An Experimental Approach to ore-free ow Soldering

Transcription

1 148 IEEE TRANSACTIONS ON COMPONENTS, PACKAGING, AND MANCTFACTURING TECHNOLOGY-PART B, VOL. 19, NO. 1, FEBRUARY 1996 An Experimental Approach to ore-free ow Soldering D. J. Xie, Y. C. Chan, Senior Member, IEEE, and J. K. L. Lai Abstract-This paper addresses the mechanism of pore growth and reports the results of an investigation on pore-free reflow soldering. An experimental approach to obtain pore-free solder joints has been successfully developed. The method employed is to split the solder joints during the reflow soldering process. For the Sn63Pb37 solder paste with metal content of 90%, the solder joints produced by this method have no detectable pores compared with 7.5% area fraction of pores in normal IR reflow soldering. The joint strength also increases by about 2040% as compared with that of normal solder joints. This method has promising applications, especially in the nitrogen reflow soldering technique, to yield pore-free and robust solder joints. A model for the pore growth is also proposed in the paper. It shows that flux residues usually occur together with the pores and are important to the pore development. Once entrapped by the molten solder, the pores and inclusions have difficulty escaping from the joints unless external forces or disturbances are applied. Index Terms-Porosity, nism, reflow soldering. pore-free, solder joint, growth mecha- I. INTRODUCTION OLDER PASTES as joining materials are widely used in electronics assemblies. When using solder pastes, however, one major type of defect found in solder joints is porosity. Porosity is easily formed because a solder paste normally contains volume percentage of volatile materials (e.g., flux and solvent). A lot of flux fumes are produced during reflow soldering due to thermal decomposition. Some gases may be entrapped between the flat surfaces of printed circuit boards (PCB) and component lead, and cause pore formation in solder joints upon cooling. Normally, there are three ways to control the porosity in solder joints: a) Controlling the source of gases, e.g., using high metal content solder pastes; this is not practicable due to constraints in the printing process. b) Controlling the soldering process, e.g., using high temperature to promote the out-gassing of entrapped gases or to the contrary, using low reflow temperature to decrease the vaporization andlor decomposition rate of fluxes. This method is effective in decreasing the porosity to some extent. However, the method to determine the optimum reflow temperature profile for different solder pastes has not yet been resolved [l]. c) Using special reflow soldering. It has been shown effective to decrease the porosity by nitrogen reflow [2] or vacuum reflow [3]. These methods have been popular recently and are helpful to obtain good solderability, since there are few problems associated with oxidation. In order to decrease the porosity, the mechanism of pore formation has been studied extensively [4], [5]. However, another important issue in dealing with the porosity, the mechanism of pore growth, has been relatively unexplored. Under what circumstances could a pore exist or grow in solder joints? Can the porosity be eliminated completely? These questions should be resolved to control the porosity. This paper attempts to investigate the mechanisms of pore formation and growth in order to find a sound method to control porosity. Experiments are conducted on the same specimen with different IR reflow cycles. The pore evolution at different reflow cycles is critically observed $y both X-ray radiography and fractography. II. EXPERIMENTAL METHODS Solder paste used in the experiments was composed of 90% by weight of 63Sd37Pb eutectic alloy balanced with rosin and solvent. In shear strength testing, however, solder paste with a metal content of 86% was also used. The specimen preparation was the same as described in an earlier paper [5]. Briefly, a FR-4 PCB plate with 14 copper pads was used in the experiments. Each copper pad had an qea of 1.25 x 2.5 mm2. Solder pastes were deposited to the pre-cleaned copper (Cu) pads by means of stencil printing. A shear specimen was made by soldering two PCB's together, which simulated the surface mount joints in electronic assembly. The specimen was then reflowed in the IR reflow machine (type: HELLER 932). The maximum measured temperature of the solder joints is 210 C with dwell-time of 60 s between 180 to 210 C. The total time in the reflow zone used in the experiment is 90 s. The porosity was detected non-destructively by X-ray radiography. Besides, it was also observed on the fracture surface after shear testing by stereo microscopy or scanning electron microscopy (SEM). The shear test was done in an Instron 4206 universal testing machine PORE FORMATION AND FLUX RESIDUE Manuscript received March 30, 1995; revised September 27, This - - Grant The authors are with the Department of Electronic Engineering, City Umversity of Hong Kong, Kowloon, Hong Kong. Publisher Item Identifier S (96) work was supported by the City University of Hong Kong under Research *' Residue Behavior in the 'Older Joints /96$ JEEE Figure 1 gives the enlarged X-ray radiographs showing typical pores in the shear solder joints. The corresponding fractographs obtained by stereo microscopy after the shear

2 XIE et al.: AN EXPERIMENTAL APPROACH TO PORE-FREE REFLOW SOLDERING m Fig. 3. SEM fractograph of SOIC-14 joints showing flux residues appearing in and around pores. Fig. 1. X-ray radiographs showing pores observed in the shear specimen (copper pad length = 2.5 mm). Joint Joint m H hint Joint 2 Joint t Fig. 2. Fractographs of the shear joints in Fig. 1 (stereo microscopical photographs). 0 testing are illustrated in Fig. 2 for comparison. The pores are denoted as white areas both in X-ray photographs and in the fractographs. It is shown that the configuration and distribution of pores obtained in these two figures are very close to each other. The X-ray radiograph is in good agreement with the fractograph in the observation of porosities. As shown in these photographs, there were a lot of pores in the shear solder joints. The average area fraction of pores was about 7.5%. (C) Fig. 4. Pore growth in the shear specimen at different reflow cycles. (a) First reflow; second reflow; (c) third reflow. The biggest curvature radius was around 0.25 mm and most of them were circular or elliptical shaped in the plan view. More importantly, it is worth noting in the fractographs that some flux residues occurred near the pores. The flux residues

3 150 IEEE TRANSACTIONS ON COMPONENTS, PACKAGING, AND MANUFACTURWG TECHNOLOGY-PART B, VOL. 19, NO. 1, FEBRUARY 1996 wetting and spreading of molten solder fkux residue layer compact wetting layer cu pad Fig. 6. Cross-section view of the shear joint showing a pore entrapped by two PCB's. Eq&d solder 0 ~ ~ 8 ~ fiuxresidue $ i ~ solder particle Fig. 5. Schemahc magrams of sphencal and ellipsoidal pores formed 111 the shear specimen. Note: PCB's are not shown. (a) Melting began. Arrows show the moving mrections of molten solder, porosity and flux residue respectively; melting completed. A big pore with flux residues inside is formed. appeared dark in Fig. 2, while they were of rosin color during real-time observation. This shows that the flux residues were easily entrapped by the solder. The flux residues were also commonly observed in practical surface mount joints. Fig. 3 is an SEM fractograph of a fatigued SOIC-14 solder joint which was taken directly from the circuit board of a used mobile phone. In Fig. 3, the flux residues appeared rod-shaped in or around the pores. B. Pore Growth Model A pore may grow in size during the reflow process once it is formed in the solder joints. X-ray photographs in Fig. 4 illustrate pore growth obtained at one to three reflow cycles for the same specimen. In normal reflolw soldering, only one reflow cycle is used. As shown in the figure, a large number of pores were observed with all three different times of reflow soldering. The variations of distribution and configurations of the pores may be summarized into three modes of pore growth as follows: i) Growing (see Joint No. 1): Some small pores in Fig. 4(a) were incorporated