Study of a dipping method for flip-chip flux coating
Introduction
Flux coating is one of the key processes during flip-chip packaging. Flux is applied on the solder bumps or substrate to remove oxides and pre-bonding flip chip on the substrate before reflow [1], [2]. Flux coating can be usually achieved through dipping flux, jetting flux and rotary fluxing [3]. Dipping flux has been a commonly used method in standard WLCSP (Wafer Level Chip Scale Packaging) and flip-chip processes, and flux is placed on rotary tray and scraped by a doctor blade in this method [4].
In order to obtain an ideal fluxing effect, dipping amount of flux must be strictly controlled, too small an amount of flux will cause a poor joint, and too much flux can impede the flow of underfill fluids. To control the dipping amount of flux precisely, the smooth of the flux surface must be ensured and the liquid level must be unchanged after repeated dipping. As to volatile flux, solvent of which is quite easy to volatilize, the flux solute content must also keep stable to avoid effecting bonding quality. Dipping method using rotary tray will inevitable cause the flux solute content increasing as the flux is placed on plate large area and exposed to the air prolonged. To reduce the evaporation of solvent, exposed area of the flux should be as small as possible.
Quality control is also an essential process in flux coating. Weighing can be used to control of film thickness quality [5]. But it presented several problems when use solvent-filled flux as the solvent is quickly to evaporate, and is become inapplicable as the film become thinner and less area. An optic flux thickness measure method using a unique dual-light interferometer is developing [6]. But it still has not good means to measure fluid in a groove quantitative.
Flux is a typical non-Newtonian fluid; mathematical models are usually established to predict rheological process [7], [8]. However, it is difficult to establish appropriate mathematical model to predict rheological process when liquid is conditioned in a groove, applied an inertia force, and influenced by properties of its own.
In response to these issues, this paper puts forward a gluing method for flux dipping, by which the flux is placed in a sealed container and transferred to the flux groove in a gluing way, makes the chip immerge into flux groove and dips just enough flux on chip solder bumps. This design can reduce the exposed area of flux and prevent the flux solvent from evaporation, thereby maintaining the stability of flux solids content; at the same time, it can avoid dropping of the liquid level effectively after repeated dipping by using one dipping one gluing approach. We also proposed an optical method to observe flux rheological process when it transferred to groove from container.
Flux flow process is recorded by an image acquisition system in this paper, images is processed based on an improved Otsu method [9], [10]. According to two indicators including reflective area ratio and flow velocity, flow process influenced by different gluing speed and viscosity is analyzed and contrasted in this paper. Through real chip experiment, the feasibility of this dipping method is verified.
Section snippets
Gluing method for dipping flux
Basic idea for this dipping method is placing the flux in a sealed container, flux is transferred into flux groove by a gluing way, and then makes the chip immerged into flux groove and dipping moderate amount of flux on chip solder bumps. The flow path of gluing method was shown in Fig. 1. This method includes the following processes. The flux container moves back and forth once on the plate, as shown in Fig. 1(a), so that the flux in container transferred to the flux groove on the plate; at
Reflective area ratio influenced by gluing speed and viscosity
Reflective area ratio is used to describe the flow process. The physical meaning of reflective area is the light source projection on the flux surface. This parameter can reflect the flow trend of the flux. For an image A with the size M × N, the gray level of the reflecting portion is higher and corresponding to the pixel value of 1 in the binary image, non-reflective part has lower gray level and corresponding to the pixel value of 0 in the binary image. Thus, we can obtain the reflective area
Conclusions
In summary, a dipping method for flip-chip flux coating is proposed. Flow process of the flux was investigated successfully by an optical detection method. Flow process influenced by gluing speed and viscosity is analyzed. Feasibility of the dipping method is verified through real chip experiment. The main conclusions are as follows:
- (1)
The changing of viscosity or gluing speed has almost no effect on the flow process if the gluing speed is ⩽5 cm/s, when the gluing speed is ⩾10 cm/s, the viscosity of
References (15)
- et al.
A Newton–Krylov finite volume algorithm for the power-law non-Newtonian fluid flow using pseudo-compressibility
J Non-Newtonian Fluid Mech
(2011) - et al.
Spreading of Boger fluid on horizontal surface
J Non-Newtonian Fluid Mech
(2013) - et al.
An improved median-based Otsu image thresholding algorithm
AASPRI Proc
(2012) - et al.
Characteristic analysis of Otsu threshold and its applications
Pattern Recognit Lett
(2011) - et al.
The pupil location based on the Otsu method and Hough transform
Proc Environ Sci
(2011) - et al.
Microstructural characteristics of Au/Al bonded interfaces
Mater Charact
(2007) - et al.
An investigation into the effects of flux residues on properties of underfill materials for flip chip packages
IEEE Trans Compon Pack Technol
(2003)