Font Size: a A A

Pattern Recognition And Statistical Process Control For Die Crack In IC Packaging Process

Posted on:2020-09-27Degree:MasterType:Thesis
Country:ChinaCandidate:J Q D a i s y C a i CaiFull Text:PDF
GTID:2428330602451049Subject:Engineering
Abstract/Summary:PDF Full Text Request
With the advancement of microelectronics technology,China's microelectronics industry is also rapidly developing.As an important back-end production in the semiconductor industry,chip packaging plays an increasingly important role,directly determining the size,weight,application convenience,lifetime,performance and cost of electronic products,and also affecting device-to-system effective links and the quality and competitiveness of microelectronics products.Statistical process control technology is widely used in the packaging process to ensure the stability of packaging production process,which is important to ensure product quality.Chip crack is one of the common failure causes in packaging,which is mainly caused by two reasons in production.One is the pressure and ultrasonic waves generated by the bonding process,or the improper usage of capillary.The other is generated by wafer back grinding,wafer saw and other process,or the collision of the chip with the cartridge during its transportation.In this paper,the former is called process cause and the latter is called material cause.And the non-conforming products rate caused by the latter is much higher than the one caused by the former.When the control chart is established according to the nonconforming product rate,a large number of alarms will occur.Since the crack can only be obtained by electrical testing after being packaged,when the rate of defective products exceeds the control limit,the bonding process needs to be stopped to find the cause.After many investigations,it was proved that most of the cases were caused by raw materials.The suspension of production without theoretical guidance caused serious economic losses.For lots of alarms caused by raw materials(false alarms for the bonding process),and the problem of production suspension without certain reasons,the author investigated the statistical process control of the nonconforming product rate caused by the above two reasons,and established a new control chart with theoretical analysis and matlab simulation in this paper.The statistical process control technology can identify the main causes of substandard products and reduce the shutdown of the bonding process.The research results in this paper can assist engineers find out the reasons for loss of control,reduce the economic losses caused by production stoppages,improve the reliability of semiconductor package products,and are of great significance to improve the market competitiveness of enterprises.The following is the main research done by the author.1)Correcting the common mistakes in the actual production of semiconductor packages when controlling the non-conforming products rate,and introducing the SPC principle of the non-conforming product rate control chart;2)Establishing the models of the non-conforming products rate of two varieties,studying the classification of the causes of unqualified products based on the non-conforming products rate,and putting forward the test methods for the correctness of the classification method;3)Studying the higher false alarm of the existing non-conforming product rate control chart under the low unqualified product rate,and giving the establishment method of the low unqualified rate control chart and the calculation method of the control limit according to the SPC principle;4)Researching the method of establishing the control chart of the non-conforming products of the two varieties when the quantity of the batch products is not fixed for the unqualified products caused by two reasons;5)Combining the control chart of non-conforming product of the two types with the identification method of the cause of non-conforming product,and giving an indication of how to judge whether there are two reasons for the unqualified products and when to trigger the control chart of non-conforming product rate of two varieties according to the production data in actual production.Finally,it is proved the correctness of the research content with 7 sets of actual data and failure analysis result.
Keywords/Search Tags:package, die crack, control chart of non-conforming product rate, classification, statistical process control
PDF Full Text Request
Related items