fig. 3 – Yield strength of ASTM F-75 processed using 75 kW and 125 kW of induction melting power and tensile tested as per ISO 5832-4.
Fig. 5 – Transverse cross section from the centre of a tensile bar processed using 75 kW.
could not be matched to a distribution due to outlying values at 4% and 19%.
3.2.
Fig. 4 – Elongation (%) of ASTM F-75 processed using 75 kW and 125 kW of induction melting power and tensile tested as per ISO 5832-4.
Microstructural evaluation
Untested tensile specimens were selected to represent the mechanical properties of the best performing and poorest performing master heats. On average (n = 14), the poorest performing master heat demonstrated 721 MPa, 445 MPa and 8%, Ultimate Tensile strength, Yield strength and Elongation respectively. This population shall be denoted as 75 kW. On average (n = 12), the best performing master heat demonstrated 832 MPa, 532 MPa and 12%, Ultimate Tensile strength, Yield strength and elongation respectively. This population shall be denoted as 125 kW.
3.2.1. Porosity 3.2.1.1. 75 kW. Fig. 5 shows a cross section of tensile bar pro(p = 0.181). Using this distribution predicts 0 parts per million below the 450 MPa lower bound defined in ISO 5832-4.
3.1.3.
Elongation
It was found that there was a statistical difference in the mean of Elongation (p = 0.001 Bonnett’s test and p = 0.043 Levene’s test), Fig. 4. Due to the resolution of the measured data, it was not possible to apply distribution fitting. To address this, a uniform data distribution was created ranging from –0.5 to 0.5 and added to the original data sets. In effect adding additional variation and, as a by-product, smoothing out the resolution of the % Elongation data. The values obtained by using 75 kW to melt the ASTM F75 material, with the additional uniform distribution added, followed a 3-Parameter Weibull distribution (p => 0.500). Using this distribution predicts 76,666 parts per million below the 8% lower bound defined in ISO 5832-4. The values obtained by using 125 kW to melt the ASTM F75 material, with the additional uniform distribution added,
cessed at 75 kW. The cross section is imaged from the centre of the tensile bar at 100×. The presence of micro gas porosity is evident in this cross section. Micro gas porosity is also evident towards the edge of the tensile bar as shown, however there is also shrinkage porosity present as shown in Fig. 6. Similar micro gas porosity is evident in tensile bars when micrographs are taken in the longitudinal orientation, as shown in Fig. 7.
3.2.1.2. 125 kW. Fig. 8 shows a cross section of tensile bar processed at 125 kW. The cross section is imaged from the centre of the tensile bar at 100×. In contrast to Fig. 6, much smaller micro gas porosity is evident in tensile bars produced using 125 kW heating power, with a max diameter of 10.4 m versus 5.7 m for those cast at 75 kW, see Table 2. Fig. 9 shows comparable results for the longitudinal cross section (Fig. 10). Figs. 5, 7–9 were subject to image analysis using Image J v1.49k [4] to determine the number and average diameter of micro gas pores per mm2 . The results are presented in Table 2.
4421
j m a t e r r e s t e c h n o l . 2 0 1 9;8(5):4417–4424
Table 2 – Micro gas pores/mm2 in the transverse and longitudinal directions of tensile bars cast using 75 kW and 125 kW of induction melting power.
125 kW 75 kW Kaiser et al. [18]
Pores/mm2
Average micro gas pore diameter (m)
Longitudinal
Transverse
3.38 5.02 3.57
29 28 -
53 14 -
Fig. 6 – Transverse cross section from the edge of a tensile bar processed using 75 kW.
Fig. 9 – Longitudinal cross section from the centre of a tensile bar processed using 125 kW.
Fig. 7 – Longitudinal cross section from the edge of a tensile bar processed using 75 kW.
Fig. 10 – Mean pore diameter of material processed using 75 kW and 125 kW of induction heating power.
Fig. 8 – Transverse cross section from the centre of a tensile bar processed using 125 kW.
The % area porosity was also calculated, and the results plotted in Fig. 11, below Analysis of the micrographs indicates that there are fewer micro gas pores present when 75 kW is used, however, those pores are larger. Conversely the material melted using 125 kW exhibited more, smaller, pores which are distributed throughout the section. Using the values presented above it was calculated that the mean pore size diameter of the 75 kW material is 2.3 times that of the 125 kW material. This approximation was verified,
4422
j m a t e r r e s t e c h n o l . 2 0 1 9;8(5):4417–4424
Fig. 11 – Area % of gas micro porosity in material processed using 75 kW and 125 kW of induction heating power.
Fig. 13 – Etched longitudinal cross section from the centre of a tensile bar processed using 75 kW.
