5.1 Experiment of adhesive distribution by changing initial distance
In order to ensure that the adhesive distribution experiment is not affected by environmental factors, the experiment is arranged in a closed space with room temperature of 25 ℃ and relative humidity of 50%. The control variable method was used in the experiment, and the initial distance was set as a variable. Other experimental conditions were unchanged, epoxy resin with a viscosity of 1000cPs was selected as the adhesive, ordinary glass capillary tube with an inner diameter of 0.8mm was selected as the capillary tube, clean glass slides with ultrasonic cleaning were selected as the base, select tungsten needle with tip diameter of 80µm as pipetting needle, the residence time was set to 2s and the stretching speed was set to 1.5mm/s, and only the initial distance of adhesive distribution was changed.
In order to ensure the quality of the transfer droplets packaging, it is necessary to test the height of the initial droplet and determine a reasonable initial distance range. A pipetting needle with a diameter of 80µm was used to complete 100 times of adhesive taking and the height of the initial droplet was measured. The measurement results of the initial droplet height are shown in Fig. 12. It can be seen from the figure that the height of the initial droplet is between 16.5µm and 17.8µm, and the average height is 17.09µm. Therefore, the upper limit of the initial distance was set to 15µm, and six groups of experiments were conducted. The initial distance of each group of experiments was set to 0µm, 3µm, 6µm, 9µm, 12µm and 15µm.
In order to reflect the universality of the experimental results, multiple sampling was carried out under the same parameters, and each group completed 100 times of adhesive distribution. The morphology of some transfer droplets under the condition of changing the initial distance is shown in Fig. 13. From the figure, the appearance of transfer droplets is round and there is no obvious defect.
The volume of each group of transfer droplets is measured by pixel method and the average value is taken. The contact area of each group of transfer droplets is calculated and the average value is taken according to the measurement results of the radius of transfer droplets. See Table 1 for specific data. It can be seen from Table 1 that the larger the initial distance, the smaller the volume of the transfer droplet, and the smaller the contact area of the transfer droplet. The average volume of the transfer droplets decreased from 47.27pL to 24.33pL, and the average contact radius of the transfer droplets decreased from 7607.76µm² to 4144.21µm².
Table 1
The related data of transfer droplet
Initial distance (µm)

0

3

6

9

12

15

Average volume of transfer droplet (pL)

47.27

39.08

31.93

28.51

26.13

24.33

Average contact area of transfer droplet (µm2)

7607.76

6889.67

5376.76

4639.71

4361.5

4144.21

The main reasons for the change of parameter configuration law are described as follows. Since the experimental adhesive and the pipetting needle are not changed, the volume of the initial droplet is almost the same after each adhesive taking, so the volume of the liquid bridge formed by the initial droplet contacting the base is almost unchanged. However, as the initial distance increases, the height of the liquid bridge increases, and the contact area between the liquid bridge and the base decreases. Since the volume of the transfer droplet is related to the viscous force [18], the larger the initial distance, the viscous force between the liquid bridge and the base will gradually weaken, the smaller the attraction of the base to the adhesive, so that the volume of the formed transfer droplet will gradually decrease.
5.2 Prediction and verification of parameter allocation law
From Section 5.1, it is concluded that the larger the initial distance, the smaller the volume of the transfer droplet, and the smaller the contact area of the transfer droplet. However, in the actual microelectronic packaging process, there are usually restrictions on the size and volume of the transfer droplet, so it is necessary to predict the relevant values before packaging.
In Section 5.1, many experiments have been carried out on the relationship between the initial distance, contact area and the volume of transfer droplet, and the experimental results are universal. Therefore, the data in Table 1 is used to predict the transfer droplet parameter configuration law. The prediction results of the parameter configuration law are shown in Fig. 14. Extract the relevant data points of the initial distance and the volume of transfer droplet and select the appropriate function and algorithm to perform nonlinear curve fitting on the data points. In order to make the fitting result reliable, 95% confidence band and 95% prediction band are set. After four iterations, the fitting converges, and the predicted results of the volume of transfer droplet are shown in Fig. 14(a). The same method is used to predict the law between the initial distance and the contact area. After five iterations, the fitting converges, and the prediction results of the contact area of the transfer droplets are shown in Fig. 14(b).
According to the fitting curve of the parameter configuration, the prediction Eq. (3) for the volume of the transfer droplet and the prediction Eq. (4) for the contact area of the transfer droplet are derived. In Eq. (3) and Eq. (4), V represents the predicted volume of the transfer droplet, S represents the predicted contact area of the transfer droplet, and x represents the initial distance.
In addition, if the parameter configuration conditions are not considered, the numerical relationship between the volume and contact area of transfer droplet can also be predicted in the adhesive distribution experiment conducted in this paper. The prediction results are shown in Fig. 15.
From the prediction results, the dispensing method proposed in this paper can roughly predict the volume and contact area of the droplet with a volume of 25pL ~ 47pL and a contact area of 4200µm2 ~ 7600µm2. This method can be used to predict the feasibility of microelectronic packaging and improve its reliability and efficiency. In order to facilitate the prediction of the VS value of the transfer droplet, the VS prediction Eq. (5) is derived according to the fitting curve, and the relevant data can be used.
$$S=19424.14945  1638.94464V+54.49034{V^2}  0.53114{V^3}$$
5
In order to verify the accuracy of the prediction results, the adhesive distribution experiment was designed to change the initial distance, and the initial distance was set to 1µm, 4µm, 7µm, 10µm and 13µm. The morphology of transfer droplets at different initial distances is shown in Fig. 16. Through observation and test, it is found that the dispensing method proposed in this paper meets the packaging requirements and has application value.
The same method as that in Section 5.1 is adopted for the test. Each group only changes the initial distance, and a total of 5 groups of experiments are conducted. Each group completes 100 sampling and takes the average value of the volume and contact area of the transfer droplets. See Table 2 for the specific experimental data.
Table 2
Comparison between predicted data and actual data
Initial distance (µm)

1

4

7

10

13

Predicted volume (pL)

44.03

36.18

31.01

27.59

25.34

Actual volume (pL)

43.88

36.21

31.13

27.41

25.06

Difference rate of volume (%)

0.34

0.08

0.39

0.66

1.1

Predicted contact area (µm2)

7562

6339.95

5062.97

4498.74

4261.7

Actual contact area (µm2)

7549.21

6341.12

5086.34

4456.27

4211.33

Difference rate of contact area (%)

0.17

0.02

0.46

0.95

1.2

It can be seen from Table 2 that the difference between the predicted value and the actual value is small, the average difference rate between the predicted volume and the actual volume is 0.51%, and the average difference rate between the predicted contact area and the actual contact area is 0.56%. Therefore, it is further proved that the proposed method can be applied to the actual microelectronic packaging.