Preparation of optimal feedstock for low-pressure injection molding of Al/SiC nanocomposite

2.1 Materials

Al and SiC powders were purchased from (Khorasan Powder Metallurgy, Iran) and (MKnano, Canada) companies and applied as metal and reinforcement powders, respectively. Prior to PIM, the starting materials were examined by densitometry, electron microscopy, and dynamic light scattering methods in order to determine their primary density, particle shape, and particle size distribution. Results of this analysis showed that the aluminum and silicon carbide powders have pycnometer density of 2.7 g/cm³ and 3.07 g/cm³, respectively, while the corresponding primary particle sizes are d50=28 μm and d50=40 nm, respectively. The morphologies of the powders are also depicted in Figure 1. The particle shapes of the powders were studied by means of Philips CM-200 FEG transmission electron microscopy (TEM) and S-4160 Hitachi SEM. Al powders were of micron sizes (average 10 μm) with spherical shape, while the SiC nanopowders were less spherical and of particle size ranging from 40 to 100 nm. A Fritch analystte 22 Particle size analyzer was utilized to measure the particle size distributions of the powders. The densities of the powders were determined by an AccuPyc 1330 helium pycnometer.

Figure 1

(A) SEM image of original Al micro-powders, (B) TEM image of SiC nanoparticles.

In this study, the wax-based binder was selected for PIM due to its facile debinding capability. The binder composition and the characteristics of its components are listed in Table 1.

2.2 Procedure

Aluminum powder (95 wt%) and silicon carbide powder (5 wt%) were mixed by a planetary ball mill. Stearic acid (2 wt%) was also added to the powders as a lubricant to avoid agglomeration of the powders during the milling. The ball-milling experiment was performed in a high-energy planetary ball mill at a rotation speed of 250 rpm in argon atmosphere with ball-to-powder ratio of 10:1 for 1 h using the chrome steel balls. The mixed powders were removed from the ball-mill chamber in the inert atmosphere to prevent aluminum oxidization.

Then the mixed powders were weighted as the values listed in Table 2 and were blended with the melted binder using a mixer at the temperature of 85°C. The mixing temperature was between the decomposition and melting temperatures of the binder components.

The viscosity and rheological properties of the feedstocks were determined using a plate-plate measuring system by means of Physica MCR 301 Anton Paar Rheometer. The rheological behaviors of the feedstocks were studied at the temperature of 85°C and a shear rate range of 0.1–1000 s^-1. This instrument has capability of temperature adjusting, and therefore the temperature dependence of reological parameters could also be investigated. The Weir model is also applied to find the optimum feedstocks.

3.1 Effect of powder loading on the rheological behavior of the feedstocks

Shear stress values versus shear rate of the prepared feedstocks are portrayed in Figure 2A. This figure shows that with increasing the shear rate, shear stress increases for all feedstocks. It can also be observed that the rates of increase associated with samples A and B are uniform, while some abrupt changes could be seen in the case of samples C and D in high shear rates. Since abrupt changes in shear stress are not desirable during injection molding, samples C and D are not suitable for injection.

Figure 2

The rheological behavior of the feedstocks. (A) The adaptability of feedstocks with power law model, (B) viscosities of samples as a function of shear rates.

The rheological behavior of feedstocks could be analyzed with the power law model to determine their type of flows:

(1)τ=k(γ˙)n (1)

where k is a constant, τ denotes the shear stress, γ˙ is the shear rate, and n denotes the flow behavior index. The parameter of n is an indicator of shear sensitivity. If n<1, the behavior is pseudoplastic, and if n>1, the feedstock behavior is dilatant. Also, Newtonian behavior is characterized by n=1. The analyses of the rheological behavior of all samples with the power law are summarized in Figure 2A and Table 3. It could be seen that all samples show pseudoplastic behavior, and the n value of sample A is the lowest. This indicates that sample A is the best sample for injection molding as the shear sensitivity of this sample is lowest.