Fig. 12 – Etched transverse cross section from the centre of a tensile bar processed using 75 kW. Fig. 14 – Etched transverse cross section from the centre of a tensile bar processed using 125 kW. within an order of magnitude, by performing individual pore diameter measurements.
3.2.2. Microstructure 3.2.2.1. 75 kW. Fig. 12 is a representative image of the microstructure observed in the transverse cross section of tensile bar processed at 75 kW. The cross section is imaged from the centre of the tensile bar at 100×. Fig. 13 is a representative image of the microstructure observed in the longitudinal cross section of tensile bar processed at 75 kW. The cross section is imaged from the centre of the tensile bar at 100×.
3.2.2.2. 125 kW. Fig. 14 is a representative image of the microstructure observed in the transverse cross section of tensile bar processed at 125 kW. The cross section is imaged from the centre of the tensile bar at 100×. Fig. 15 is a representative image of the microstructure observed in the transverse cross section of tensile bar processed at 125 kW. The cross section is imaged from the centre of the tensile bar at 100× magnification.
Fig. 15 – Etched longitudinal cross section from the centre of a tensile bar processed using 125 kW.
j m a t e r r e s t e c h n o l . 2 0 1 9;8(5):4417–4424
Fig. 16 – Area % of carbides in material processed using 75 kW and 125 kW of induction heating power.
4423
Fig. 18 – Secondary electron image of the fracture surface.
Table 3 – Area % carbides in the transverse and longitudinal directions of tensile bars cast using 75 kW and 125 kW of induction melting power. Area % Carbide Longitudinal 125 kW 75 kW Kaiser et al. [18]
8.6 7.5 ˜ 3.5
Transverse 8.5 8.4
Fig. 19 – Secondary electron image of the suspected fracture initiation site.
mean went from 49.42 m to 37.88 m and the spread from 13.79 m to 4.25 m, respectively. Kaiser et al. [18] measured an ˜ m on cast cylindrical samples which were 12 mm SDAS of 35 in diameter. Fig. 17 – Secondary Dendrite Arm Spacing (SDAS) in material processed using 75 kW and 125 kW of induction heating power.
A visual appraisal of the microstructures reveals that they are quite similar. All etched micrographs exhibited an austenitic microstructure with a dispersion of carbides, typical of as cast ASTM F-75 alloy. There was no evidence of continuous blocky carbides on the grain boundaries, which would be typical of slow cooling rates. The micrographs were again analysed using ImageJ to determine the area % carbides, Fig. 16 and Table 3, and the secondary dendrite arm spacing (SDAS), Fig. 17. The SDAS of the etched microstructures were measured. It was found that an increase in melt power resulted in a reduction in both the spread, and the mean, of the data points. The
3.2.3.
Fracture surfaces
The bar which exhibited the lowest elongation value, 4%, also demonstrated values of 674 MPa Ultimate Tensile Strength and 436 MPa Yield strength. To assess if this inferior performance was related to the microstructure, the remnants were returned from the external test laboratory in order to examine the fracture surface. It was found that the fracture surface exhibits regions of surface texture which are typical of inter and intra grain fracture mechanisms, Fig. 18. There is also a large fissure which may also be associated with the shrinkage porosity observed in Fig. 6. A more detailed examination was performed of the fracture ˜ m defect was observed on the periphsurface, Fig. 19. A 150 ery of the bar, the directionality of the striations indicates a potential fracture initiation site.
4424
4.
j m a t e r r e s t e c h n o l . 2 0 1 9;8(5):4417–4424
Conclusions
It has been found that increasing the supplied induction power, from 75 kW to 125 kW, has had the effect of both reducing the variability and increasing the mean in the mechanical properties of ASTM F-75. The increase in heating rate shortened the time to reach the designated casting temperature, reducing atmospheric entrainment. This is consistent with the observed reduction in size of individual micro gas porosity, and elimination of shrinkage porosity. Both mechanisms serve to minimise internal defect size and thus could contribute to the improvement and increased consistency of mechanical properties. There was also an inferred grain size reduction from the SDAS measurements. Therefore, the authors recommend that to reduce the variability and increase the mean in the mechanical properties of cast ASTM F-75 CoCrMo alloy for use in orthopaedic implants, higher induction melting power should be used. There is an added benefit of reducing cycle time and cost. As mentioned in the introduction, Campbell [14] notes that the rate of induction stirring of the melt due to high induction power densities may exacerbate oxide entrainment. This effect was not observed for the charge size and power regime used in this study.