into a large one [Fig. 4] or a new pore formed [Fig. 4(c)]. ii) Stable (see Joint No. 2, 3, and 5): Pores remain unchanged in size and configuration. iii) Escaping (see Joint No. 4): Some pores agglomerate to escape as is shown in the variations from Fig. 4(a) and. This is also considered as a special case of growing mode. It should be noted that no pores in the joints in Fig. 4 split into small ones or contracted during further reflow cycles. Fig. 7. Pores agglomerated at the surface of the non-wettable PCB. (a) Surface wew showing the pores no porosity found inside the photograph, surface pores have been rejected and not shown) Growth mechanism of a pore is controlled by the pressure equilibrium and takes the following two forms if the effect of shrinkage is neglected [6] Stable: Growth: Pg = Ph + P, + y(l/rl + ~ /Tz) Pg > Ph + P, + y(1/?"1 f I /Tz) where Pg, Ph, and P, are the pressure for internal gas, hydrostatic and atmosphere, respectively; y represents the surface tension in a bubble; r1 and rz are the principal radii

4 XIE et al.: AN EXPERIMENTAL APPROACH TO PORE-FREE REFLOW SOLDERING 151 P.. 4! :?D %I : r111 mi : 83 $ f : Qii Fig. 9. The solidification surface on the remaining half of the splitted shear joints. (a) Surface view showing a big pore; X-ray radiograph showing no porosity inside the $2 SAY: $4'1 M: M: Ilii (C) Fig. 8. Effect of splitting on the pore formation (X-ray radiographs). (a) Normal IR reflow soldering; pore-free joints obtained by splitting; (c) second reflow of. of curvature of the free surface. Pg is contributed directly by the evaporation of flux and solvent and depends greatly on the temperature. Ph and Pa can prohibit the gas bubble from growing. When reaching equilibrium at a certain temperature, the fluxes in a pore stop evaporating and the pore also stops growing. The pore may increase in size with the progress of gas evaporation when the reflow temperature increases. Since the up floating pore is stopped by the top PCB, the pore is unable to reach the free surface of the joint. Therefore, the pore grows horizontally after reaching the maximum vertical length. The pore size may vary from zero to the pad width or length according to its gas volume and the temperature distribution of the solder joint. Iv. MECHANISM OF PORE FORMATION According to the above observations, the mechanism of pore formation is summarized schematically in Fig. 5. At the beginning of melting, solder melts and wets the copper pads which have been cleared by the flux. On the other hand, the liquid solder is unable to wet the flux residue which is composed of resin acid, copper dioxide and their combination [7], [SI. The molten solder spreads on the copper surface by pushing the flux residue away so that a compact wetting layer is formed. The flux residue and gases formed move away from the copper pad surface and may encounter the molten solder coming from the upper side. Some of them may be pushed horizontally by the disturbance of the fluid and escape to the free surface. Some may be entrapped in the solder and remain as inclusions and pores after freezing. As seen in the fractographs in Figs. 2 and 3, flux inclusions usually occur together with the pores. This is because the inclusions are easily pushed to the interfaces where the pores are often concentrated. Less surface energy is required if the pore and flux residue grow together. Prolonging reflow time or increasing reflow temperature may enhance the evaporation of gases and the growth of the pore. The pore can escape if it grows big enough to the edge of the solder joint. In most cases, however, a pore will stop moving vertically and can only grow horizontally when it reaches the compact wetting layer at the top or bottom. Few pores can grow to escape from the solder joint. Fig. 6 is a cross-section view which shows a pore growing up to take the whole width (0.19 mm) of the gap between two PCB's. The pore grew along the copper pads to about 0.6 mm. It is also shown that the wetting layer is not smooth and integrated, indicating the bad wettability of the copper pads. The pores and inclusions are often stable at those pads.

