Effect of Sintering Process Parameters in Metal Injection Molding (MIM) Process on Impact Toughness of Sintered Parts

Transcription

1 Cloud Publications International Journal of Advanced Materials and Metallurgical Engineering 2016, Volume 2, Issue 1, pp , Article ID Tech-583 Research Article Open Access Effect of Sintering Process Parameters in Metal Injection Molding (MIM) Process on Impact Toughness of Sintered Parts Praveen Pachauri 1 and Mohammed Hamiuddin 2 1 Department of Mechanical Engineering, Noida Institute of Engineering & Technology, Greater Noida, Uttar Pradesh, India 2 Department of Mechanical Engineering Aligarh Muslim University, Aligarh, Uttar Pradesh, India Publication Date: 11 June 2016 Article Link: Copyright 2016 Praveen Pachauri and Mohammed Hamiuddin. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Abstract The identification of significant process parameters during sintering of brown parts in Metal Injection Moulding (MIM) is essential because improper control of these parameters can result poor parts. The controlled parameters used for optimization in this work include sintering temperature, heating rate, sintering time, cooling rate and sintering atmosphere. These parameters have been optimized using analysis of variance (ANOVA) for signal to noise ratios obtained in experiment performed by following Taguchi L 16 orthogonal array. The ANOVA also provides the contribution of significant process parameters to impact toughness. Results show that out of the five factors four are the significant factors. The sintering temperature (T s ) has a contribution of 23.64% at a confidence level of 95%, heating rate ( ) has a contribution of 9.76% at a confidence level of 95%, sintering time (t s ) has a contribution of 61.51% at a confidence level of 99%, cooling rate ( ) has a contribution of 3.82% with a confidence level of 90%, and sintering atmosphere is observed to be insignificant factor for impact energy absorbed by the specimen. Keywords Analysis of Variance (ANOVA); Impact Toughness; Orthogonal Array; Powder Injection Moulding (PIM); Sintering; Taguchi Method 1. Introduction Metal injection moulding (MIM) is a promising technology to process metal powders into parts of desired shapes. The MIM process combines the traditional shape-making capability of plastic injection moulding and materials flexibility of powder metallurgy [1]. The process consists of four main steps: mixing, injection moulding, debinding and sintering. In the mixing step, the powder is mixed with a binder to form a homogeneous feedstock. The binder is key component, which provides the necessary flowability and formability for moulding [2]. During injection moulding a green part with the desired shape is formed by the feedstock flow into a mold under pressure. After moulding, the binder holds the particles in place. The binder is then removed in the debinding step and the debound brown part is sintered to achieve the required mechanical properties. The quality of the brown parts affects the quality of the sintered parts. The geometrical accuracy and mechanical properties of the final parts

2 after sintering depend strongly on the process parameters in the different stages [3, 4]. The classical design of experiment (DOE) technique has been used by many authors [5-7] for optimization of single process parameters at a time. In order to obtain high efficiency in the planning and analysis of experimental data, the Taguchi method is recognized as a systematic approach for design and analysis of experiments to improve product quality. This method has been applied by many authors to investigate and optimize the process parameters [8-13]. The success of MIM process is highly dependent on sintering process [16-21]. The objective of this paper is to optimize the sintering process parameters to attain high impact toughness. In this paper, the experiment is conducted by following Taguchi L 16 orthogonal array and data is analyzed by using analysis of variance (ANOVA) to find the significant factors and their contribution in impact toughness of final part. 2. Materials and Methods To make the working material, the SS316L stainless steel powder was mixed with the binder comprised of polyethylene glycol (PEG), polymethyl methacrylate (PMMA), paraffin wax and stearic acid (SA) to form the feedstock for moulding. The main advantage of using PMMA/PEG binder is that it can be removed from the mouldings in a comparatively short time. The SS316L metal powder used in this research was supplied by Osprey. The chemical composition of the steel is presented in Table 1. The size distribution of metal powder is given in Table 2. The percentage concentration of constituents by weight and densities are given in Table 3. The details of the binder ingredients are given in Table 4. Table 1: Composition of SS316 L powder (Report given by Osprey alongwith the powder) Element % C Si Mn P S Cr Ni Mo Fe balance Table 2: Size distribution of SS316 L powder (Report given by Osprey ) Powder Tests Report by Sandvik Osprey Ltd. d10 d50 d90-53 µm Tap Density 3.9 µm 13.0 µm 36.6 µm 99.2 % 5.0 gm/cc Table 3: Theoretical density of constituents of SS316 L powder Element Percentage Concentration Theoretical Density C Si Mn P S Cr Ni Mo Fe SS 316 L Table 4: The Binder Ingredients Designation Manufacturer Amount (%) Melting temperature C Boiling point C International Journal of Advanced Materials and Metallurgical Engineering 24 Density(gm/ cc) Polymethyl Methacrylate (PMMA) Vetec Polyethylene Glycol(PEG) Rankem Paraffin Wax Thermo Fischer