references
[1] Pike LM. 100+ Years of Wrought Alloy Development at Haynes International, in 8th International Symposium on Superalloy 718 and Derivatives. Hoboken, NJ: John Wiley & Sons, Inc.; 2014. p. 15–30. [2] Davis JR. Handbook of Materials for Medical Devices. Materials Park Ohio ASM International; 2003. [3] Zhuang LZ, Langer EW. Effects of cooling rate control during the solidification process on the microstructure and mechanical properties of cast Co-Cr-Mo alloy used for surgical implants. J Mat Sci 1989;24(2):381–8. [4] Huang Q, Wang L, Wang J. Mechanical properties of artificial materials for bone repair. J Shanghai Jiaotong University 2014;19(6):675–80. [5] ASTM. Standard Specification for Cobalt-28 Chromium-6 Molybdenum Alloy Castings and Casting Alloy for Surgical Implants (UNS R30075). PA: West Conshohocken; 2012.
[6] Zhuang LZ, Langer EW. Effects of alloy additions on the microstructures and tensile properties of cast Co-Cr-Mo alloy used for surgical implants. J Mat Sci 1989;24(12):4324–30. ˜ [7] Mischler S, Munoz AI. Wear of CoCrMo alloys used in metal-on-metal hip joints: A tribocorrosion appraisal. J Wear 2013;297(1–2):1081–94. [8] Niinomi M, Nakai M, Hieda J. Development of new metallic alloys for biomedical applications. J Acta Biomater 2012;8(11):3888–903. [9] Wataha JC, Lockwood PE, Marek M, et al. Ability of Ni-containing biomedical alloys to activate monocytes and endothelial cells in vitro. J Biomed Mater Res 1999;45(3):251–7. [10] Hiromoto S, Onodera E, Chiba A, et al. Microstructure and corrosion behaviour in biological environments of the new forged low-Ni Co–Cr–Mo alloys. Biomaterials 2005;26(24):4912–23. [11] Patil Digvijay D, Ghatge Dayanand. Parametric Evaluation of Melting Practice on Induction Furnace to Improve Efficiency and System Productivity of CI and SGI Foundry- A Review. IARJSET 2017;4:159–63. [12] Inc., B. Advanced Melting Technologies:Energy Saving Concepts and Opportunities for the Metal Casting Industry; 2005. [13] Mishra A, Bag S, Pal S. Induction Heating in Sustainable Manufacturing and Material Processing Technologies—A State of the Art Literature Review. In: Reference Module in Materials Science and Materials Engineering. Elsevier; 2019. [14] Campbell J. Castings. Elsevier; 2003. [15] Salonitis K, Jolly MR, Zeng B, et al. Improvements in energy consumption and environmental impact by novel single shot melting process for casting. J Clean Prod 2016;137:1532–42. [16] Salonitis K, Zeng B, Mehrabi HA, et al. The challenges for energy efficient casting processes. Procedia CIRP 2016;40:24–9. [17] Swamy NHS, Joseph MA, Nagarajan NM, et al. Effects of rapid heating on microstructure and mechanical properties of modified vibrated Al–Si–Mg alloys using a fluidized bed. Mater Manuf Processes 2011;26(2):210–7. [18] Kaiser R, Williamson K, O’Brien C, et al. The influence of cooling conditions on grain size, secondary phase precipitates and mechanical properties of biomedical alloy specimens produced by investment casting. J Mech Behav Biomed Mater 2013;24:53–63.
Available online at www.sciencedirect.com
www.jmrt.com.br
Original article
The effect of induction heating power on the microstructural and physical properties of investment cast ASTM-F75 CoCrMo alloy Thomas J. Fleming a,∗ , Alan Kavanagh a , Greg Duggan a , Brian O’Mahony a , Mackenzie Higgens b a b
DePuy Synthes, Loughbeg, Ringaskiddy, Co. Cork P43 ED82, Ireland DePuy Orthopaedics, Inc., 325 Paramount Drive, Raynham, 02767 MA, USA
a r t i c l e
i n f o
a b s t r a c t
Article history:
Investment cast ASTM F-75 Cobalt Chromium Molybdenum (CoCrMo) alloy is commonly
Received 21 September 2018
used for orthopaedic implants due to its biocompatibility and wear characteristics. Apart
Accepted 24 July 2019
from the values listed in the standard, the relationship between individual processing
Available online 13 August 2019
parameters and the resulting physio-mechanical properties are not well documented. This
Keywords:
ical properties of investment cast CoCrMo alloy. Furnace loads ranging from 30 kg to 45 kg
Investment casting
of alloy were inductively melted using 75 kW and 125 kW of power and cast into tensile test
study characterises the effect of induction heating power on the microstructural and phys-
Cobalt chrome molybdenum
bars. The bars were ground to create samples conforming to ASTM E8M and tested per ISO
ASTM F-75
6892-1. It was found the mean ultimate tensile strength, yield strength and % elongation
Orthopaedic devices
increased when the material was melted at 125 kW. There may be scope to improve the mean
Induction heating
mechanical properties of F75 by using higher induction melting power whilst reducing cycle
ISO 5832-4
time and cost. © 2019 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
1.