5 152 EEE TRANSACTIONS ON COMPONENTS, PACKAGING, AND MANUFAC"G TECHNOLOGY-PMT B, VOL. 19, NO. 1, FEBRUARY 1996 A 2% X XXX A AA. X A- spiitted x -normal 86% Metal content by weight in the solder paste. 86% 90% Metal content by weight in the solder paste (a) Fig. 10. Effects of splitting on area fraction of pores and shear strength of the solder joints. (a) Area fraction of pores; shear strength. If the upper PCB is nonwettable, e.g., it has no copper pad on its surface, the solder cannot wet the surface. In this circumstance, the flux inclusions and pores may float up and agglomerate at the top surface as the inclusions can wet the PCB board. Therefore, the porosity and inclusions occur at the surface but not inside the solder. This is confirmed by the observations shown in Fig. 7. Fig. 7(a) shows a surface between solder and a non-wettable board, where a lot of surface pores and flux residues appeared. At the same time, there was no porosity inside the solder as shown in the X-ray radiograph of Fig. 7. This implied that it was possible to eliminate the porosity entirely in the solder joint. v. ELIMINATING POROSITY BY SPLITTING An attempt has been made to ge rid of the pores. From the above discussion, the porosity would escape from the solder if placed under turbulence. One way to do this was to split the joint for a very short time during IR reflow soldering. Splitting was done in the cooling zone at a temperature over the melting point of the solder. The results are shown in Fig. 8. As noted from Fig. X(a), a lot of pores are present in the shear specimen under normal IR reflow conditions. The pores were almost completely eliminated after the splitting was done in the shear specimen [Fig. 8]. The solder joints had very few pores after reflowing twice, which showed that most of the inclusions had also escaped during the splitting as shown in Fig. 8(c). This is because the pores and inclusions agglomerated at the center escape easily to the edge of the joints during splitting. Figure 9(a) is an in-situ surface view of the remaining half of a split solder joint, where a big gas hole and some flux residues are detectable. In Fig. 9@), there was no porosity inside the solder joints after splitting. This indicates that all the pores have floated out of the solder. Therefore, splitting is an effective way to eliminate porosity in solder joints. Besides, splitting is useful to increase the cooling and freezing rates of the solder. That helps to eliminate shrinkage pores in the center of the joints and promotes the formation of a fine grain structure and a higher joint strength. Fig. 10 shows quantitative effects of the splitting on the area fraction of pores and shear strength. Two levels of metal content in the solder pastes, 86% and f 90%, were used. As shown in Fig. lo(a), the area fraction of pores for solder paste with 86% metal content was about 9.3% in the normal joints and became 0.8% after splitting. For solder paste with 90% metal content, it was in the range of 2-7.5% in the normal joints and became 0 N 0.3% after splitting. As shown in Fig. 10, the strength of the solder joint has increased greatly due to splitting. The average shear strength has increased by 22% and 40% for solder pastes with metal contents of 90% and 86%, respectively. During reflow soldering, the splitting time should be short enough so that no oxidization or solidification of solder occurs. It was found that there is difficulty in soldering the joint when the splitting lasts longer than 4 s. When soldering joints with many pins, a machine hand, similar to the pick and place machine, is required to pick the component up to ensure a good alignment between PCB leads and copper pads. The joint quality should be further improved if nitrogen IR reflow soldering is used, where few oxidation problems would occur. This implies a promising improvement in the present reflow method and is worthy of m e r investigation. VI. CONCLUSIONS An effective approach to obtain a pore-free joint in a conventional IR reflow soldering is proposed. The method is based on the knowledge of the pore growth mechanism. Pores usually grow together with flux inclusions and are hard to escape once they are entrapped in solder joints. This is because the forces on the pores, both external and internal, create an equilibrium which prevents pores from moving vertically or horizontally. Pores remain stable or grow in size during reheating, but few pores can grow large enough to escape from the joint. Prolonging the reflow time in normal IR reflow soldering cannot decrease the area fraction of pores. A feasible way to eliminate the pores entirely is to split the solder joint for a very short time during reflow soldering. By using this method, a pore-free solder joint can be obtained for solder paste with 90% metal content. For solder paste with 86% metal content, the area fraction of pores in the solder joint decreases from 9.3% to 0.8% by the splitting method. The average shear strength of the solder joint obtained by splitting has increased

