Suction test sample quantity determination method for bonding tension sampling coverage test

Abstract

The invention discloses a method for determining the number of test samples for a bonding tension sampling coverage test. The method comprises the following steps: obtaining a reference sample and test samples; determining the standard deviation of the reference sample and the standard deviation of the test sample; determining a judgment factor; judging the judgment factor, and if the judgment factor meets the judgment condition, for the submitted products of which the total number is not greater than the reference sample number, carrying out spot check by using the number of the test samples, so as to cover the quality qualification requirements of all the submitted products; and otherwise, increasing the number of the test samples, and carrying out circular judgment until the judgment factors meet the judgment conditions. According to the invention, the working efficiency is improved, and the quality qualification requirements of all submitted products can be covered with the number of samples as few as possible.

Description

Technical Field

[0001] The invention belongs to the technical field of quality inspection, and in particular to a method for determining the number of random test samples for a bonding tension sampling coverage test. Background Art

[0002] Wire bonding is a technology that uses the two ends of a metal wire to connect the chip and the pin to form an electrical connection. It has been 55 years since the invention of wire bonding technology. As the longest-lasting and most widely used technology for interconnecting pins and chips, wire bonding technology has the advantages of high flexibility and low cost, which is enough to take the lead in interconnection technology. At present, wire bonding accounts for 90% of the chip interconnection methods in the entire microelectronic packaging field, and it is expected that it will continue to occupy the first place in the microelectronic packaging field for some time to come.

[0003] With the rapid development of modern packaging technology, the requirements for the inspection of packaging product quality are becoming more and more stringent. Microelectronic packaging includes several important components such as dicing, bonding, bonding, and capping, and bonding is an important step to connect the internal chip with the external leads. The quality of bonding directly affects the performance of components and the yield of packaging. Bonding strength is one of the important indicators for determining bonding quality, so it is very important to use appropriate methods to count the bonding strength of the same device and explore the law of bonding tension.

[0004] Traditional bonding strength test method (all hooks) In traditional methods, wire bonding uses 100% bonding test, or the number of hooks is manually defined. This is unscientific and imprecise, and wastes work efficiency. Therefore, it is urgent to use information technology to conduct bond strength statistics and analysis. Summary of the invention

[0005] The technical problem solved by the present invention is: to overcome the shortcomings of the prior art, and to provide a method for determining the number of test samples for a bonding tension sampling coverage test. The purpose of the present invention is to improve work efficiency, and to carry out research on intelligent interpretation of bonding tension results based on historical data, and to provide a bonding tension sampling coverage test method. By using a model built with big data, correlation calculations can be performed in real time during the bonding tension test. When an extreme correlation is found at a certain moment (the difference between the determination factors is not greater than 0.2), the current test data can be used to put into the model for identification and draw conclusions. If the statistic of the test data approaches or exceeds the warning value, it is considered that the test should be continued until the entire test is completed.

[0006] The technical solution of the present invention is: A method for determining the number of random test samples for a bonding tension sampling coverage test, comprising: 1) Perform bonding tension test on all bonding wires in the same batch of products to obtain the bonding tension value of each bonding wire and use it as a benchmark sample; 2) Randomly select some bonding wires from all the bonding wires in the same batch of products to be inspected as test samples and conduct bonding tension test to obtain the bonding tension value of each bonding wire in the test samples; 3) Determine the standard deviation of the reference sample based on the bonding tension value of the bonding wire obtained in step 1); determine the standard deviation of the sampled sample based on the bonding tension value of each bonding wire in the sampled sample; 4) Determine the absolute value of the difference between the standard deviation of the benchmark sample and the standard deviation of the test sample as the determination factor; 5) Determine the determination factor. If the determination factor meets the determination condition, then for the total number of products to be inspected that is not greater than the number of benchmark samples, the results of the random inspection using the number of random samples can cover the quality requirements of all products to be inspected; otherwise, proceed to step 6); 6) Increase the number of random samples and return to step 3) for cyclic determination until the determination factor meets the determination condition.

[0007] Preferably, the product to be inspected is a semiconductor integrated circuit or a semiconductor discrete device.

[0008] Preferably, the diameter of the bonding wire in the product to be inspected is within a range of no more than 75 μm.

[0009] Preferably, step 1) further includes performing data processing to eliminate errors in the bonding tension value of the reference sample, specifically: If the product to be inspected is a semiconductor integrated circuit, if the bonding force value is lower than the failure criterion standard shown in Table 1 of Method 2011.1 of GJB548B-2005, the bonding force value is determined to be error data and the bonding force value is rejected; if the product to be inspected is a semiconductor discrete device, if the bonding force value is lower than the failure criterion shown in Table 2037-1 of Method 2037 of GJB128A-97, the bonding force value is determined to be error data and the bonding force value is rejected; If the bonding tension value is higher than twice the average value of the reference sample, the bonding tension value is determined to be error data and is eliminated.

