Dipping Process Characteristics Based on Image Processing of Pictures Captured by High-speed Cameras
- 1.6k Downloads
The dipping process was recorded firstly by high-speed camera system; acceleration time, speed, and dipping time were set by the control system of dipping bed, respectively. By image processing of dipping process based on Otsu’s method, it was found that low-viscosity flux glue eliminates the micelle effectively, very low speed also leads to small micelle hidden between the bumps, and this small micelle and hidden phenomenon disappeared when the speed is ≥0.2 cm s−1. Dipping flux quantity of the bump decreases by about 100 square pixels when flux viscosity is reduced from 4,500 to 3,500 mpa s. For the 3,500 mpa s viscosity glue, dipping flux quantity increases with the increase of the speed and decreases with the increase of the speed after the speed is up to 0.8 cm s−1. The stable time of dipping glue can be obtained by real-time curve of dipping flux quantity and is only 80–90 ms when dipping speed is from 1.6 to 4.0 cm s−1. Dipping flux quantity has an increasing trend for acceleration time and has a decreasing trend for acceleration. Dipping flux quantity increases with the increase of dipping time, and is becoming saturated when the time is ≥55 ms.
KeywordsDipping acceleration Dipping speed Dipping time Viscosity Image processing
Flip-chip technology has been widely used in high-performance and high-density microelectronics packaging [1, 2, 3, 4, 5, 6, 7, 8] due to shorter possible leads, lower inductance, higher frequency, better noise control, smaller device footprints, and a lower profile, such as smart card, laser emitting diode, and surface-acoustic-wave filter in telecom applications [9, 10, 11, 12, 13]. Flux coating is one of the key processes during flip-chip packaging; flux is applied on the solder bumps or substrate to remove the oxides, pre-bonding flip chip on the substrate before reflow, and increases the wettability of the solder bump and improve assembly reliability [14, 15, 16]. Usually, flux coating can be achieved through dipping flux, ultrasonic flux, printing flux, and so on . Chip to wafer (C2W) flip-chip bonding is more suitable at present for IC integration with expected process flexibility . A flux dipping method was applied to C2W bonding . Manna  and Nyamannavar et al.  reported that effect of fluxing chemical on wire surface by hot dip process and Heat Flux Transients at Solder/Substrate interface. The dipping quantity of flip-chip bumps affects the flip-chip bonding effect. Little flux can result in unbonding, excessive flux influence reliability of flip-chip.
However, from the literature, studies on the dipping process are quite lacking. Expected to develop new methods to observe the dipping process, adopt new ways to express the dipping quantity of tiny bumps and precisely control the dipping quantity.
In the paper, flux viscosity, dipping acceleration, dipping time, and flux stable time were firstly investigated using high-speed camera system and Matlab data processing. Based on the experimental images and data, the dipping process is discussed further.
3 Results and Discussion
3.1 The Influence of Flux Viscosity on Dipping
In order to compare the dipping effects of different viscosities, dipping experiments of 4,500 and 3,500 mpa s viscosity fluxes were carried out, respectively.
Step I is extracting image from the original images including image before dipping, flux separating image, and completed dipping image.
Step II is converting into a binary image using Otsu’s method. Otsu’s method is used to segment the image and extract image futures. Otsu’s method is an adaptive threshold determination method, and it is clustering based . The basic principle of the method was shown as follows.
Specifying the value of T from 0 to L sequentially, the number which makes σ2(T) has a maximum value is the threshold needed.
Step III is filling holes in a small area.
Step IV is identifying four holes’ center and fitting a straight line: firstly, finding out the centroid of reflective part on the solder bumps and, secondly, obtaining equation of the dividing line through X-direction coordinates of four centroids.
X-direction coordinate of dividing line can be calculated by after four bumps centroids are obtained, and the function expression of the dividing line can be represented by x = x0.
StepV is filling holes and calculating bump area, dipping process area, and dipping bump area.
3.2 The Real-time Characteristics of the Dipping
The average stable time at different speeds
Speed (cm s−1)
Stable time (ms)
3.3 Acceleration & Acceleration Time Effect on the Dipping
In the above experiment, acceleration time is set to 10 ms; accelerations of the speed (0.2, 0.8, 1.2, 1.6, 3.2, and 4.0 cm s−1) are 28.5, 115, 171, 228.5, 460, and 570 cm s−2, respectively. The relationship between acceleration value and dipping flux quantity is shown in Fig. 17.
3.4 Dipping Time Influence on the Dipping
The influence of flux viscosity on dipping is very obvious. Low-viscosity flux glue eliminates effectively the micelle. Dipping flux quantity of the bump decreases by about 100 square pixels when flux viscosity is reduced from 4,500 to 3,500 mpa s.
Real-time curve indicates the dipping process. The stable time of dipping glue decreases with the increase of the speed and has the same limit stable time as 80–90 ms when the speed is from 1.6 to 4 cm s−1. Dipping flux quantity increases with the increase of the speed in the beginning and decreases with the increase of the speed after the speed is up to 0.8 cm s−1.
Very low speed sometimes leads to the small micelle hidden between the bumps during dipping process, and the phenomenon disappearing when the speed is ≥0.2 cm s−1.
Dipping flux quantity has an increasing trend for acceleration time or has a decreasing trend for acceleration.
Dipping flux quantity increases with the increase of dipping time and is becoming saturated when the dipping time is ≥55 ms.
In summary, experimental results have provided important real-time data for developing dipping technology.
This work was supported by National Natural Science Foundation of China (No. 51275536) and the China High Technology R&D Program 973 (No. 2015CB057206).
- 13.K. Shen, W. Lin, D. Wuu, S. Huang, K. Wen, S. Pai, L. Wu, R. Horng, An 83 % enhancement in the external quantum efficiency of ultraviolet flip-chip light-emitting diodes with the incorporation of a self-textured oxide mask. IEEE Electr. Device L. 34(2), 274–276 (2013). doi: 10.1109/LED.2012.2228462CrossRefGoogle Scholar
- 19.J. Lu, H. Takagi, R. Maeda, Chip to wafer temporary bonding with self-alignment by patterned FDTS layer for size-free MEMS integration. Proceedings of IEEE Sensors Conference, 1121–1124 (2011)Google Scholar
- 23.Z. Fu, Z. Xie, Y. Zhao, Quality evaluation of adhesive coating on space solar cells based on image processing. Acta Energize Solaris Sonica 28(8), 844–848 (2007)Google Scholar
Open AccessThis article is distributed under the terms of the Creative Commons Attribution License which permits any use, distribution, and reproduction in any medium, provided the original author(s) and the source are credited.