Introduction
Investment casting, or ‘lost wax’ casting as it is also known, is based on one of the oldest metal-forming process developed over 5000 years ago by the Egyptians for the production of gold jewellery. In the investment casting process a ceramic slurry is applied, or ‘invested’, around a disposable pattern, usually
∗
wax, and hardens to form a disposable casting mould. The wax pattern is ‘lost’ when it is melted from the ceramic mould, which later disintegrates further in the process to recover the casting. The most common metals used for total joint replacements are Cobalt-based alloys, Titanium-based alloys and 316 L series stainless steels. This is due in large part to their mechanical stability and biocompatibility in service. Cobalt base alloys were first patented in 1907 by Elwood Haynes and were referred to commercially as the Stellite® alloys. They were originally developed to produce fine cutlery, but soon found widespread application as tool materials for cutting and
Corresponding author. E-mail: tfl[email protected] (T.J. Fleming). https://doi.org/10.1016/j.jmrt.2019.07.052 2238-7854/© 2019 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http:// creativecommons.org/licenses/by-nc-nd/4.0/).
4418
j m a t e r r e s t e c h n o l . 2 0 1 9;8(5):4417–4424
high-speed machining due to their corrosion resistance and elevated temperature applications [1]. Several manufacturing methods, such as metal forming (e.g., rolling, forging), powder metallurgy (e.g., metal injection moulding), additive manufacturing and casting have been used successfully to produce Cobalt alloy articles [2]. The continued use of Cobalt alloys in biomedical applications is due to its long clinical heritage [3]. The first Cobalt Chromium (Co-Cr) alloy applied to orthopaedic surgery dates to 1937 [4]. This family of Co-Cr, as specified by ASTM F75 [5], is widely used in the manufacture of knee and hip orthopaedic implants and in dental appliances. ASTM F-75 is closely related to the Co-Cr-Mo-C and Co-Cr-W-Ni alloys named Stellite-21 and Stellite-25, respectively. These were developed as anti-corrosion alloys for use in gas turbine blades and vanes in the 1930 s. ASTM F-75 alloy was developed to provide outstanding corrosion resistance, wear behaviour [6–8] and biocompatibility—primarily due to the minimisation of the Nickel (Ni) content. Ni has been shown to be highly toxic in vivo [9,10]. To date, much of the research in the area of induction heating with respect to casting has been focused primarily on foundry productivity and energy consumption [11–13] with limited emphasis on casting quality or cast metallurgy. In Castings [14] John Campbell states “induction furnaces enjoy the great advantage of extremely rapid melting”, suggesting the benefits of high melt rates. However, Campbell also notes that the rate of induction stirring of the melt due to high induction power densities may exacerbate oxide entrainment, although he was aware of no supporting evidence at the time of writing. In recent years the novel Constrained Rapid Induction Melting Single Shot Up-Casting “CRIMSON” process was developed for light metal casting to reduce energy consumption and improve metal yield through a just in time, single shot melting process [15,16]. By rapidly melting the correct amount of metal for a single mould, the molten metal has less chance to react with the atmosphere to form an oxide film or to absorb hydrogen. Improvements in casting quality and defect occurrence are purported, however this is also attributed to the countergravity running system adopted for this process and not the induction melting process specifically. Swamy et al. [17] found superior mechanical properties and microstructure exhibited by test Al–Si–Mg Alloys castings when produced by means of rapid heating combined with vibration during solidification. While literature exists in this field, very little information has been gathered in an industrial setting, let alone the medical device field. The intent of this paper is to apply sound statistical techniques to quantify the relative effect of induction heating power on the physical and mechanical properties of cast CoCrMo alloy in an industrial environment. Thereby providing recommendations on best foundry practice to maximise mechanical performance.
2.
