Influence of the hottest packaging technology on t

The influence of packaging technology on the quality and reliability of medical electronic equipment

due to the particularity of application, medical electronic equipment puts forward very high requirements for reliability. It not only needs to work stably for a long time, but also can withstand some special conditions, such as the instantaneous impact caused by high-speed impact

the main factors affecting the decline of reliability

nickel/gold electroplated coating

our concerns about nickel/gold are very different from Kirkendall void effect. First of all, gold plated on the surface of nickel is used to prevent nickel oxidation, but due to process differences, nickel may have been oxidized in advance, or porous in the gold plating layer, which cannot effectively protect it afterwards. In addition, poor process control may also leave organic matter or other pollutants on the surface or inside of gold. No matter what happens, the welding quality will be reduced and a "cold" solder joint will be formed. However, such defects can generally be detected, which only requires slight deformation of the assembly (i.e. mechanical screening) or aging test. Similar to the Kirkendall effect, our most concerned problem is that the formation of this kind of solder joint with our family background is not thick. In the initial mechanical test, it can not be distinguished from the good solder joint, but it will gradually deteriorate in the process of use

ENIG (nickel/Au) gold plating is a self passivation process commonly used to protect chemical nickel coatings, and its coating thickness is automatically limited to 0.075 ~ 0.20 μ M. The ni/au process of nickel melting and gold leaching is widely used in the electronic industry, for example, to eliminate the missing plating vacancy left at the energized electrode in the electroplating process. Dipping is a corrosive process, which replaces the exposed nickel atom with two gold atoms; In order to prevent excessive corrosion, it is necessary to add a considerable amount of protective materials, such as phosphorous substances, to the nickel layer below. In fact, if the corrosion process is not properly controlled, serious problems will also occur

in addition to the common problems of all nickel/gold coatings, there is also a widely recognized unique phenomenon in the gold plating process, namely the so-called "black disk" phenomenon. In fact, "black disk" is a ubiquitous term. Many of the phenomena it includes are related to the fracture of solder joints on or near the Ni (P)/ni3sn4 interface. In most cases, it refers to the phenomenon of lack of weldability of Ni (P) surface due to excessive corrosion in the process of gold plating, but it usually also includes the effect caused by the presence of various alloys near the surface or the presence of alloy pollution

in short, the "black disk" effect usually refers to the "zero time defect" effect, that is, the resulting defects can be detected, and even can be detected only by slight deformation (mechanical screening) or aging test on the assembly. However, there is also a mechanism related to the corrosive "black disk" effect, that is, a seemingly intact solder joint will deteriorate over time. This mechanism has puzzled researchers in the industry for many years. According to this mechanism, the phenomenon of strength weakening and transition to embrittlement fatigue can be measured in the solder joints after aging treatment, and its behavior is almost indistinguishable from that caused by (Ni, AU) Sn4 discussed above

we have carried out aging treatment and shear tests on the solder joints on ENIG pads provided by many manufacturers for various lengths of time at 150 ℃. Experiments show that some samples begin to appear embrittlement fatigue after 250 ~ 1000 hours of aging treatment. Thickness is 0.15 μ A gold coating of more than m is sufficient to form 1.3 on the surface of the pad μ M thick (Ni, AU) Sn4, i.e. embrittlement fatigue can be observed in the shear test. However, for solder balls with a height of 24mil, the aging time required to reach such a level is much longer, and careful analysis does not form such a level. In fact, we found that at least some gold atoms are still bound in (Ni, AU) Sn4 deposits in the solder joint; Unlike gold in ausn4 deposits, these gold have no effect on the growth of (Ni, AU) Sn4 layers on the pad surface

many studies have revealed various correlations between corrosivity and "black disk" effect in gold plating process, one of which is that gold atoms locally melt into nickel grain boundaries, thus increasing the overall thickness of gold (as measured by XRF). The reason may be that excessive corrosion is easy to leave too much crystalline Ni3P under the gold coating, which cannot be welded. Despite the local deep melting of gold atoms, we have observed a large number of well formed solder joints. Since the corrosion degree will decrease rapidly with the increase of phosphorus content in nickel, and the "black disk" effect will also decrease significantly, it is natural that the external correlation between the measured effect and the phosphorus concentration on the surface of Ni (P) is more complex: when the concentration is less than 8 ~ 10%, there will be "problems", and when the concentration reaches 15 ~ 16%, it seems to become "sometimes problems, but not necessarily"

the microstructure analysis of shear test samples showed that there was no simple correlation between aging treatment and phosphorus concentration near the surface. Some samples whose gold atoms are obviously locally fused to the nickel grain boundary fail due to embrittlement fatigue after 250 hours, while others fail in the solder joint after 1000 hours. The expansionary growth of ni3sn4 during reflow welding and aging treatment is easy to leave Ni3P under the gold coating, but experiments show that the ni3p/ni3sn4 interface may still be stable

