Postmolding Shrinkage Evaluation of 316L Feedstock Micromolded

Sep 22, 2014 - Cranfield University, Building 61, Wharley End, Cranfield, Bedfordshire MK43 0AL, United Kingdom. ABSTRACT: The purpose of this paper ...
0 downloads 0 Views 1MB Size
Download PDF
Article pubs.acs.org/IECR

Postmolding Shrinkage Evaluation of 316L Feedstock Micromolded Parts D. Annicchiarico,*,† U. M. Attia,*,‡ and J. R. Alcock*,§ †

Weber Saint Gobain Ltd., Dickens House, Enterprise Way, Flitwick, Bedford MK45 5BY, United Kingdom Manufacturing Technology Centre Ltd., Ansty Park, Pilot Way, Coventry CV7 9JU, United Kingdom § Cranfield University, Building 61, Wharley End, Cranfield, Bedfordshire MK43 0AL, United Kingdom ‡

ABSTRACT: The purpose of this paper was to evaluate the shrinkage behavior of a 316L molding feedstock. The methodology adopted a statistical approach (design of experiment) and a standard microshrinkage measurement approach. The statistical approach identified the mold temperatureparallel to the flow directionand the combined effect of the holding and injection pressurenormal to the flow directionas critical factors. In comparison with the polymer on which the feedstock was based, lower shrinkage values and fewer critical factors were observed. In conclusion, the lower shrinkage values were a consequence of the powder loading. The critical factors identified in the present work have found confirmation in the literature, except the absence of melt temperature between feedstock critical factors. of nanoceramic powders were used by Huang and Chiu10 for filling square geometries. The work did not adopt a statistical approach, but the results showed that shrinkage can be significantly reduced by increasing the powder loading, and the injection pressure and mold temperature had to be increased to improve the cavity filling. The 316L molded shrinkage was compared with respect to its pure binder shrinkage [poly(oxymethylene) (POM)].11 The comparison between 316L feedstock and POM was performed in terms of different shrinkage values as a consequence of the powder loading and in terms of different critical factors. Previous studies investigated the powder loading effect on the morphology12 and thermal properties:13,14 the latter aspect is important also because of the influence the crystalline arrangement of molded parts.15 With regard to the shrinkage value, there is little prior work in the literature to compare with this trend. However, Huang and Chiu10 demonstrated that high powder loading has been shown to lead to lower shrinkage at the microscale (although without any estimate of statistical significance). The comparison between 316L feedstock and POM shrinkage considered also the processing parameters: mold temperature, holding pressure, injection pressure, holding time, and melt temperature. Previous studies demonstrated that the mold temperature is a critical factor for pure polymers16,17 and for feedstocks.8,10 As a trend, an increase of the mold temperature leads to a decrease in shrinkage. The holding pressure is a critical factor both for POM11,18 and for feedstock:8 as a general trend, high holding values lead to reduced shrinkage.

1. INTRODUCTION Powder injection molding (PIM) is a near-net-shape manufacturing technique for ceramic or metals.1 It typically consists of four steps: feedstock (binder mixed with powder) formulation, molding, debinding, and sintering.2 The potential of net-shape manufacturingin particular at the microscale (μ-PIM)3is key to the future development of several technologies, notably the nonsilicon microelectromechanical system.4 The μ-PIM process still faces several technical challenges3 such as the mechanical stability of the microstructures (especially those with high aspect ratio) and the accuracy and reproducibility of final dimensions. Both of these aspects are affected by shrinkage.5 Total shrinkage in μ-PIM components can be thought of as comprised of three parts: postmolding shrinkage, postdebinding shrinkage, and postsintering shrinkage. However, in comparison to the microshrinkage of nonfilled polymers, postmolding shrinkage has been little studied in the μ-PIM literature, and it is likely that this behavior will be different from that observed for macroscale PIM.6 Because shrinkage will form a part of the overall replication quality at the microscale,7 some indications with regard to factors affecting shrinkage may be taken from microscale replication studies. Tay et al.8 considered the replication quality of micropillar arrays molded using 316L powder mixed with a low-density polyethylene binder base: the mold temperature, melt temperature, and packing pressure were shown to affect the replication quality. Fu et al.9 investigated the influence of the same processing parameters as those adopted in the present paper on micropillar molded parts by injecting a 316L feedstock. The study was not performed using a statistical approach, but the final results showed that high pressures (injection and holding) and molding temperature values improved and affected the filling of microcavities. The injection pressure and mold temperature also were considered to be critical parameters by Huang and Chiu10 and Tay et al.,8 respectively. Feedstocks with different percentages © 2014 American Chemical Society