Materials and methods
The study was carried out in the investment casting process at DePuy Synthes with the objective to evaluate the broad
Table 1 – Chemical composition by weight % of ASTM F-75 Co-C alloy with BS7252 and Delores Stellite 21 Co-Cr alloys also listed. Element
ASTM F-75
BS7252,IV
Chromium, Cr Molybdenum, Mo Nickel, Ni Iron, Fe Carbon, C Silicon, Si Manganese, Mn Tungsten, W Phosphorus, P Sulphur, S Nitrogen, N Aluminium, Al Titanium, Ti Boron, B Cobalt, Co
27–30 % 5–7 % <0,5 % <0.75 % <0.35 % <1 % <1 % <0.2 % <0.02 % <0.01 % <0.25 % <0.1 % <0.1 % <0.01 % Balance
26.5-30.0% 4.5–7.0% 1.0 % max 1.0% max 0.35% max 1.0% max 1.0% max Balance
Stellite 21 26–29% 4.5–6.0% 2.0–3.0% 0.20–0.35% Balance
effect of induction heating power on the microstructural and physical properties of ASTM F-75. The chemical composition of the ASTM F-75 Co-Cr alloy is shown in Table 1 alongside similar Co-Cr alloys BS7252 Part 4 and Delores Stellite 21 for comparison. Tensile test bar patterns were produced in wax, Fig. 1, and assembled to a wax down sprue to form a casting tree. This tree was shelled using a proprietary ceramic system, primarily composed of zircon slurry and alumino-silicate stucco, and once dried, the wax casting tree was removed using an autoclave leaving a ceramic casting mould with a wall-thicknesses ˜ of 5–10 mm. Prior to casting, the ceramic shells were fired at temperatures up to 1100 ◦ C to sinter the ceramic matrix, burn out residual wax and pre-heat the moulds to accept the molten metal. To capture typically observed raw material variation, alloy for this study was drawn from 21 independent master heats provided by Arconic (USA). Three individual investment casts were performed per independent master heat (N = 63). Each runner system comprised 8 tensile specimens for redundancy. The study comprised of n = 256 individual data points. A small-scale induction furnace was used for this study, and two induction heating powers were used for this experiment, 125 kW and 75 kW. Given the limited access to production equipment in a live manufacturing environment and, as the intent of this paper is to study the broad effect of melt power, only two levels were chosen. It is not the intention of this work to perform response surface modelling of the entire power range, however this may be an avenue for future work. Of the n = 256, 170 were produced using 125 kW of induction heating power and the remaining 86 produced using 75 kW of induction heating power. Furnace charge sizes ranged from 30 kg to 45 kg. The charges were loaded into a refractory crucible and air-melted using a medium frequency coreless induction furnace (VIP 3000 Hz Power-Trak, Inductotherm UK). An inert fluid system was used to minimise atmospheric interactions with the melt. A type-R immersion thermocouple was used to target a melt ◦ C to 1600 ◦ C. Once the temperature ˜ ˜ temperature range of 1500
j m a t e r r e s t e c h n o l . 2 0 1 9;8(5):4417–4424
4419
Fig. 1 – Tensile test bar wax pattern, units in mm.
range was achieved, the material was cast into the moulds. By supplying 75 kW and 125 kW of power, the charges reached casting temperature in approximately 20 and 10 min, respectively. This corresponds to melting rates of 3.00–4.50 kg/min & 1.50–2.25 kg/min, respectively. The resultant cast test bars were machine ground to possess a nominal gauge length of 32 mm and a nominal diameter of 6.3 mm with a 1% taper applied to the gauge as per ASTM E8M Fig. 8, specimen 3. The samples were then tested with respect to ISO 5832-4 using strain rates per method A of ISO 6892-1 of 0.00025 s−1 through yield at 0.2% offset and 0.00007 s−1 after yield until fracture. ¨ ISO 5832-4 provides the following direction Should any of the test pieces not meet the specified requirements, or should they break outside the gauge limits, two further test pieces representative of the same batch shall be tested in the same manner. The alloy shall be deemed to comply only if both addi¨ as tional test pieces meet the specified requirementshowever the intention of this experiment is not to qualify alloy or to perform manufacturing batch release, all test results have been used for data analysis. Post tensile testing several of the untested tensile specimens were subject to metallurgical evaluation. Samples were selected to represent populations of the lowest and highest observed mechanical properties. Specimens were mounted both in the transverse (radial) and longitudinal (axial) directions. The mounts were planar ground and polished using successively finer grits of silicon carbide paper and a mirror finish achieved using colloidal silica. The mounts were subject to optical examination at 100× magnification to assess porosity. Subsequently the mounts were electrolytically etched using an aqueous solution ammonium persulphate and 6 V DC. The etched mounts were re-examined to assess the resultant microstructures.
3.
Results and discussion
3.1.
Mechanical properties ISO-5832-4
3.1.1.