[0010] Preferably, step 2) further includes performing data processing to eliminate errors in the bonding tension values ​​of the sample being tested, specifically: If the product to be inspected is a semiconductor integrated circuit, if the bonding force value is lower than the failure criterion standard shown in Table 1 of Method 2011.1 of GJB548B-2005, the bonding force value is determined to be error data and the bonding force value is rejected; if the product to be inspected is a semiconductor discrete device, if the bonding force value is lower than the failure criterion shown in Table 2037-1 of Method 2037 of GJB128A-97, the bonding force value is determined to be error data and the bonding force value is rejected; If the bond tension value is higher than twice the average value of the sample, the bond tension value is judged as error data and is eliminated.

[0011] Preferably, in step 5), the determination factor is determined, specifically: If the difference between the determination factors in the two cycles before and after is not greater than 0.2, the determination factor meets the determination condition; otherwise, the determination condition is not met.

[0012] Preferably, the number of samples increased for random testing in step 6) is in the range of 5% to 10% of the total number of all bonding wires in the same batch of products to be tested.

[0013] The advantages of the present invention compared with the prior art are: 1) The present invention performs error data elimination processing to eliminate values ​​whose bonding tensile force values ​​are higher than twice the average value of the reference samples and values ​​whose bonding tensile force values ​​are lower than the failure criterion, thereby ensuring data validity.

[0014] 2) The present invention can determine the minimum number of random test samples to cover the quality requirements of the random test samples, thereby greatly improving engineering efficiency.

[0015] 3) The determination factors designed in the present invention are reasonable, ensuring that the device is qualified and reliable. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 It is a flow chart of the present invention. DETAILED DESCRIPTION

[0017] The following detailed description of the present invention will make the features and advantages of the present invention more clear and explicit.

[0018] like Figure 1 As shown, the present invention provides a method for determining the number of random samples for a bonding tension sampling coverage test, comprising: 1) Perform bonding tension test on all bonding wires in the same batch of products to be inspected, obtain the bonding tension value of each bonding wire and use it as a reference sample; the products to be inspected are semiconductor integrated circuits or semiconductor discrete devices. The diameter range of the bonding wires in the products to be inspected is not greater than 75μm. Step 1) also includes data processing to eliminate errors in the bonding tension value of the reference sample, specifically: If the product to be inspected is a semiconductor integrated circuit, if the bonding force value is lower than the failure criterion standard shown in Table 1 of Method 2011.1 of GJB548B-2005, the bonding force value is determined to be error data and the bonding force value is rejected; if the product to be inspected is a semiconductor discrete device, if the bonding force value is lower than the failure criterion shown in Table 2037-1 of Method 2037 of GJB128A-97, the bonding force value is determined to be error data and the bonding force value is rejected; If the bonding tension value is higher than twice the average value of the reference sample, the bonding tension value is determined to be error data and is eliminated.

[0019] 2) From all the bonding wires in the same batch of products to be inspected, randomly select some bonding wires (in the embodiment of the present invention, the number of bonding wires selected is no more than 3% of the total number, as the initial number) as test samples and perform bonding tension test to obtain the bonding tension value of each bonding wire in the test sample; step 2) also includes eliminating error data processing for the bonding tension value in the test sample, specifically: If the product to be inspected is a semiconductor integrated circuit, if the bonding force value is lower than the failure criterion standard shown in Table 1 of Method 2011.1 of GJB548B-2005, the bonding force value is determined to be error data and the bonding force value is rejected; if the product to be inspected is a semiconductor discrete device, if the bonding force value is lower than the failure criterion shown in Table 2037-1 of Method 2037 of GJB128A-97, the bonding force value is determined to be error data and the bonding force value is rejected; If the bond tension value is higher than twice the average value of the sample, the bond tension value is judged as error data and is eliminated.

[0020] 3) Determine the standard deviation of the reference sample based on the bonding tension value of the bonding wire obtained in step 1); determine the standard deviation of the sampled sample based on the bonding tension value of each bonding wire in the sampled sample; 4) Determine the absolute value of the difference between the standard deviation of the benchmark sample and the standard deviation of the test sample as the determination factor; 5) Determine the determination factor. If the determination factor meets the determination condition, then for the total number of products to be inspected that is not greater than the number of benchmark samples, the results of the random inspection using the number of random samples can cover the quality requirements of all products to be inspected; otherwise, proceed to step 6); in step 5), determine the determination factor, specifically: If the difference between the determination factors in the two cycles is not greater than 0.2, the determination factor meets the determination conditions, and the results of random inspection using the number of samples in the previous cycle can cover the qualified conditions of all the products submitted for inspection; otherwise, it is determined that the determination conditions are not met.