Received: Revised: Accepted: Published: 16559

November 26, 2013 September 21, 2014 September 22, 2014 September 22, 2014 dx.doi.org/10.1021/ie4040048 | Ind. Eng. Chem. Res. 2014, 53, 16559−16567

Industrial & Engineering Chemistry Research

Article

The injection pressure affected polymer19,20 and feedstock10 shrinkage as well. However, the injection pressure can be affected by the viscosity,21 which can be affected by powder loading.22 Moreover, Seokyoung et al.23 reported that the critical influence of the injection pressure is not simply predictable because of the pseudoplastic feedstock behavior and temperature influences. The holding time emerged as critical only for pure polymers,24,25 and no work reported a critical influence of this parameter for the feedstock. The melt temperature is an important parameter both for feedstock8 and for pure polymers.25 1.1. Purpose of the Paper. The aim of this paper was to evaluate shrinkage of a common molded feedstock in the form of microscale green parts. According to the literature review reported in ref 26, few data concerning shrinkage of green 316L feedstock molded parts are available. Besides, no paper that investigated 316L feedstock shrinkage adopted the standardized methodology proposed in ref 11.

Table 2. 316L Feedstock Processing Values Investigated processing parameters organized according to the half-fractional factorial matrix

2. METHODOLOGY 2.1. Material. The feedstock used for the present study was the Catamold 316LS BASF. According to the manufacturer data sheet, the material properties of the sintered part were as follows: density, ≥7.9 g/cm3; yield strength, ≥180 MPa; ultimate tensile strength, ≥510 MPa; elongation, ≥50%; hardness, 120 HV10. The mean 316L powder size was about D50 = 4−5 μm, with a volume powder loading of around 60%. The binder composition was a mixture of POM and other different polymers in lower percentages [poly(methyl methacrylate), polypropylene, poly(butylene terephthalate), polystyrene, and elastomers]. 2.2. Choice of Processing Parameters. The processing parameters were chosen by reviewing different papers in the area of microinjection molding: the works of Zhao et al.,18 Vasco et al.,27 and Shen and Wu.28 The same parameters were used in ref 11: injection pressure, holding pressure, melt temperature, mold temperature, and holding time. 2.3. Experimental Design and Procedure. The shrinkage study followed the procedure and mold design already adopted in ref 11. A micromold was manufactured. The final dimensions of the single square cavity were length = 9.987 ± 0.001 mm, breadth = 9.980 ± 0.001 mm, and height = 0.350 ± 0.001 mm. It is important to report that experimental evidence have shown that the scale can affect shrinkage, as summarized in the review in ref 26. In the present paper, five processing parameters (reported in Table 1) were investigated according

processing parameter

initial value

value +

value −

900 250 194 135 3

930 300 198 140 4

870 200 190 130 2

holding time [s]

holding pressure [bar]

injection pressure [bar]

mold temp [°C]

melt temp [°C]

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

4 2 4 2 4 2 2 2 4 2 2 4 4 4 2 4

300 300 200 300 200 300 200 200 200 300 300 200 300 300 200 300

930 930 930 870 870 930 870 930 870 870 870 930 870 870 930 930

140 140 130 140 130 130 130 140 140 130 140 140 130 140 130 130

198 190 198 190 190 198 198 198 198 190 198 190 198 190 190 190

Figure 1. Measurement protocol adopted.

Because the corner close to the gate was always present, this was chosen as the zero point. The specimen was moved 5 mm to point 1. A line position was measured. The specimen was moved in the flow direction until the opposite edge was reached. A second line position was measured (2). The specimen was moved back by 5 mm and then moved in a cross direction until the edge was found. A line position was measured (4). The specimen was moved in the direction opposite to the other edge. A final line position was measured (5). The dimensions were converted to shrinkage values (molding, postmolding, and total shrinkages) according to ISO 294-4. A statistical approach (DoE, design of experiment) was adopted for studying shrinkage, and the statistical program Minitab 16 was used for detecting the critical factors. The experiments were conducted without replications, with a Battenfeld Microsystem 50 injection machine. This approach was decided because of the molding conditions. The sources of noise, such as humidity, were believed to be constant. With regard to the humidity, the injection machine allows one to control the polymer temperature stored in a sealed hopper (set at 70 °C), and this assured constant polymer conditions throughout the entire test campaign. Regarding other variables connected to the machine setup,

