*3.2.2 Process simulation and validation*

In order to better understand the connection between product quality and process parameters, process simulation of injection molding has developed over several decades. Simulation in the early stages of part and mold design is relatively cost-efficient and offers the capability to evaluate various design options, such as runner design and gate designs. However, microscale effects, such as altering heat transfer coefficient (HTC), wall slip behavior, mold surface roughness, venting operations, which tend to be ignored in conventional injection molding simulation, should be considered in the simulation of microinjection molding effectively [52–54]. Another important concern is that most studies only used nominal machine processes, but they do not include the actual machine dynamics, which is important for microinjection molding of surface structures because microscale surface structures are much sensitive to process variation. As a result, simulation results are more or less unconvincing and cannot be adapted for real-world applications, and the simulation inaccuracy needs to be addressed.

#### **Figure 11.**

*Filling mechanism of surface structure on a substrate: (a) the injection stage, (b) the end of the injection stage, (c) the packing stage [50]. Copyright (2018) with permission from IOP publishing, LTD.*

Based on an experimental study on the filling behavior of microfluidic surface structures (**Figures 10(b)** and **11(a)**), the real responses of the injection molding machine are acquired and adopted in the process settings of the simulation with the help of process monitoring [55]. In addition, the effect of microscale sensitive parameters on the replication of surface structures using simulation is systematically studied and validated by flow front profile, cross-sectional profile, and replication of the structures. Consequently, the combination including a relatively higher heat transfer coefficient (30,000 W/(m2·K)) of the injection stage, standard atmospheric pressure (0.1 MPa) as the initial air pressure of venting, 0.7 as the friction coefficient for wall slip and a freezing temperature of 20 degrees above the glass transition temperature is selected. In terms of the flow cytometer surface structures, replication defects in experiments (circled in **Figure 12(a)**) are successfully predicted after the optimization as the blue parts shown in **Figure 12(c)**. Besides, the insufficient replication of the droplet cylinders (the areas in white in **Figure 12(c)**) is also predicted after the selected parameters are applied.