3 Stearic acid(sa) Scientific Thermo Fischer Scientific Feedstock Formulation The metal powder and binder were mixed thoroughly for 90 minutes with the help of a Brookfield Rheometer in the desired proportion under precise weight and temperature control condition. The calculated amount of metal powder, PMMA, PEG, paraffin wax and stearic acid were weighed and mixed together. The mixing was carried out at 160 C and 40 rpm to achieve a homogeneous distribution of the powder particles and binder in feedstock. After thorough mixing, the mixture was first dried in air at ambient temperature for 2 hours and then in an oven at a temperature of 50 C for 1 hour. After compounding the feedstock was allowed to cool to at ambient temperature and then granulated in a rotary feedstock granulator Production of Test Specimen A four-cavity mould was specifically designed and made by National Small Industries Corporation (NSIC), Aligarh according to the specifications of the Demag injection moulding machine. The cavities were created in accordance with MPIF Standard 50 and ASTM Standard E8-98. The impact specimen geometry needed for this study was a 5mm x 10mm x 55mm (0.197 x x ) unnotched bar. Using the known shrinkage factor for the given feedstock, the impact bars were molded to produce oversize green parts as shown in Figure 1. The subsequent processing produced parts with the final dimensions as specified in MPIF Standard 59 [14]. Either no machining or very fine machining was needed to prepare the Charpy impact bars for testing. The green parts and sintered parts are shown in Figure 2 and Figure 3 respectively. Figure 1: Impact Test Bar (Specimen Size for Green Part) Figure 2: The Samples Produced by Injection Moulding International Journal of Advanced Materials and Metallurgical Engineering 25

4 Figure 3: The Samples Produced after Sintering 2.3. Injection Moulding Procedure Injection moulding process includes heating of the feedstock material to binder melting temperature, forcing the molten material into the mould cavities, holding at high pressure, then cooling and ejection of the molded parts out of the mould cavity. In the experimental work, a Demag injection moulding machine with microprocessor control was used. On the machine, the injection pressure, injection temperature, mould temperature, holding pressure, injection speed, holding time and cooling time were set at the desired values. Since, the powder loading is an external factor; it is not to be taken care by the machine control. Feedstock was developed before the start of the experiment with fine weight control and homogeneous mixing. Each set of values was repeated five times to make samples at each processing conditions after the machine has come to smooth functioning. All the test parts were produced using only virgin feedstock Debinding Procedure The green parts were debinded through the solvent and thermal debinding techniques to remove the binders effectively. In the first step, solvent extraction was used to extract out the PEG and paraffin wax from the green parts. The green specimens were immersed in high temperature distilled water with continuous stirring. The leached specimens were then dried in an oven to completely remove the remains of water and then cooled. The second step, referred to as thermal debinding was used to remove the PMMA and stearic acid after solvent debinding. The leached specimens were put into an alumina tray in which the surrounding space was filled with alumina powder to avoid any distortion of the specimens. The thermal debinding was achieved in a vacuum furnace. The brown part was allowed for slow cooling to ambient temperature (27 C) to release the residual stress from the part Sintering Procedure For sintering of brown parts was conducted in two steps. In first step the parts were presintered. The peak temperature for presintering after debinding was kept 900 C. The heating rate was 3 C/min and the holding time at peak temperature was 60 minutes. The cooling rate was 5 C/min. The presintered specimens were sintered afterwards in a batch furnace. The sintering was carried out at desired process parameters as shown in Table Design of Experiments and Testing Procedure The experiment was designed using Taguchi modified L 16 orthogonal array consisting of 16 experiment trials and 5 columns/experimental parameters to obtain the signal to noise ratio (S/N ratio) of final part quality. The method is based on balanced orthogonal arrays [15]. The sixteen different trial runs of sintering cycles were tested in this experiment to find the optimum level. The modified orthogonal array was used to accommodate four main factors with four levels and one two level factor without any International Journal of Advanced Materials and Metallurgical Engineering 26