Table 1. 316L Feedstock Processing Parameter Values injection pressure [bar] holding pressure [bar] melt temperature [°C] mold temperature [°C] holding time [s]

run

to the half-fractional factorial matrix (Table 2); five specimens for each processing combination were chosen after 20 cycles (considered sufficient for stabilization of the molding parameters); the specimens were optically measured using the TESA VISIO 300, equipment with a precision of ±1 μm (parallel and normal to the flow direction) following the protocol depicted in Figure 1. 16560

dx.doi.org/10.1021/ie4040048 | Ind. Eng. Chem. Res. 2014, 53, 16559−16567

Industrial & Engineering Chemistry Research

Article

Table 3. 316L Feedstock Shrinkage Results run 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

SMp [%] 0.651 0.806 0.834 0.758 0.843 0.726 0.686 0.842 0.656 0.771 0.879 1.052 0.599 0.961 0.800 1.289

± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±

0.002 0.002 0.007 0.005 0.002 0.002 0.003 0.001 0.001 0.001 0.003 0.003 0.002 0.003 0.004 0.004

SMn [%] 0.447 0.621 0.699 0.603 0.586 0.684 0.374 0.497 0.524 0.560 0.583 0.383 0.690 0.542 0.344 0.493

± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±

0.003 0.002 0.006 0.002 0.001 0.003 0.002 0.002 0.001 0.003 0.002 0.004 0.001 0.001 0.001 0.003

SPp [%] −0.136 −0.095 0.789 −0.128 −0.252 0.126 0.010 −0.128 0.084 0.192 −0.097 −0.028 0.535 −0.084 0.031 −0.100

± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±

0.005 0.005 0.019 0.008 0.004 0.004 0.006 0.001 0.003 0.003 0.005 0.005 0.007 0.008 0.006 0.007

SPn [%] 0.119 −0.237 0.325 −0.083 −0.484 −0.328 0.573 0.192 −0.053 0.144 −0.007 0.362 0.037 0.006 0.331 −0.088

± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±

0.005 0.004 0.020 0.004 0.004 0.004 0.006 0.002 0.004 0.005 0.003 0.005 0.004 0.003 0.006 0.005

STp [%] 0.515 0.712 1.616 0.631 0.598 0.851 0.696 0.715 0.740 0.961 0.783 1.024 1.130 0.877 0.830 1.191

± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±

0.004 0.005 0.018 0.004 0.003 0.003 0.003 0.001 0.002 0.003 0.002 0.001 0.006 0.007 0.002 0.005

STn [%] 0.566 0.385 1.022 0.521 0.104 0.358 0.945 0.688 0.471 0.702 0.575 0.743 0.727 0.548 0.674 0.406

± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±

0.003 0.003 0.017 0.003 0.003 0.002 0.005 0.001 0.003 0.003 0.002 0.002 0.003 0.003 0.006 0.002

The red line represents the statistical significance level determined by the α value, the maximum acceptable level of risk, and was expressed as a probability ranging between 0 and 1. The noise due to the microscale makes it difficult to estimate the critical factors because of low feedstock shrinkage (signal values). When the uncertainty of the results was reduced, the α value was increased from the common value of 0.05−0.1 (consequently, the confidence threshold decreased from 95 to 90%). A higher α value means that rejection of a true null hypothesis is more likely, but it also means that detection of real critical factors affecting the shrinkage is more likely. Main Effect Plot. The main effect graph helps to visualize the effect of the factors on the response and to compare the relative strength of the effects. Each main effect plot shows the single factors: the slope of the line representing the magnitude and direction of the effect on the response. The average of all of the data was the reference line. If the line represented on the main effect graph was horizontal (parallel to the reference line), then the response does not change depending on the factor level. The greater the slope of the line, the stronger the effect. The main plots on this paper reported only the critical factors that crossed the statistical significance line. Interaction Plot. Sometimes the critical factor can be determined by a combination of two single factors. In that case, when the effect of a one factor depends on the level of the other factor, instead of the main effect plot, it was necessary to analyze the interaction plot. An interaction plot is a plot of the means for each level of a factor, with the level of a second factor held constant.