even if this crystalline phase leading to failure is successfully found, it is still necessary to find out what kind of process (if any) is controlling the rate of deterioration, that is, determining the temperature dependence of deterioration. On the other hand, direct measurement of the activation energy required for embrittlement transition requires a considerable number of similar samples. As expected, due to the poor controllability and great variability of the plating process, the differences between different pads will be very large, even in a component point. Therefore, whether a single activation energy value can be applied to all pads with deterioration remains to be discussed

tin whiskers

in 1986, the U.S. Food and Drug Administration (FDA) warned that cardiac pacemakers made by a manufacturer would fail due to whisker shaped tin crystals growing on the white iron shell and short circuiting the crystals; Tin whiskers disabled a commercial satellite in 1988. Whisker crystal growth on pure tin layer has therefore become a concern in high reliability applications

it was observed in the experiment that the tin coating stored for a long time can grow whiskers up to 9mm at room temperature or slightly higher temperature. This phenomenon occurs in an environment without humidity and electric field, and does not belong to dendritic crystal growth. Sensitivity tests of various parameters show that the growth of this whisker may be greatly enhanced under certain types of mechanical loads. At present, there is no widely accepted explanation for this phenomenon, but an explanation based on recrystallization theory seems to be reasonable. According to this explanation, the nucleation and growth of whiskers at least depend on the local combination of high activation energy and low activation energy grain boundaries with appropriate direction. Generally speaking, in order to prevent the growth of whiskers on the tin layer, it is necessary to control multiple factors in the manufacturing environment at the same time, but although such a control level is possible, it may be unrealistic

it is not practical to limit the final length of whiskers by limiting the thickness of tin layer. Experiments show that if there are many whiskers in the region where the thickness does not decrease significantly, there will be a long-distance diffusion. Perhaps removing the pure tin layer from the final assembly is one of the few safer remedies. It is generally believed that the addition of 3% or more lead by weight can prevent the formation of whiskers; And this phenomenon rarely occurs on tin lead solder joints. Although it is occasionally observed that there are up to 25~30 growth in tin lead solder μ M whiskers, but it may be the abnormal precipitation phenomenon caused by electromigration effect under high current

recently, the microelectronics industry has been promoting the use of pure tin plated components, such as ceramic capacitors or resistors, which is one of the work of the industry's transition to lead-free welding solutions. In addition, the tin coating is also highly recommended to protect the solderability of the organic cushion and the copper pad on the circuit board. In short, people are concerned about whether the solderability of the tin coating will occur. Of course, it is impossible to know which kind of new lead-free solder (if any) can resist the formation of whiskers like tin and lead

there are indications that the coating with uniform thickness may help to reduce the growth of whiskers, but it is not clear to what extent it is reduced or how thick the coating should be

bottom filling technology

anti defect (anti gap) sealing technology with hard materials, such as the traditional flip chip bottom filling technology, can effectively inhibit the growth of whiskers and prevent circuit short circuit. It can well protect the wire joints and solder joints from the influence of mechanical load, and effectively eliminate the fracture caused by the strength degradation of the connection points. The only thing that cannot be improved by underfill technology is the solder joint fracture failure caused by Kirkendall effect, but this effect can usually be avoided by careful manufacturing process control

flip chip assembly generally adopts the bottom filling technology, that is, the solder joint is filled with thermosetting materials with relatively good thermal fatigue strength, so as to minimize the fatigue of the solder joint in the process of thermal drift. For BGA and CSP, repairable heat softened materials are usually used. This material will not increase the thermal resistance, but it is not a big problem for medical implants. The repaired material can even be used for flip chip assembly. Softer filler materials also make the whole assembly more flexible, and usually have a good effect on resisting delamination caused by mechanical load; This may be very important, because once a problem occurs at a welding point, it will lead to catastrophic damage

however, underfill introduces another potential failure mechanism. The flow and solidification of the filler will certainly occasionally form some small bubbles, some of which will be easily bound between two adjacent solder joints. If there is a small enough exposed area on the surface of a solder joint that should be completely sealed, its effective melting point will be reduced due to capillary action and the effective release of medium pressure stress in the solder joint. According to the existing reports, it was observed that the solder was squeezed into these gaps during the reflow treatment to be carried out next and the long-term storage at 80 ℃

after the solder is squeezed into the bubble, the thermal resistance of the solder joint will be reduced, but as mentioned above, this is not a big problem. The solder squeezed into the gap is obviously connected with the adjacent solder joint, which will pose a real threat to precision devices. In fact, it is extremely difficult to quantitatively describe the risk of connectivity in a statistical sense. It is impossible for two voids to appear on the same solder joint surface and at the same position, so it is difficult to predict the occurrence and evolution of solder extrusion

finally, it should be emphasized that in the manufacturing of medical implant devices, most filling materials will be greatly degraded after erosion and cleaning. The selection of materials must consider whether the materials have the characteristics of delamination, fracture and solder extrusion in use

remedial measures and accelerated testing

the Joint Commission on electronic device engineering (JEDEC) of the United States published an article entitled "tin whisker growth test", which made detailed provisions for two storage tests and thermal aging cycle tests. However, due to the complex sensitivity of tin whisker growth to many parameters, how to determine the "acceleration factor" so that the aging test