5 interaction. The experimental results were converted in S/N values for optimization of parameters. The S/N ratio for higher the better was used. ANOVA is used to find the confidence level and variance of the data. The confidence level is measured from the variance of each parameter. Table 5: Response Table for Mean value of Impact Energy Absorbed Process Parameters Symbol Level 1 Level 2 Level 3 Level 4 Sintering temperature ( C) Ts Heating rate ( C/min) Sintering time (minutes) ts Cooling rate ( C/min) Sintering atmosphere p Vacuum N Results, Discussion and Conclusion The effect of variable controllable parameters on the mean values of impact toughness is measured by impact energy absorbed by the specimen during unnotched Charpy test. The calculated values for mean and S/N ratio at all process parameter levels are shown in Table 6 and Table 7. The analysis of variance made by using S/N ratio to find the significant factors is expressed in Table 8. The level average response required for the analysis of the trend of performance characteristic with respect to the variation of the factor under study is shown in Figure 4. From Figure 5 and 6 it can be observed that the impact toughness has maximum value when sintering temperature is 1380 Calongwithsintering time of 100 minute, heating rate of 8 C/min and cooling rate of 5 C/min. The dependence of impact toughness on significant process parameters can also be observed from Figures 7 to10. It can be observed from Figure7 that high impact energy is absorbed by the specimen when the combination of heating rate and sintering temperature is in dark green zone. From Figure8and 9 it can be observed that highest impact energy is absorbed when the heating rate is in the range is kept6.5 to 8.5 C/min whereas a heating rate of 11 to 13.5 causes poor results. From Figure 10 it can be observed that a slow cooling rate and a sintering time above 90 minute gives favourable results. Table 6: Response Table for Mean value of Impact Energy Absorbed Level T s t s p Delta Rank Table 7: Response Table for Signal to Noise Ratios (Larger is better) Level T s t s p Delta Rank International Journal of Advanced Materials and Metallurgical Engineering 27

6 Factors/ Source Table 8: Analysis of Variance using S/N ratios for Impact Energy Absorbed DOF, v Sums of squares Variance, V n Variance Ratio, F n Significance Level, α Pure Sum Square T s Contribution, P in % t s p (1) Pooled Residual Error Total Figure 4: Main Effects plot for Mean values of Impact Energy Absorbed (J) Figure 5: Interaction Plot for Sintering Temperature ( º C), Heating Rate ( º C/min), and Sintering Time (minutes) International Journal of Advanced Materials and Metallurgical Engineering 28

7 Sintering time (minutes) Heating rate ( C/min) IJAMME An Open Access Journal Figure 6: Interaction Plot for Heating Rate ( º C/min), Sintering Time (minutes), and Cooling Rate ( º C/min) Contour Curve for sintering temperature and heating rate MEAN3 < > Sintering temperature ( C) Figure 7: Contour Curve for Sintering Temperature and Heating Rate Contour Curve for heating rate and sintering time MEAN3 < > Heating rate ( C/min) 15.0 Figure 8: Contour Curve for Heating Rate and Sintering Time International Journal of Advanced Materials and Metallurgical Engineering 29