previous replicated works have shown that the machine does not affect the trends in the variable observed.29 Indeed in this paper the authors plotted the average masses for three replications, and considering the worst run, the values are spread within an interval of 1.7% with respect to the average. Parameters held constant during the 316L feedstock molding stage were the cooling time (15 s), metering volume (185 mm3), and injection speed (300 mm/s). 2.4. Shrinkage Definitions. As summarized in the review published in ref 26, different factors (processing parameters, mold and specimen design, material properties, and scale effects) can affect shrinkage. However, as reported in section 1.1, the present paper analyzed only the influence of processing parameters. The ISO standard adopted30 allowed one to conduct shrinkage measurements as a linear dimensional variation. The measurements were reported in percentage values and were measured at two instances of time and in two directions: (a) immediately after the component is ejected from the mold and measured with respect to the cavity mold, referred to as “molding shrinkage”; (b) after 24 h of ejection and molding shrinkage measurements, when the component has reached its final dimensions, referred to as “postmolding shrinkage”. Also the “total shrinkage” was calculated, referred to as the 24 h specimen dimensions with respect to the cavity mold. The two shrinkage directions were (a) parallel to the flow direction of the material, referred to as the “parallel-flow” direction, and (b) normal to the flow direction, referred to as the “normal-flow” direction. Therefore, the results shown in Table 3 were at six different conditions combining the three shrinkages and the two shrinking directions: (a) “mold shrinkage” at the “parallel direction” (SMp); (b) “mold shrinkage” at the “normal direction” (SMn); (c) “postmolding shrinkage” at the “parallel direction” (SPp); (d) “postmolding shrinkage” at the “normal direction” (SPn); (e) “total shrinkage” at the “parallel direction” (STp), and (f) “total shrinkage” at the “normal direction” (STn). 2.5. Statistical Representations. The statistical study was performed using the Pareto chart, the main effect plot, and the interaction plot, using the Minitab 16.31 Pareto Chart. The Pareto chart is a bar chart that graphically ranks the criticality of factors from largest to smallest. This representation helps to determine the magnitude and importance of the processing parameters and to prioritize the problems.

3. RESULTS Table 3 reports the 316L feedstock shrinkage values. 3.1. Statistical Analysis of Molding Shrinkage Parallel to the Flow Direction. The Pareto chart depicted in Figure 1 represents the molding shrinkage in the “flow direction” (SMp). As reported in Figure 2, each processing parameter was represented by a letter. If the significant effect is associated with a combination of more than one parameter, relevant letters are grouped together. The processing parameters are labeled as A (holding time), B (holding pressure), C (injection pressure), D (mold temperature), and E (melt temperature). Pareto’s results showed that there were no critical factors that affect the 16561

dx.doi.org/10.1021/ie4040048 | Ind. Eng. Chem. Res. 2014, 53, 16559−16567

Industrial & Engineering Chemistry Research

Article

Figure 2. Pareto chart of 316L feedstock molding shrinkage in parallel to the flow direction.

Figure 3. Pareto chart of 316L feedstock molding shrinkage normal to the flow direction.

molding shrinkage parallel to the flow direction within the 90% confidence level. 3.2. Statistical Analysis of Molding Shrinkage Normal to the Flow Direction. Figure 3 represents the Pareto chart of molding shrinkage normal to the flow direction: no statistically significant factors were detected within the 90% confidence level. 3.3. Statistical Analysis of Postmolding Shrinkage Parallel to the Flow Direction. Figure 4 depicts the Pareto chart for postmolding shrinkage parallel to the flow direction: the mold temperature was identified as a critical parameter. Figure 5 reports the main effect plot: the slope of the critical factor within the interval of confidence shows that SPp decreases when the mold temperature increases. 3.4. Statistical Analysis of Postmolding Shrinkage Normal to the Flow Direction. Figure 6 reports the Pareto chart of postmolding shrinkage normal to the flow direction: no critical factors were detected within the 90% of confidence level.