Ultimate tensile strength
It was found that there was a statistical difference in the means and standard deviations of Ultimate Tensile Strength (p = 0.000 and p = 0.000, respectively). The material cast using 125 kW demonstrated values of 811.5 ± 18.8 MPa
Fig. 2 – Ultimate tensile strength of ASTM F-75 processed using 75 kW and 125 kW of induction melting power and tensile tested as per ISO 5832-4.
whereas the material cast using 75 kW demonstrated values of 791.0 ± 36.2 MPa, as shown in Fig. 2. The values obtained by using 75 kW to melt the ASTM F-75 material followed a 3-Parameter Weibull distribution (p = 0.050). Using this distribution predicts 5314 parts per million below the 665 MPa lower bound defined in ISO 5832-4. The values obtained by using 125 kW to melt the ASTM F-75 material followed a 3-Parameter Weibull distribution (p = 0.453). Using this distribution predicts 0 parts per million below the 665 MPa lower bound defined in ISO 5832-4.
3.1.2.
Yield strength
It was found that there was a statistical difference in the means and standard deviations of Yield Strength (p = 0.000 and p = 0.000, respectively). The material cast using 125 kW demonstrated values of 518.9 ± 12.8 MPa whereas the material cast using 75 kW demonstrated values of 495.0 ± 25.2 MPa, Fig. 3. The values obtained by using 75 kW to melt the ASTM F-75 material followed a 3-Parameter Weibull distribution (p = 0.286). Using this distribution predicts 43,182 parts per million below the 450 MPa lower bound defined in ISO 5832-4. The values obtained by using 125 kW to melt the ASTM F-75 material followed a 3-Parameter Weibull distribution
4420
j m a t e r r e s t e c h n o l . 2 0 1 9;8(5):4417–4424
Fig. 3 – Yield strength of ASTM F-75 processed using 75 kW and 125 kW of induction melting power and tensile tested as per ISO 5832-4.
Fig. 5 – Transverse cross section from the centre of a tensile bar processed using 75 kW.
could not be matched to a distribution due to outlying values at 4% and 19%.
3.2.
Fig. 4 – Elongation (%) of ASTM F-75 processed using 75 kW and 125 kW of induction melting power and tensile tested as per ISO 5832-4.
Microstructural evaluation
Untested tensile specimens were selected to represent the mechanical properties of the best performing and poorest performing master heats. On average (n = 14), the poorest performing master heat demonstrated 721 MPa, 445 MPa and 8%, Ultimate Tensile strength, Yield strength and Elongation respectively. This population shall be denoted as 75 kW. On average (n = 12), the best performing master heat demonstrated 832 MPa, 532 MPa and 12%, Ultimate Tensile strength, Yield strength and elongation respectively. This population shall be denoted as 125 kW.
3.2.1. Porosity 3.2.1.1. 75 kW. Fig. 5 shows a cross section of tensile bar pro(p = 0.181). Using this distribution predicts 0 parts per million below the 450 MPa lower bound defined in ISO 5832-4.
3.1.3.
Elongation
It was found that there was a statistical difference in the mean of Elongation (p = 0.001 Bonnett’s test and p = 0.043 Levene’s test), Fig. 4. Due to the resolution of the measured data, it was not possible to apply distribution fitting. To address this, a uniform data distribution was created ranging from –0.5 to 0.5 and added to the original data sets. In effect adding additional variation and, as a by-product, smoothing out the resolution of the % Elongation data. The values obtained by using 75 kW to melt the ASTM F75 material, with the additional uniform distribution added, followed a 3-Parameter Weibull distribution (p => 0.500). Using this distribution predicts 76,666 parts per million below the 8% lower bound defined in ISO 5832-4. The values obtained by using 125 kW to melt the ASTM F75 material, with the additional uniform distribution added,
cessed at 75 kW. The cross section is imaged from the centre of the tensile bar at 100×. The presence of micro gas porosity is evident in this cross section. Micro gas porosity is also evident towards the edge of the tensile bar as shown, however there is also shrinkage porosity present as shown in Fig. 6. Similar micro gas porosity is evident in tensile bars when micrographs are taken in the longitudinal orientation, as shown in Fig. 7.
3.2.1.2. 125 kW. Fig. 8 shows a cross section of tensile bar processed at 125 kW. The cross section is imaged from the centre of the tensile bar at 100×. In contrast to Fig. 6, much smaller micro gas porosity is evident in tensile bars produced using 125 kW heating power, with a max diameter of 10.4 m versus 5.7 m for those cast at 75 kW, see Table 2. Fig. 9 shows comparable results for the longitudinal cross section (Fig. 10). Figs. 5, 7–9 were subject to image analysis using Image J v1.49k [4] to determine the number and average diameter of micro gas pores per mm2 . The results are presented in Table 2.
4421
j m a t e r r e s t e c h n o l . 2 0 1 9;8(5):4417–4424
Table 2 – Micro gas pores/mm2 in the transverse and longitudinal directions of tensile bars cast using 75 kW and 125 kW of induction melting power.
