Mechanism and solutions of appearance defects on microfluidic chips manufactured by UV-curing assisted injection molding

2.1 Generation mechanism of bubble defects

Bubble defects could be captured by camera through visualization quartz mold insert. The mixture resin was loaded into the storage cylinder for 4 h to vent gas. Smooth pipeline and good sealing performance were ensured. The pipelines were pre-filled completely before injection molding. The average degree of functionality of the mixture resin was 3.0. The holding pressure was set at 0.2 MPa after filling stage. Meanwhile, UV-curing started with irradiation intensity of 1000 mW/cm². The generation process of bubbles was recorded and displayed in Figure 3.

Figure 3:

Bubble defects generation process of visualization: (A) at the beginning of UV light when mixture resin fully filled cavity, (B) after 16 s of UV light, (C) after 32 s of UV light, (D) after 70 s of UV light no more bubbles.

Figure 3A–D represents the beginning of UV light irradiation at 0 s and the duration of 16 s, 32 s, and 70 s, respectively. It could be seen that no bubbles occurred at the beginning. After that, bubbles appeared in the center of cavity when the mixture resin began to cure, which showed a tendency to the edge. As shown in Figure 3D, no more bubbles increased at that moment because the resin was cured totally. Due to the point UV light source employed in this study, the irradiation energy at the center was relatively higher than that at the edge, resulting in bubbles generated at the center.

Visualization results showed that bubbles always appeared on the surface close to the mold side, while the opposite side surface near the glass was smooth without any defects. It could be explained that the UV light irradiates the resin from the top surface to the bottom of cavity. Curing reaction occurred on the top surface firstly, which caused shrinkage along the vertical direction in undesirable packing stage. Because there was no gas produced during curing reaction process, it could be inferred that the bubbles were caused by the shrinkage of ultraviolet-curing resin during solidification. Previous dynamic test results utilizing dilatometer revealed that the volumetric shrinkage of various ultraviolet-curing resin could reach 2.91%–6.96% [17]. Polarization detection results as shown in Figure 2C indicated that residual stress appeared around the bubbles in shades of brown. There were still certain shrinkage stress areas in the field of vision because of the brown bubbles, but we cannot see any bubble hole without polarizing optical instrument, which was a potential shrinkage bubble hole.

2.2 Reaction principle

The principle of polyurethane acrylate UV light curing is as follows:

Photoinitiator-184 cleavage reacts firstly to produce benzene formyl radical and hydroxyl cyclohexyl radical, which is polymerization initiated by free radicals such as that of the formula (1). The addition reaction between free radical and double bond generates another free radical, such as the reaction of the double bond in the free radical and the polyurethane acrylate in formula (2). After benzoyl radical reacts completely, product in formula (2) continues to react with the excess polyurethane acrylate follow formula (3). The reaction route of hydroxyl cyclohexyl radical is shown in formula (4) and formula (5), which keep the identical reaction principle mentioned above until the complete reaction of polyurethane acrylate. Then a polymerization reaction occurs among free radicals following formula (6). It can be found clearly that there is no gas generated in the whole reaction process.

3.1 Experimental equipment

Our laboratory has developed UV light curing injection molding machine (shown in Figure 4) including injection cylinder, mold clamping unit, light radiation equipment, storage tanks, and pneumatic-electric control system. The dimension of the prototype is 1300×600×400 mm. The mold clamping unit and injection unit are both driven by air cylinder. The maximum clamping force is up to 14 kN. Injection and holding pressure could increase from 0 to 3.2 MPa. The maximum injection rate could reach 150 mm³/s. The dimension of molding tool applied in this study is 196×160×35 mm. The photography of the cavity and the drawing of the microfluidic chip are shown in Figure 5. A point UV light source with a diameter of 12 mm (purchased from Shenzhen Ulamp Co. Ltd., China) was employed, and the power of UV light source can vary from 0 to 1000 mW/cm². The wavelength of UV light is 365 nm.

Figure 4:

The photo of UV light curing injection molding machine: 1. clamping unit, 2. storage tank, 3. injection unit, 4. UV lighting controller, 5. electropneumatic controller.

Figure 5:

The structures of the mold cavity and product: (A) mold cavity, (B) product.

UV light curing injection molding consists of the following stages: 1. Store the mixture resin in the storage tank, and deliver it into the injection unit after metering. 2. Injection piston was actuated by air cylinder to inject the mixture resin to the cavity. 3. Turn on the UV light source for irradiation, the injection piston continues to provide holding pressure for feeding in the whole process of the reaction. 4. Open the mold when the resin cured completely to obtain product.

3.2 Experimental procedure

3.2.2 Effect of irradiation pattern on the bubble defects

In order to avoid the premature curing of the mixture resin around the gate, which would lead to decrease of feeding capacity, a dynamic irradiation pattern was proposed. A moveable curing light source was carried out and scanned following the path shown in Figure 6, which could effectively control the curing time and improve the filling behavior in holding stage. The average function degree of the mixture resin was 2.0. The viscosity was 3300 cP at 20°C, and holding pressure was set at 0.6 MPa. The intensity of irradiation was 100%. The experimental results of the dynamic UV-curing irradiation pattern were compared with the ones of stationary pattern.

Figure 6:

Dynamic irradiation mode and scanning path.

4.1 Effect of injection molding parameters on the bubble defect

4.1.1 The effect of pressure on the number of bubbles

Experimental results showed that the average mass of the products with holding pressure at 0.6 MPa, 0.4 MPa, and 0.2 MPa was 5.65 g, 5.62 g, and 5.58 g, respectively. The mass decreased with the decrease of the holding pressure, but the range was not large. This kind of effect was similar to the effect of traditional injection molding. The surface quality of the product and the micro morphology of the bubble were shown in Figure 7.