8 Cooling rate ( C/min) Cooling rate ( C/min) IJAMME An Open Access Journal Contour Curve for heating rate and cooling rate MEAN3 < > Heating rate ( C/min) 15.0 Figure 9: Contour Curve for Heating Rate and Cooling Rate Contour Curve for sintering time and cooling rate MEAN3 < > Sintering time (minutes) Figure 10: Contour Curve for Sintering Time and Cooling Rate From Table 8, it can be observed that out of the five factors four are significant factors. Since, there is one insignificant factor, so pooling is required. The sintering temperature (T s ) has a contribution of 23.64% at a confidence level of 95%, heating rate ( ) has a contribution of 9.76% at a confidence level of 95%, sintering time (t s ) has a contribution of 61.51% at a confidence level of 99%, cooling rate ( ) has a contribution of 3.82% with a confidence level of 90%, and sintering atmosphere is observed to be insignificant factor for impact energy absorbed by the specimen. From Table 7, it can further be noted from the rank of the parameters that variation in the value of S/N ratio with the change in the value of parameter is maximum for sintering time and least for sintering atmosphere. The optimum level of parameters can be obtained by selecting the highest values of S/N ratios from respective column in Table 7. The optimum level for impact energy absorbed occurs at (T s ) 4 ( ) 2 (t s ) 4 ( ) 1 (p) 1. From Table 7 it can also be noted that the rank of the variation in S/N ratio values with the change in the value of the parameter is maximum for sintering time and minimum for sintering atmosphere. International Journal of Advanced Materials and Metallurgical Engineering 30

9 Sintering temperature ( C) Table 9: Optimum Factor Levels for Highest Impact Energy Absorption Heating rate ( C/min) Sintering time (minutes) Cooling rate ( C/min) Sintering atmosphere Vacuum 3.1. Confirmation Test Since, the sintering temperature (T s ), heating rate ( ), sintering time (t s ), and cooling rate ( ) are all significant factors, the optimum value of Impact energy absorbed will depend on all these factors and could be estimated by eq. (1) at the optimum levels shown in Table 9. µ IT = + [(T s ) [( ) [(t s ) [( ) 1 - (1) Where, is the overall mean of Impact energy absorbed= J (T s ) 4 is the mean value of Impact energy absorbed at level 4 of factor T s = J, ( ) 2 is the mean value of Impact energy absorbed at level 2 of factor = J, (t s ) 4 is the mean value of Impact energy absorbed at level 4 of factor t s = J, ( ) 1 is the mean value of Impact energy absorbed at level 1 of factor = J, Hence, the expected Impact energy absorbed at optimum condition is: µ IT = (3 x ) = J The 95% confidence interval (CI) for the expected yield from the confirmation experiment can be calculated using eq. (2) as follows: (2) Where, n eff = (N/ (1+ total degree of freedom of all factors used for estimating µ) r= sample size for the confirmation experiment, r 0. is the variance ratio of and at level of significance α. The confidence level is (1-α), is the degree of freedom of mean (equal to 1) and is the degree of freedom for the pooled error. Variance for pooled error is V e. The confidence interval indicates the maximum and minimum levels of the optimum performance. Tabulated F-ratio at 95% confidence level (α = 0.05): F 0.05 ;(1,3) = 10.1and n eff = [16 x 5/13] = 6.15 CI = {10.1 x [(1/6.15) + (1/5)]} ½ = ± Therefore, expected impact energy absorbed at optimum condition = ± i.e < µ IT < To confirm the prediction, another five samples were made at the recommended setting (T s ) 4 ( ) 2 (t s ) 4 ( ) 1 and (p) 1 as shown in Table 9. The experimental observations are shown in Table 10. International Journal of Advanced Materials and Metallurgical Engineering 31