3.5. Statistical Analysis of Total Shrinkage Parallel to the Flow Direction. Figure 7 reports the Pareto chart for otal shrinkage parallel to the flow direction: no critical processing factors were detected within the 90% confidence level. 3.6. Statistical Analysis of Total Shrinkage Normal to the Flow Direction. The total shrinkage normal to the flow direction was critically affected by the combination of the holding pressure and injection pressure, as reported in Figure 8. Figure 9 represents the interaction plot between the injection pressure and holding pressure (labeled as BC in Figure 8), considered critical factors within the interval of confidence. The boxes show the changing of total shrinkage normal to the flow direction STn. The top right box plots shrinkage as a function of the injection pressure, and the bottom left box plots shrinkage as a function of the holding pressure. With regard to the injection pressure (top right) moving from low pressure (870 bar) to high pressure (930 bar), shrinkage improves with a low holding pressure value (200 bar) 16562

dx.doi.org/10.1021/ie4040048 | Ind. Eng. Chem. Res. 2014, 53, 16559−16567

Industrial & Engineering Chemistry Research

Article

Figure 4. Pareto chart of 316L feedstock postmolding shrinkage parallel to the flow direction.

Figure 5. Main effect of 316L feedstock postmolding shrinkage parallel to the flow direction.

and decreases with a high holding pressure value (300 bar). With regard to the holding pressure (bottom left) moving from a low pressure (200 bar) to a high pressure (300 bar), shrinkage improves with a low injection pressure value (870 bar) and decreases with a high injection pressure value (930 bar).

This statistical study identified the mold temperature as a critical parameter for postmolding shrinkage parallel to the flow direction, and the combination of the holding pressure and injection pressure as an interaction that affected total shrinkage normal to the flow direction: the same processing parameters were identified as critical factors by previous studies, as reported in the Introduction. With regard to the trend shown in Table 4, an increase in the values of the critical factors within the interval of confidence led to an increase in the cavity filling (hence, to decreased shrinkage): this trend was confirmed by literature results as well. 4.2. 316L Feedstock and Pure Polymer Molding Shrinkage Differences. Table 6 compares the average 316L feedstock shrinkage values in Table 3 and POM shrinkage average values from previous authors’ papers.11 Table 5 reports POM molding conditions.11 With respect to the conventional injection molding, micromolding requires precise metering control and high-speed

4. DISCUSSION This discussion is divided into three parts: (a) factors affecting 316L feedstock shrinkage; (b) a comparison between shrinkages of 316L feedstock and POM, the polymer on which the feedstock is based; (c) difference in terms of critical factors between the feedstock and pure binder. The comparison between 316L feedstock and POM was performed because it can highlight the consequences of powder loading. 4.1. 316L Feedstock Shrinkage. Table 4 reports the critical factors within the interval of confidence that affect 316L feedstock molding shrinkage. 16563

dx.doi.org/10.1021/ie4040048 | Ind. Eng. Chem. Res. 2014, 53, 16559−16567

Industrial & Engineering Chemistry Research

Article

Figure 6. Pareto chart of 316L feedstock postmolding shrinkage normal to the flow direction.

Figure 7. Pareto chart of 316L feedstock total shrinkage parallel to the flow direction.

injection.32 Under these conditions, also the transition from velocity to pressure control (V−P transition) during the injection stage is an important parameter, as stated by Zhang and Gilchrist.32 Generally, the V−P switchover is determined by the injection screw position: the injection machine used in this study adopted this method, and the switchover position is 5 mm. POM is the baseline polymer of the feedstock: compared to 316L feedstock shrinkage, polymer shrinkage is approximately 5 times greater parallel to the flow direction and 6 times greater normal to the flow direction (Table 6). This result was confirmed by the literature. The shrinkage behavior for the feedstock is more isotropic. For the polymer, the total shrinkage parallel is 68% greater than that normal to the flow, and for the feedstock, it is 47%. A spherical powder shape is thought to influence shrinkage because it makes the feedstock more isotropic (no preferred molecular orientations) with respect to crystalline POM.33 This may be the underlying mechanism here.