125 kW 75 kW Kaiser et al. [18]
Pores/mm2
Average micro gas pore diameter (m)
Longitudinal
Transverse
3.38 5.02 3.57
29 28 -
53 14 -
Fig. 6 – Transverse cross section from the edge of a tensile bar processed using 75 kW.
Fig. 9 – Longitudinal cross section from the centre of a tensile bar processed using 125 kW.
Fig. 7 – Longitudinal cross section from the edge of a tensile bar processed using 75 kW.
Fig. 10 – Mean pore diameter of material processed using 75 kW and 125 kW of induction heating power.
Fig. 8 – Transverse cross section from the centre of a tensile bar processed using 125 kW.
The % area porosity was also calculated, and the results plotted in Fig. 11, below Analysis of the micrographs indicates that there are fewer micro gas pores present when 75 kW is used, however, those pores are larger. Conversely the material melted using 125 kW exhibited more, smaller, pores which are distributed throughout the section. Using the values presented above it was calculated that the mean pore size diameter of the 75 kW material is 2.3 times that of the 125 kW material. This approximation was verified,
4422
j m a t e r r e s t e c h n o l . 2 0 1 9;8(5):4417–4424
Fig. 11 – Area % of gas micro porosity in material processed using 75 kW and 125 kW of induction heating power.
Fig. 13 – Etched longitudinal cross section from the centre of a tensile bar processed using 75 kW.
Fig. 12 – Etched transverse cross section from the centre of a tensile bar processed using 75 kW. Fig. 14 – Etched transverse cross section from the centre of a tensile bar processed using 125 kW. within an order of magnitude, by performing individual pore diameter measurements.
3.2.2. Microstructure 3.2.2.1. 75 kW. Fig. 12 is a representative image of the microstructure observed in the transverse cross section of tensile bar processed at 75 kW. The cross section is imaged from the centre of the tensile bar at 100×. Fig. 13 is a representative image of the microstructure observed in the longitudinal cross section of tensile bar processed at 75 kW. The cross section is imaged from the centre of the tensile bar at 100×.
3.2.2.2. 125 kW. Fig. 14 is a representative image of the microstructure observed in the transverse cross section of tensile bar processed at 125 kW. The cross section is imaged from the centre of the tensile bar at 100×. Fig. 15 is a representative image of the microstructure observed in the transverse cross section of tensile bar processed at 125 kW. The cross section is imaged from the centre of the tensile bar at 100× magnification.
Fig. 15 – Etched longitudinal cross section from the centre of a tensile bar processed using 125 kW.
j m a t e r r e s t e c h n o l . 2 0 1 9;8(5):4417–4424
Fig. 16 – Area % of carbides in material processed using 75 kW and 125 kW of induction heating power.
4423
Fig. 18 – Secondary electron image of the fracture surface.
Table 3 – Area % carbides in the transverse and longitudinal directions of tensile bars cast using 75 kW and 125 kW of induction melting power. Area % Carbide Longitudinal 125 kW 75 kW Kaiser et al. [18]
8.6 7.5 ˜ 3.5
Transverse 8.5 8.4
Fig. 19 – Secondary electron image of the suspected fracture initiation site.
mean went from 49.42 m to 37.88 m and the spread from 13.79 m to 4.25 m, respectively. Kaiser et al. [18] measured an ˜ m on cast cylindrical samples which were 12 mm SDAS of 35 in diameter. Fig. 17 – Secondary Dendrite Arm Spacing (SDAS) in material processed using 75 kW and 125 kW of induction heating power.
A visual appraisal of the microstructures reveals that they are quite similar. All etched micrographs exhibited an austenitic microstructure with a dispersion of carbides, typical of as cast ASTM F-75 alloy. There was no evidence of continuous blocky carbides on the grain boundaries, which would be typical of slow cooling rates. The micrographs were again analysed using ImageJ to determine the area % carbides, Fig. 16 and Table 3, and the secondary dendrite arm spacing (SDAS), Fig. 17. The SDAS of the etched microstructures were measured. It was found that an increase in melt power resulted in a reduction in both the spread, and the mean, of the data points. The
3.2.3.
Fracture surfaces
The bar which exhibited the lowest elongation value, 4%, also demonstrated values of 674 MPa Ultimate Tensile Strength and 436 MPa Yield strength. To assess if this inferior performance was related to the microstructure, the remnants were returned from the external test laboratory in order to examine the fracture surface. It was found that the fracture surface exhibits regions of surface texture which are typical of inter and intra grain fracture mechanisms, Fig. 18. There is also a large fissure which may also be associated with the shrinkage porosity observed in Fig. 6. A more detailed examination was performed of the fracture ˜ m defect was observed on the periphsurface, Fig. 19. A 150 ery of the bar, the directionality of the striations indicates a potential fracture initiation site.