6 XE et al.: AN EXPERIMENTAL APPROACH TO PORE-FREE REFLOW SOLDERING 153 by 22% and 40% for solder pastes with metal contents of 90% and 86%, respectively. This method shows a promising application to obtain a robust solder joint if incorporated into the nitrogen reflow soldering process. REFERENCES [1] Y. C. Chan, D. J. Xie, J. K. L. Lai, F. Yeung, and H. Wong, 6th International Symposium on Non-Destructive Characterization of Materials. New York Plenum Press, 1993, pp [2] J. H. Lau, Handbook of Fine Pitch Surface Mount Technology. New York Van Nostrand Reinhold, 1994, pp [3] K. Mizuishi, M. Tokuda, and Y. Fujita, IEEE Trans. Comp., Hybrids, Manufact. Technol., vol. 11, no. 4, pp , [4] G. Humpston and D. M. Jacobson, Principles of Soldering and Brazing. ASM International, 1993, pp [5] D. J. Xie, Y. C. Chan, J. K. L. Lai, and I. K. Hui, Porosity formation and its effects on mechanical properties of SMT solder joints, in Proc. 44th ECTC, 1994, pp [6] J. Campbell, Solidijkation of Metals. London, U.K.: Iron & Steel Institute, 1968, p [7] Y. S. Yin, Soldering Handbook. P.R.O.C.: Science and Technology Press of Helongjiang, 1989, pp [8] J. S. Hwang, Solder Paste in Electronics Packaging. New York Van Nostrand Reinhold, 1989, p r D. J. Xie received the B. Eng. and M.Eng. from Huazhong University of Science and Technology, China, in 1982 and 1985, respectively. From 1987 to 1992, he was employed as a Lecturer at Department of Materials Science at the same university. From August 1992 to date, he has worked as a Research Assistant and registered as a Ph.D. student in the Department of Electronic Engineering at City University of Hong Kong. His current research focuses on fatigue life analysis, solder joint defect studies and reflow soldering. Y. C. Chan (SM 95) received the B.S. degree in electrical engineering, the M.S. degree in materials science, and the Ph.D. degree in electrical engineering, from the Imperial College of Science and Technology, London, U.K., in 1977, 1978, and 1983, respectively. In 1983, he joined the Advanced Technology Department of Fairchild Semiconductor in California as a Senior Engineer. In 1985, he was appointed Lectureship in Electronics at the Chinese University of Hong Kong. Between 1987 and 1991, he worked in various senior operations &d engineering management functions in electronics manufacturing (including SAE Magnetics (HK) Ltd. and Seagate Technology). He set up the Failure Analysis and Reliability Engineering Laboratory for SMT PCBA in Seagate Technology (Singapore) prior to joining the City Polytechnic of Hong Kong as a Senior Lecturer in Electronic Engineering in 1991 and was promoted to University Senior Lecturer in His current technical interests include surface mount technology and electronic materials and component reliability. Dr. Chan is currently Chairman of the IEE Hong Kong Center. J. K. L. Lai graduated with first class honors from Keble College, Oxford University in From 1974 to 1985, he was employed as a Research Officer at the Central Electricity Research Laboratories at Leadherhead, Surrey, UK. In 1984, he was appointed Project Leader of the Remaining Life Study Group and a member of the Remnant Life Task Force of the Central Electricity Generating Board, UK. He is now a Professor in the Department of Physics and Materials Science and Head of the? Materials Research Centre, City University of Hong Kong. He is a member of the following committees: the International Institute of Welding s Working Group on Creep, the Pressure Equipment Advisory Committee of the Labour Department, the Plastics Training Board of the Vocational Training Council and the Consumer Council of Hong Kong. He holds one patent and has published over 50 technical papers in refereed intemational journals and conference proceedings. He is also the author of over twenty industrial reports on failure analysis and materials for engineering design, issued by the Central Electricity Generating Board, UK. He has performed over forty cases of consultancy in the area of failure analysis of metallic components in Hong Kong.