4.3. Difference of the Critical Factors between the Feedstock and Pure Binder. A comparison of 316L feedstock critical factors with its binder (POM) can be made by comparing Tables 4 and 7. A lower number of critical factors within the interval of confidence is apparent for the feedstock in comparison to the binder. The differences in the critical factors are discussed below. The mold temperature is present as a critical factor both in Table 4 for the 316L feedstock (SPp) and in Table 7 for the POM polymer (SMp, SMn, and STp). The trend of this factor is confirmed by the literature: an increase of the mold temperature leads to a decrease in shrinkage. The holding pressure factor is present in both the 316L feedstock (STn, with injection pressure) and POM (SPn, with mold temperature). Similar to the trend reported in the present paper, literature results evidenced that high holding values lead to reduced shrinkage. 16564

dx.doi.org/10.1021/ie4040048 | Ind. Eng. Chem. Res. 2014, 53, 16559−16567

Industrial & Engineering Chemistry Research

Article

Figure 8. Pareto chart of 316L feedstock total shrinkage normal to the flow direction.

Figure 9. Interaction plot of 316L feedstock total shrinkage normal to the flow direction.

Table 4. 316L Feedstock Shrinkage Critical Factorsa

Table 6. Total Shrinkage (ST) Average Values between POM11 and 316L Feedstock Parallel (p) and Normal (n) to the Flow Direction

316L feedstock SPp STn

↓ ↓

mold temperature holding and injection pressure

a

POM 316L feedstock

The arrows indicate that the factor increase causes a decrease (↓) in shrinkage.

initial values

value +

value −

injection pressure [bar] holding pressure [bar] melt temperature [°C] mold temperature [°C] holding time [s]

850 500 195 100 3

900 550 200 115 4

800 450 190 85 2

STn [%] 3.026 ± 0.266 0.590 ± 0.227

The injection pressure is present only in Table 4 for the 316L feedstock (STn, with holding pressure). As stated in the Introduction, this result contrasts with the literature. However, the presence of injection pressure as a critical parameter for the 316L feedstock and not for POM might be connected to the feedstock powder loading because literature results confirmed that powder loading leads to an increase in the feedstock viscosity, and viscosity affects the injection pressure. It is likely that the pseudoplastic feedstock behavior makes the relationship between the viscosity and injection pressure not simply predictable, as confirmed by the literature.

Table 5. POM Processing Parameter Values As Reported in Reference 11 POM processing parameters

STp [%] 5.089 ± 1.658 0.867 ± 0.274

16565

dx.doi.org/10.1021/ie4040048 | Ind. Eng. Chem. Res. 2014, 53, 16559−16567

Industrial & Engineering Chemistry Research

Article

Table 7. Critical Factors That Affect Shrinkage of POMa

direction presented similar percentage results and trends, even if feedstock shrinkage values were lower with respect to that of the pure polymer as a consequence of powder loading. The comparison of critical factors in some cases presented different situations with respect to the literature review, i.e., the absence of the melt temperature between the feedstock critical factors.

POM SMp SMn SPn STp

mold temperature mold temperature holding time and mold temp mold temp, holding pressure, mold temp and holding pressure, melt temperature, and melt temp and mold temp

↓ ↑ ↓

■

a

The arrows indicate whether an increasing factor causes an increase (↑) or a decrease (↓) in shrinkage; − indicates no clear trend.

Corresponding Authors

*E-mail: [email protected]. *E-mail: [email protected]. *E-mail: j.r.alcock@cranfield.ac.uk.

Table 8. Comparison between Pure Polymer and Feedstock Critical Factorsa mold temperature holding pressure injection pressure holding time melt temperature

316L feedstock

POM

literature

yes yes yes no no

yes yes no yes yes

both both both only pure polymers both

AUTHOR INFORMATION

Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The authors thank Cranfield University for enabling conducting of this research. The mold was realized in the Ultra Precision machining laboratories: we want to thank J. Hedge and Dr. I. Walton for technical support during the manufacturing step.

“yes” indicates the presence and “no” the absence of the critical factor. The third column reports whether the factor affects are reported in the literature.

a

■

The 316L feedstock did not show as critical parameters either the holding time or melt temperature (both present in POM). This result was confirmed by the literature. The absence of the melt temperature is not simply explainable. However, a combined effect with the melt temperature is a second ranking effect for feedstock SPp (Figure 3). In a summary of the results reported above, it is possible to say that the presence of the mold temperature and holding pressure for both and the holding time only for the polymer was confirmed by the literature results. In contrast to literature indications, results have shown that the injection pressure is not critical for POM and the melt temperature is not critical for the feedstock. Regarding the injection pressure, it is likely that the powder loading caused the criticality of this parameter in the feedstock and not in the polymer. Regarding the melt temperature, further studies have to be performed even if in the literature only one work in nine identified this parameter as critical for the feedstock. Table 8 reports these considerations. The critical factors related to 316L feedstock and POM derived from Tables 4 and 7, respectively: the critical factors were compared with trends available in the literature.

