The IPC Solder Product Value Association (SPVC) conducted a reliability study to determine whether the performance of various tin silver copper (SAC) alloys in lead-free alloys is equivalent in assembly, metallographic analysis and basic characteristics, thermal shock and temperature cycling. The conclusion of SPVC is that the performance of the three SAC alloys is not significantly different, and it is recommended to use SAC 305 as the preferred lead-free alloy in the electronics industry. Another result of this study is that when analyzing test boards using SAC lead-free solder paste, voids were detected. The cavity was marked on the test circuit board, but it was not repaired. In this case, the test circuit board was subjected to thermal shock and temperature cycling. The voids have also undergone thermal shock and temperature cycling, and SPVC also compares the voids in the solder joints.

The debate about solder joints

It is widely recognized in the electronics manufacturing industry that voids commonly seen in X-ray images may be a factor in classifying solder joints as non-conforming, especially when the size of the void exceeds 25% of the solder joint - especially when voids appear in BGA.

There has always been controversy in the electronic assembly industry regarding voids in solder joints. With the shift towards lead-free, controversy is becoming increasingly intense. The reason is that lead-free solder joints are more prone to forming voids than tin lead solder joints; Moreover, the percentage of voids in SAC alloy is higher than that in other lead-free alloys.

Test plan

Two types of lead-free circuit boards were used for testing in the IPC SPVC testing project: lead-free test board A * and lead-free test board B * *. The solder paste used for these two test boards is not specifically selected, but the same soldering flux is used. This way? Where is the beauty of Yan Mu? Private meal with hairy toes and sandpipers? Lai Mou Zhen Tian N Yi knocked and waved his surname, "Father of the Imperial Academy, how do you complain and fight? Apricot vomit??

In the lead-free test board A *, voids with an area exceeding 25% were observed on all circuit boards assembled using three types of SAC alloy solder paste, particularly on CSP84 lead-free components with a spacing of 0.5mm. The circuit board assembled with tin lead alloy clearly has voids, but on CSP84 components assembled with tin lead alloy with a spacing of 0.5mm, the void area is less than 25%.

On lead-free test board B * *, it was observed that the void area exceeded 25% on all circuit boards assembled with SAC alloy. On the circuit board, the voids in PBGA196, C-CSP224, and chip level package CSP8 all exceed 25%.

In order to determine the impact of voids on the reliability of solder joints, the testing scheme adopted by IPC SPVC includes the commonly recognized conventional temperature cycling and thermal shock in the entire industry. Environmental tests were conducted on both sets of test boards, and their functionality was monitored during the testing period. The test board includes circuit boards from two companies, each with four sets of forty boards. Three types of SAC solder alloy compositions and one low melting point (tin lead) solder are used for comparison. Each group took out a plate for destructive metallographic analysis, and this test plate did not participate in the temperature cycling study.

The temperature cycling scheme used reflects the IPC testing method. This temperature cycling scheme first maintains a low temperature (0 ℃) for ten minutes, then slowly rises the temperature to 100 ℃, and then maintains this high temperature for ten minutes before gradually returning to the low temperature state. The entire temperature cycle usually takes about 60 minutes. The cycle time is related to the time it takes for the furnace temperature to rise and the temperature stabilization process of the test board.

The thermal shock test plan is very similar to the one specified by JEDEC, which first maintains a low temperature (-55 ℃) for five minutes, then maintains a high temperature (125 ℃) for ten minutes, and then returns to the low temperature. The total cycle time is about 20 minutes. This cyclic process repeats continuously. Two sets of test samples were subjected to metallographic analysis after 500 temperature cycles per set.

test result

After environmental testing, this study compared the X-ray analysis results of the cavity, comparing the failure situation after 6000 temperature cycles, thermal shock, metallographic examination of both failed and normal solder joints, and comparing the results of metallographic examination. From these comparisons, and using several different statistical methods to compare temperature failure data and void location and size, we can clearly see that voids will not have any impact on the integrity of the solder joint.

After every 500 temperature cycles, remove the circuit board and perform penetrative X-ray imaging on each component on each circuit board. From this image, we can see that there are far more voids in SAC alloy solder joints than in tin lead alloy solder joints. In terms of quantity and size, CSP84 packaging solder joints have more voids than other array packaging solder joints. Comparing the cross-sectional images of packages that have undergone temperature cycling, we cannot see any clear correlation between voids and connection failure. For example, large voids can be seen in the cross-sectional view of the SAC 305 solder joints on test board A *, but this does not necessarily mean that these voids will cause connection failure - although this package has undergone 4500 temperature cycles.

On lead-free test board A *, for 84 input/output CPS packages with a spacing of 0.5mm, there are quite a few solder joint failures caused by creep fatigue caused by temperature cycling. From this, two parameters can be obtained, namely Weibull slope (β) and characteristic life (η) values. Figure 1 is a Weibull plot of the failure distribution of CSP84 packaging. The Weibull distribution indicates that the characteristic lifespan of SAC alloy solder joints is longer than that of tin lead solder joints (SAC alloy has 4713 to 6810 temperature cycles, while tin lead alloy has 1595 temperature cycles). However, on the circuit board without monitoring, the through X-ray images and cross-sectional views of every 500 temperature cycles indicate that there are significantly more and larger voids in the SAC alloy solder joints than in the SnPb solder joints. It is precisely because of this that each CSP84 package was inspected with X-rays after 6000 temperature cycles. We attempted to correlate the number of temperature cycles at which voids occurred with the occurrence of failure. Figure 2 shows Weibull values and η values. These statistical results were obtained from 24 CSP84 packages and 60 CSP84 packages. The 24 CSP84 packages were part of the IPC SPVC reliability test, while the 60 CSP84 packages were part of the reliability test for test board A. In an analysis, there were significant voids in the SAC387 solder joints assembled using conventional remelting techniques. However, there is no particularly significant difference in the characteristic lifespan of solder joints.

conclusion

We compared the number of voids and temperature cycles at which failure occurred in SAC connections based on the data provided by the IPC SPVC reliability research project on SAC alloys. We used eight independent statistical analysis methods (box plot, one-way ANOM, main rendering, matrix plot, etc.) to compare the number of cycles when a void exceeds 25% of the connection area and the number of temperature cycles when a void occurs. The comparison between the distribution of voids and the number of failure cycles indicates that voids do not affect the reliability of solder joints.

The number and size of voids in the solder joints were compared with the connection failure in temperature cycling data, and the conclusion is that there is no evidence to prove that voids in such SAC alloy solder joints will affect the reliability of the solder joints.