[0021] 6) Increase the number of random test samples and return to step 3) for cyclic determination until the determination factor meets the determination condition. The value range of the increased number of random test samples in step 6) is 5%~10% of the total number of all bonding wires in the same batch of products to be inspected. That is, the increased number of random test samples is equal to the sum of 5%~10% of the total number of all bonding wires and the number of random test samples in the previous cycle.

[0022] The recognition model designed by the present invention can be applied to the actual test process. In subsequent tests, the model can be continuously optimized to continuously provide sample data, thereby optimizing data preprocessing and algorithms to obtain more accurate conclusions. The number of sampling tests can also be further optimized to improve test efficiency, forming a bonding tension sampling coverage test method.

[0023] Example This method mainly uses the determination factor analysis method to look at the absolute value of the standard deviation to perform key analysis of the determination factor. In the early stage of this case, data was sorted for typical devices in the range of 2014-2022. A total of 53 batches of samples (each batch contains 800-1000 bonding experiments) test results, including 13 unqualified samples and 40 qualified samples. The number of tests for each sample ranged from a dozen to hundreds.

[0024] The data attributes are numerical variables, namely the bonding tension values ​​of 53 batches of samples.

[0025] The bonding tensile force value is tested using a bonding tensile force testing instrument, and statistical tools are used to calculate the statistics of each sample test tensile force data, including: mean and standard deviation.

[0026] After eliminating unavailable data, perform the above calculations for each batch of products, starting with 3% probability and increasing by 5% until a suitable number of coverage is found.

[0027] Although the present invention has been disclosed as above with preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make possible changes and modifications to the technical solution of the present invention by using the above disclosed methods and technical contents without departing from the spirit and scope of the present invention. Therefore, any simple modification, equivalent change and modification made to the above embodiments according to the technical essence of the present invention without departing from the content of the technical solution of the present invention shall belong to the protection scope of the technical solution of the present invention. In the absence of conflict, the embodiments of the present application and the technical features in the embodiments can be combined with each other.

[0028] The contents not described in detail in the specification of the present invention belong to the common knowledge of those skilled in the art.

Claims

1. A method for determining the number of samples for a bonding tensile sampling coverage test, characterized in that: include: 1) Perform bonding tension test on all bonding wires in the same batch of products to obtain the bonding tension value of each bonding wire and use it as a benchmark sample; 2) Randomly select some bonding wires from all the bonding wires in the same batch of products to be inspected as test samples and conduct bonding tension test to obtain the bonding tension value of each bonding wire in the test samples; 3) Determine the standard deviation of the reference sample based on the bonding tension value of the bonding wire obtained in step 1); determine the standard deviation of the sampled sample based on the bonding tension value of each bonding wire in the sampled sample; 4) Determine the absolute value of the difference between the standard deviation of the benchmark sample and the standard deviation of the test sample as the determination factor; 5) Determine the determination factors. If the determination factors meet the determination conditions, then for the total number of products to be inspected that is not greater than the number of benchmark samples, the results of the random inspection using the number of random samples can cover the quality requirements of all products to be inspected; Otherwise, go to step 6); 6) Increase the number of random samples and return to step 3) for cyclic determination until the determination factor meets the determination condition.

2. The method for determining the number of random samples for bonding tension sampling coverage test according to claim 1, characterized in that: The products submitted for inspection are semiconductor integrated circuits or semiconductor discrete devices.

3. The method for determining the number of random samples for bonding tension sampling coverage test according to claim 2, characterized in that: The diameter range of the bonding wire in the product to be inspected shall not exceed 75μm.

4. A method for determining the number of test samples for a bonding tension sampling coverage test according to any one of claims 1 to 3, characterized in that: Step 1) also includes processing the bonding tension value in the reference sample to eliminate the error data, specifically: If the bonding tensile force value is lower than the failure criterion standard, the bonding tensile force value is determined to be error data and is eliminated; If the bonding tension value is higher than twice the average value of the reference sample, the bonding tension value is determined to be error data and is eliminated.

5. The method for determining the number of random samples for bonding tension sampling coverage test according to claim 4, characterized in that: Step 2) also includes processing the bonding tension values ​​in the sample for eliminating errors, specifically: If the bonding tensile force value is lower than the failure criterion standard, the bonding tensile force value is determined to be error data and is eliminated; If the bond tension value is higher than twice the average value of the sample, the bond tension value is judged as error data and is eliminated.

6. The method for determining the number of random samples for bonding tension sampling coverage test according to claim 5, characterized in that: In step 5), the determination factor is determined, specifically: If the difference between the determination factors in the two cycles before and after is not greater than 0.2, the determination factor meets the determination condition; otherwise, the determination condition is not met.

7. The method for determining the number of random samples for bonding tension sampling coverage test according to claim 6, characterized in that: The value range of the number of samples to be added for random testing in step 6) is 5%~10% of the total number of all bonding wires in the same batch of products submitted for inspection.

Images