REFERENCES

(1) Stanimirovic, Z.; Stanimirovic, I. Ceramic Injection Moulding, available at http://cdn.intechopen.com/pdfs/33648/InTech-Ceramic_ injection_molding.pdf (accessed July 2013). (2) Sotomayor, M. E.; Várez, A.; Levenfeld, B. Influence of powder particle size distribution on rheological properties of 316L powder injection moulding feedstocks. Powder Technol. 2010, 200 (1−2), 30− 36. (3) Nishiyabu, K. Micro metal powder injection molding, available at http://cdn.intechopen.com/pdfs/33647/InTech-Micro_metal_ powder_injection_molding.pdf (accessed June 2013). (4) Piotter, V.; Bauer, W.; Benzler, T.; Emde, A. Injection molding of components for microsystems. Microsyst. Technol. 2001, 7, 99−102. (5) Chen, R.-H.; Ho, C.-H.; Fan, H.-C. Shrinkage properties of ceramic injection moulding part with a step-contracted cross-section in the filling direction. Ceram. Int. 2004, 30 (6), 991−996. (6) Fanghui, L.; Chao, G.; Xian, W.; Xinyuan, Q.; Hong, L.; Jie, Z. Morphological comparison of isotactic polypropylene parts prepared by micro-injection molding and conventional injection moulding. Polym. Adv. Technol. 2012, No. 23, 686−694. (7) Liu, Y.; Song, M. C.; Wang, M. J.; Zhang, C. Z. Quality defects and analysis of the microfluidic chip injection molding. Mater. Sci. Forum 2009, 628 (629), 417−422. (8) Tay, B. Y.; Liu, L.; Loh, N. H.; Tor, S. B.; Murakoshi, Y.; Maeda, R. Injection molding of 3D microstructures by μpIM. Microsyst. Technol. 2005, 11 (2−3), 210−213. (9) Fu, G.; Loh, N. H.; Tor, S. B.; Murakoshi, Y.; Maeda, R. Replication of metal microstructures by micro powder injection molding. Mater. Des. 2004, 25 (8), 729−733. (10) Huang, C. K.; Chiu, S. W. Formability and accuracy of micropolymer compound with added nanomaterials in microinjection molding. J. Appl. Polym. Sci. 2005, 98 (5), 1865−1874. (11) Annicchiarico, D.; Attia, U. M.; Alcock, J. R. A methodology for shrinkage measurements in micro-injection moulding. Polym. Test. 2013, 32, 769−777. (12) Yang, S. Y.; Huang, C. K.; Lin, B. C.; Wei, W. C. J. Kneading and molding of ceramic microparts by precision powder injection molding (PIM). J. Appl. Polym. Sci. 2006, 100 (2), 892−899. (13) Binet, C.; Heaney, D. F.; Spina, R.; Tricarico, L. Experimental and numerical analysis of metal injection molded products. J. Mater. Process. Technol. 2005, 164−165, 1160−1166. (14) Weidenfeller, B.; Höfer, M.; Schillingb, F. R. Cooling behaviour of particle filled polypropylene during injection moulding process. Composites, Part A 2005, 36, 345−351.

5. CONCLUSIONS The influence of five processing parameters (injection and holding pressure, holding time, and melt and mold temperature) in terms of shrinkage for a 316L molding feedstock was investigated. The statistical treatment identified the critical factors that affected feedstock shrinkage: the mold temperature for postmolding shrinkage parallel to the flow direction, and the combination holding pressure−injection pressure for total shrinkage normal to the flow direction. The results were discussed by considering different aspects: factors affecting 316L feedstock shrinkage, a comparison between 316L feedstock and pure polymer shrinkage values, and a discussion by comparing each critical factor in pure polymer and feedstock. By considering feedstock and pure polymer, the comparison between shrinkage values parallel and normal to the flow 16566