Figure 7:

Holding pressure at 0.6 MPa, the surface quality and the micro morphology: (A) appearance of products, (B) local magnification, (C) micro morphology of bubbles, 100×, the diameter of the bubble is about 300 μm.

Table 5 lists the distribution of bubbles according to diameter. The decreasing of feeding capacity caused by holding pressure led to the significant increase in the amount of surface bubbles, especially in smaller diameter (d<150 μm) bubbles. These tiny bubbles affect the appearance quality more obviously. In other word, the bubbles could be improved through increasing holding pressure.

Table 5:

The number of the bubbles and distribution.

Group	Sum	d≥300 μm	300 μm ≥d≥150 μm	150 μm≥d≥75 μm
1	110	20	30	60
2	190	40	40	110
3	300	40	70	190

4.1.2 The effect of irradiation intensity on the number of bubbles

Figure 8 showed the photo of bubble defect under different irradiation intensity. It was indicated that the bubbles in central area were dense and the bubble diameter decreased along the radial direction. Figure 8A–C showed the bubble holes around “cross” structure under irradiation intensity of 100%, 50%, and 25%, respectively. Obviously, the number of bubbles decreased along with the decrease of irradiation intensity.

Figure 8:

Effect of irradiation intensity on the number of bubbles: (A) the bubble holes around “cross” under 100%, (B) the bubble holes around “cross” under 50%, (C) the bubble holes around “cross” under 25%.

The average bubbles sum of the every group was 270, 160, and 60, which decreased with the decrease of irradiation intensity apparently. However, the bubbles of large diameter (d≥150 μm) still existed in the low irradiation intensity, as shown in Figure 8C. Irradiation intensity directly affects the cure rate. The low curing rate makes the curing duration long enough for feeding. Hence, the shrinkage and the resulting residue stress become smaller. The number of bubbles reduced as well. Although the bubble defect could be improved by reducing the irradiation intensity, the effect is limited. Moreover, lower irradiation intensity results in longer the molding cycle. The results showed that the cycle time was 70 s at irradiation intensity of 100%, 100 s at 50%, and much more time at 25%. Considering the forming efficiency, too low irradiation intensity is unacceptable in practical production.

4.1.3 The effect of material viscosity on the bubble defect

The experimental results (Figure 9C, F, I) showed that the bubbles decreased obviously with the dropping of material viscosity. The average diameter of bubbles reduced from 300 μm to 180 μm as well. As shown in Figure 9B, E, and H, it was demonstrated that the lower residue stress could be achieved at higher material viscosity. Higher viscosity mixture resin with higher adhesion results in lower curing rate and longer cycle time, which provide sufficient feeding capacity. Consequently, the stress concentration and vacuum bubbles decreased. However, the high viscosity material may influence the dimensional accuracy of micro structure. Therefore, the improvement method through adjusting the viscosity did not apply to production of microfluidic chips.

Figure 9:

Experimental results of material viscosity on the number of shrinkage bubbles. (A, D and G) The number of bubbles in the product decreases with the decrease of the viscosity of the material. (B, E and H) The residual stress of the product varies with the increase of the viscosity of the material. (C, F and I) Local magnification of the picture A, D and G.

4.2 Effect of irradiation pattern on the bubble defect

Figure 10 showed the different products obtained by different irradiation patterns under the identity process conditions and formula. Compared with the part by stationary irradiation pattern as shown in Figure 10, the apparent bubble defects obtained by dynamic irradiation pattern were obviously eliminated.

Figure 10:

Products by scanning mode of light (B) and fixed light (A).

Figure 11 shows the micrograms on the equal position of the products gained via different irradiation patterns. The bubble defect becomes hardly invisible due to the dynamic irradiation pattern. With a laser microscope on both surfaces of roughness measurement, the blue line in Figure 11 was employed for line roughness sampling trajectory; the results showed that the rough degree Ra value was approximately 118 nm through the dynamic irradiation pattern, while Ra value was around 562 nm through stationary pattern. It further demonstrated that the apparent bubble defects were eliminated and greatly enhanced the appearance quality and transparency. It could be attributed to the difference of feeding process. The feeding direction is from the plane of mold core to the glass by stationary pattern, while it is along length direction from the gate to the internal core by dynamic pattern. It indicated that the feeding direction along the length is more suitable for method of UV-curing injection.

Figure 11:

Surface structures under the microscope: (A) the surface of product obtained by fixing irradiation, (B) the surface of product obtained by scanning irradiation.

4.3 Preparation of microfluidic chips of cross flow channel

We have successfully fabricated microfluidic chips by employing UV-curing injection molding method. As shown in Figure 12, the microstructure basically meets the design requirements. As a contrary, we used the same mold core insert to fabricate microfluidic chips by traditional injection molding method with PMMA. The contour of the microstructure was compared with the one fabricated by UV-curing injection molding in Figure 13. It could be found that the dimensional accuracy of the product by UV-curing injection molding was higher than that by traditional injection molding. Considering the cost, efficiency, and molding precision, the method of UV light curing injection molding revealed remarkable advantages for manufacturing of microfluidic chip. Preliminary fluid mixing experiments were carried out successfully by applying the microfluidic chips we prepared in this paper (Figure 12B–C).

Figure 12:

Microfluidic chips and test: (A) microfluidic chip successfully bond, (B) microfluidic chip has been used in fluid mixing experiments, (C) microscopic photo of the flow channel.

Figure 13:

Micro channel profiles from laser scanning microscope.