10 Table 10: Results of Confirmation Experiments Characteristic Replication at Optimum Process Parameters Average Minitab Predicted R1 R2 R3 R4 R5 Value Impact Energy absorbed (J) It can be observed that average impact toughness obtained from the confirmation experiment is close to the predicted 95% confidence interval. It is also evident from Table 10 that the experimental results are close to predicted result by Minitab 17 software. The difference between the measured and predicted values is about 3.30%. The difference can be due to linear approximation computed by Minitab 17 software. Acknowledgement The authors would like to acknowledge the financial support and facilities provided by Noida Institute of Engineering and Technology, Greater Noida for this research work. References [1] German, R.M., and Bose, A., 1997: Injection Moulding of Metals and Ceramics. Metal Powder Industries Federation [2] Berginc, B., Kampus, Z., and Sustarsic, B. The Use of Taguchi Approach to Determine the Influence of Injection Moulding Parameters on the Properties of Green Parts. Journal of Achievements in Materials and Manufacturing Engineering ; [3] German, R.M. Homogeneity Effects on Feedstock Viscosity in Powder Injection Moulding. Journal of American Ceramic Society (1) [4] Barriere, T., Liu, B., and Gelin, J.C. Determination of the Optimal Parameters in the Metal Injection Moulding from Experiments and Numerical Modeling. Journal of Materials Processing Technology ; [5] Heaney, D.F., Zauner, R., Binet, C., Cowan, K., and Piemme, J. Variability of Powder Characteristics and their Effect on Dimensional Variability of Powder Injection Moulded Components. Powder Metallurgy (2) [6] Wei, W.C.J., Wu, R.Y., and Ho, S.J. Effects of Pressure Parameters on Alumina Made by Powder Injection Moulding. Journal of European Ceramic Society ; [7] Li, Y., Li, L., and Khalil, K.A. Effect of Powder Loading on Metal Injection Moulding Stainless Steel. Journal of Materials Processing Technology ; [8] Ross, P.J. 1996: Taguchi Techniques for Quality Engineering. Tata McGraw Hill. [9] Roy, R.K. 2001: Design of Experiments Using the Taguchi Approach. John Wiley & Sons [10] Zu, Y.S., and Lin, S.T. Optimising the Mechanical Properties of Injection Molded W-4.9% Ni-2, 1% Fe in Debinding. Journal of Materials Processing Technology ; [11] Ji, C.H., Loh, N.H., Khor, K.A., and Tor, S.B. Sintering Study of Stainless Steel Metal Injection Moulding Parts using Taguchi Method: Final Density. Materials Science & Engineering A311; International Journal of Advanced Materials and Metallurgical Engineering 32

11 [12] Tseng, W.J. Statistical Analysis of Process Parameters Influencing Dimensional Control in Ceramic Injection Moulding. Journal of Materials Processing Technology ; [13] Jamaludin, K.R., Muhamad, N., Rahman, M.A., Amin, S.R.M., and Muntadhahadi. 2008: Analysis of Variance on the Metal Injection Moulding Parameters using a Bimodal Particle Size Distribution Feedstock. Proceedings of International Conference on Mechanical & Manufacturing Engineering. [14] MPIF 2002: Method for Determination of Charpy Impact Energy of Unnotched Metal Injection Molded Test Specimens. Standard 59, Standard Test Methods for Metal Powders and Powder Metallurgy Products. [15] Roy, R.K., 1990: A Primer on the Taguchi Method. Competitive Manufacturing Series, Van Nostrand Reinhold [16] German R.M. 1996: Sintering Theory and Practice. Wiley, New York. [17] Nayar, H.S., and Wasiczko B. Nitrogen Absorption Control during Sintering of Stainless Steel Parts. Metal Powder Report (9) [18] Talacchia, S., et al. Increasing Sintering Gate and Avoiding Grain Growth in High Speed Steels by Sintering in Nitrogen Rich Atmospheres. Powder metallurgy (4) [19] Urrutibeaskoa, I., and Urcola J.J. Sintering Behaviour of Grade M Water Atomised High Speed Steel Powders Under Vacuum and Nitrogen Rich Atmosphere. Powder metallurgy (1) [20] Song, J., et al. Experiments and Numerical Modelling of Solid State Sintering for 316L Stainless Steel Components. Journal of materials processing technology (1) [21] Li, S., et al. Influences of Sintering Atmospheres on Densification Process of Injection Moulded Gas Atomised 316L Stainless Steel. Powder metallurgy (3) International Journal of Advanced Materials and Metallurgical Engineering 33