dx.doi.org/10.1021/ie4040048 | Ind. Eng. Chem. Res. 2014, 53, 16559−16567

Industrial & Engineering Chemistry Research

Article

(15) De Santis, F.; Pantani, R.; Speranza, V.; Titomanlio, G. Analysis of shrinkage development of a semicrystalline polymer during injection moulding. Ind. Eng. Chem. Res. 2010, 49, 2469−2476. (16) Chen, C.-C.; Su, P.-L.; Chiou, C.-B.; Chiang, K.-T. Experimental Investigation of Designed Parameters on Dimension Shrinkage of Injection Molded Thin-Wall Part by Integrated Response Surface Methodology and Genetic Algorithm: A Case Study. Mater. Manuf. Processes 2011, 26, 534−540. (17) Shaa, B.; Dimova, S.; Griffithsa, C.; Packianather, M. S. Investigation of micro-injection moulding: Factors affecting the replication quality. J. Mater. Process. Technol. 2007, 12, 284−296. (18) Zhao, J.; Mayes, R. H.; Chen, G.; Xie, H.; Chan, P. S. Effects of Process Parameters on the Micro Molding Process. Polym. Eng. Sci. 2003, 43 (9), 1542−1554. (19) Ong, N. S.; Zhang, H.; Woo, W. H. Plastic Injection Molding of High-Aspect Ratio Micro-Rods. Mater. Manuf. Process. 2006, 21 (8), 824−831. (20) Thakur, V.; Angstadt, D. C. Influence of temperature on microfeature replication at ambient pressure in micro-molding. Annu. Tech. Conf.ANTEC, Conf. Proc. 2007, 5, 2936. (21) Zhao, J.; Lu, X.; Lin, M.; Chen, G.; Liu, S.; Yong, M. S. Effects of rheological properties of polymer blends on micro mold filling behavior. Mater. Res. Innovations 2006, 10 (4), 111−112. (22) Nor, N. H. M.; Muhamadb, N.; Ismail, M. H.; Jamaludinc, K. R.; Ahmadd, S.; Ibrahimd, M. H. I. Flow Behaviour To Determine the Defects of Green Part in Metal Injection Moulding. Int. J. Mech. Mater. Eng. 2009, 4 (1), 70−75. (23) Seokyoung, A.; Seong, J. P.; Shiwoo, L.; Sundar, V. A.; Randall, M. G. Effect of powders and binders on material properties and molding parameters in iron and stainless steel powder injection molding process. Powder Technol. 2009, 193, 162−169. (24) Wen, J.; Wen, P. Simulation and optimization of aspheric plastic lens injection molding. J. Wuhan Univ. Technol., Mater. Sci. Ed. 2005, 20 (2), 86−89. (25) Chang, T. C.; Faison, E. A study of the effect of process conditionson the shrinkage of plastic parts in injection molding by the Taguchi method. Polym. Eng. Sci. 2001, 41 (5), 703−710. (26) Annicchiarico, D.; Alcock, J. R. Review of factors that affect shrinkage of moulded part in injection moulding. Mater. Manuf. Process. 2014. (27) Vasco, J. C., Selada, A., Neves, T. and Pouzada, A. S. A study on the mouldability of POM micro details in moulding blocks using micro manufacturing technologies, available at http://repositorium.sdum. uminho.pt/bitstream/1822/20365/1/Vasco%20JC%202010.pdf (accessed June 2013). (28) Shen, Y. K.; Wu, W. Y. An analysis of the three-dimensional micro injection moulding. Int. Comm. Heat Mass Transfer 2002, 29 (3), 423−431. (29) Attia, M.; Alcock, J. R. Optimising process conditions for multiple quality criteria in micro-injection moulding. Int. J. Adv. Manuf. Technol. 2010, 50 (5−8), 533−542. (30) PlasticsInjection moulding of test specimens of thermoplastic materialsPart 3: Small plates; 2003; BS EN ISO 294-3:2003. (31) Minitab web page, http://www.minitab.com/en-US/products/ minitab/default.aspx?WT.srch=1&WT.mc_id=SE004815 (accessed Nov 2012). (32) Zhang, N.; Gilchrist, M. D.Characterization of microinjection molding process for milligram polymer microparts. Polym. Eng. Sci. 2013; available online, DOI 10.1002/pen.23677. (33) Heunisch, A.; Dellert, A.; Roosen, A. Effect of powder, binder and process parameters on anisotropic shrinkage in tape cast ceramic products. J. Eur. Ceram. Soc. 2010, 30, 3397−3406.

16567

dx.doi.org/10.1021/ie4040048 | Ind. Eng. Chem. Res. 2014, 53, 16559−16567