4424
4.
j m a t e r r e s t e c h n o l . 2 0 1 9;8(5):4417–4424
Conclusions
It has been found that increasing the supplied induction power, from 75 kW to 125 kW, has had the effect of both reducing the variability and increasing the mean in the mechanical properties of ASTM F-75. The increase in heating rate shortened the time to reach the designated casting temperature, reducing atmospheric entrainment. This is consistent with the observed reduction in size of individual micro gas porosity, and elimination of shrinkage porosity. Both mechanisms serve to minimise internal defect size and thus could contribute to the improvement and increased consistency of mechanical properties. There was also an inferred grain size reduction from the SDAS measurements. Therefore, the authors recommend that to reduce the variability and increase the mean in the mechanical properties of cast ASTM F-75 CoCrMo alloy for use in orthopaedic implants, higher induction melting power should be used. There is an added benefit of reducing cycle time and cost. As mentioned in the introduction, Campbell [14] notes that the rate of induction stirring of the melt due to high induction power densities may exacerbate oxide entrainment. This effect was not observed for the charge size and power regime used in this study.
references
[1] Pike LM. 100+ Years of Wrought Alloy Development at Haynes International, in 8th International Symposium on Superalloy 718 and Derivatives. Hoboken, NJ: John Wiley & Sons, Inc.; 2014. p. 15–30. [2] Davis JR. Handbook of Materials for Medical Devices. Materials Park Ohio ASM International; 2003. [3] Zhuang LZ, Langer EW. Effects of cooling rate control during the solidification process on the microstructure and mechanical properties of cast Co-Cr-Mo alloy used for surgical implants. J Mat Sci 1989;24(2):381–8. [4] Huang Q, Wang L, Wang J. Mechanical properties of artificial materials for bone repair. J Shanghai Jiaotong University 2014;19(6):675–80. [5] ASTM. Standard Specification for Cobalt-28 Chromium-6 Molybdenum Alloy Castings and Casting Alloy for Surgical Implants (UNS R30075). PA: West Conshohocken; 2012.
[6] Zhuang LZ, Langer EW. Effects of alloy additions on the microstructures and tensile properties of cast Co-Cr-Mo alloy used for surgical implants. J Mat Sci 1989;24(12):4324–30. ˜ [7] Mischler S, Munoz AI. Wear of CoCrMo alloys used in metal-on-metal hip joints: A tribocorrosion appraisal. J Wear 2013;297(1–2):1081–94. [8] Niinomi M, Nakai M, Hieda J. Development of new metallic alloys for biomedical applications. J Acta Biomater 2012;8(11):3888–903. [9] Wataha JC, Lockwood PE, Marek M, et al. Ability of Ni-containing biomedical alloys to activate monocytes and endothelial cells in vitro. J Biomed Mater Res 1999;45(3):251–7. [10] Hiromoto S, Onodera E, Chiba A, et al. Microstructure and corrosion behaviour in biological environments of the new forged low-Ni Co–Cr–Mo alloys. Biomaterials 2005;26(24):4912–23. [11] Patil Digvijay D, Ghatge Dayanand. Parametric Evaluation of Melting Practice on Induction Furnace to Improve Efficiency and System Productivity of CI and SGI Foundry- A Review. IARJSET 2017;4:159–63. [12] Inc., B. Advanced Melting Technologies:Energy Saving Concepts and Opportunities for the Metal Casting Industry; 2005. [13] Mishra A, Bag S, Pal S. Induction Heating in Sustainable Manufacturing and Material Processing Technologies—A State of the Art Literature Review. In: Reference Module in Materials Science and Materials Engineering. Elsevier; 2019. [14] Campbell J. Castings. Elsevier; 2003. [15] Salonitis K, Jolly MR, Zeng B, et al. Improvements in energy consumption and environmental impact by novel single shot melting process for casting. J Clean Prod 2016;137:1532–42. [16] Salonitis K, Zeng B, Mehrabi HA, et al. The challenges for energy efficient casting processes. Procedia CIRP 2016;40:24–9. [17] Swamy NHS, Joseph MA, Nagarajan NM, et al. Effects of rapid heating on microstructure and mechanical properties of modified vibrated Al–Si–Mg alloys using a fluidized bed. Mater Manuf Processes 2011;26(2):210–7. [18] Kaiser R, Williamson K, O’Brien C, et al. The influence of cooling conditions on grain size, secondary phase precipitates and mechanical properties of biomedical alloy specimens produced by investment casting. J Mech Behav Biomed Mater 2013;24:53–63.