The viscosities (η) of the feedstocks as a function of shear rates have also been depicted in Figure 2B. Typically, the viscosity of samples with higher powder loading should be higher than the others, while some deviation from this perception could be observed in Figure 2B. These deviations especially in low shear rates may be due to the lubricant nature of SiC particles as well as their texturing, which decreases the viscosity in feedstocks with high loading (sample D).

3.2 The homogeneity of the feedstocks

As mentioned earlier, one of the major concerns in injection molding process is the homogeneity of feedstock. One way of dealing with this is the observation of its torque value. In this regard, the homogeneity of feedstock can be predicted by observing the torque variation during mixing versus time. Uniform mixing was assumed to occur when the torque reached a steady state value [18]. The changes of torque value of feedstocks versus time are shown in Figure 3A. Results of this figure indicate that the stability of feedstock has been decreased by the increase of powder loading. It seems that we had the stability decrease after having the separation of the binder and powder. For further investigation, changes of the feedstock’s viscosity with time are shown in Figure 3B. It can be seen that the viscosity first lowered and then went up in the case of sample C. The behavior could be related to the loss of structures which results in the reduction of viscosity in the first step. But as time passed, reformation of the structures occurred and the viscosity rose. In feedstock B, the viscosity increased with time until it reached a stable point. The phenomenon may be attributed to the formation of structures in the feedstock which could cause hard shearing in it. Since the stability of viscosity, even over short periods of time, is necessary for the mold to be filled, feedstock A is the desirable one for PIM due to its stable viscosity.

Figure 3

The estimation of feedstocks’s homogeneity. (A) Variation of feedstock’s torques and (B) variation of feedstock’s viscosity, respectively.

The SEM micrograph of feedstock A is illustrated in Figure 4. This image illustrates that the powder and binder were uniformly mixed in this feedstock without any bubble which verifies the results of the rheometry analyses.

Figure 4

The scanning electron micrograph of feedstock denoted as A in Table 2.

3.3 Weir model

In order to select the most suitable powder loading in injection molding by rheometry results, the Weir model [19], which is proposed for polymers and semisolid materials, could be used. Equation (2) portrays the mathematical representation of this model.

(2)α=1η0|∂logη/∂logγ˙|∂logη/∂(1/T) (2)

In Equation (2), α is known as Weir model’s constant and also moldability index, η denotes the viscosity, η₀ is referred to as a reference viscosity, and T and γ are temperature and shear rate, respectively. The higher value of moldability index (α) is pleasing since feedstocks with low η values are prone to the separation of powder and binder [16, 19, 20].

The moldability index for different feedstocks at the shear rate of 100(1/s) and the temperature of 85°C were calculated as a stability indicator of those, and the results are reported in Figure 5. Considering these figures, sample A has the highest moldability index, and therefore it would be chosen as the desired critical powder loading.

Figure 5

Evaluation of alpha values to determine the optimal powder loading.

3.4 Calculation of flow activation energy

Viscosity of the feedstocks depends on temperature by Arrhenius’ equation:

(3)η=η0exp (E/RT) (3)

where η₀ denotes viscosity at reference temperature, E is flow activation energy, R is the universal gas constant, and T denotes the temperature. The E value is indicator of the sensitivity of viscosity to temperature. If the value of E is low, the viscosity is less sensitive to temperature variations, and this results in fabrication of defect-free parts. The slopes of Ln (η) versus 1/T for all samples, as shown in Figure 6, are equal to the flow activation energy (E). The obtained results are also listed in Table 4. It can be seen that sample D has the lowest E value, but the regression coefficient of linear fit is low (<0.9), which indicates the high uncertainty of E value calculated for this sample. Omitting this sample, sample A has the lowest E value which indicates that the viscosity of this sample has less sensitivity to temperature. Therefore, it is suitable for injection molding in industrial process where the temperature control is not easy and high stability of viscosity is crucial.

Figure 6

Plot of Ln(η) versus 1000/T for the prepared feedstocks.