TWI570654B - Use the test data for quality control - Google Patents

Description

Using test data for quality control methods

The invention relates to a quality control method, in particular to a quality control method using test data.

As shown in Fig. 11, the quality control method of the conventional wafer is mainly based on the wafer pick-up and preparation of the wafer test arrangement (CP-setup), and then the circuit probing (CP) is started. The wafer is divided into a plurality of device under test (DUT) and tested separately. After the wafer test, the unqualified test component is inking, and after baking, the test wafer is processed through the wafer test whole goods station, and then the shipment is ready for shipment. After the item's out-going quality control (OQC) is sent to the warehouse until shipment.

However, in the process of wafer testing, the yield of the component under test contains other factors besides the process of the component under test, such as the problem of the tester, which means that the probe of the test machine is not correct. The calibration, or the test machine's program is not based on the correct test parameters, which will affect the yield of the component under test. However, the quality control process of the wafer test cannot analyze the failure of the component under test. The test machine is not able to correctly solve the problem of the unqualified component under test, and the test data of the qualified test component is also unrecordable. If the final product needs to be corrected, it will not be able to Provide relevant data for calibration.

Therefore, how to solve the problem of the above-mentioned conventional wafer quality control method is the main focus of the present invention.

One of the objects of the present invention is to solve the above problems and provide a quality control method using test data, collecting data of test results after wafer testing by a collecting unit, and sending the data to a test data analysis station. Analysis, using this data to produce a normal distribution map of various collected data, calculate the difference number and standard deviation, as a quality control monitoring means or test program stability analysis.

The second object of the present invention is to solve the above problems and provide a quality control method using test data, and collect data of test results after performing wafer test on a wafer through a collecting unit, and send the data to a test data analysis station. Analysis, with this data can be re-adjusted according to the customer's needs parameters of the component under test, this parameter is the relative value or other difference value rather than the absolute value, but can achieve the effect of re-grading.

The third object of the present invention is to solve the above problems and provide a quality control method using test data, collecting data of test results after performing wafer test on a wafer through a collecting unit, and sending the data to a test data analysis station. Analysis, through the perspective analysis of multiple wafers, can observe product reliability and stability, and help process and design.

The fourth object of the present invention is to solve the above problems and provide a quality control method using test data, and the data can be presented in a three-dimensional space by displaying a normal distribution map in a three-dimensional space coordinate, thereby achieving data. The viewer can see the convenience at a glance.

The fifth object of the present invention is to solve the above problems and provide a quality control method using test data, collect data of test results after performing wafer test on a wafer through a collecting unit, and send the data to a test data analysis station. Analysis, if the finished product requires a large amount of original data, the data can be published in the cloud.

For the purposes of the foregoing, the present invention includes: a wafer is divided into a plurality of components to be tested, and each of the components to be tested is tested for electrical characteristics and performance by a wafer, and the wafer is collected by a collection unit after wafer testing. The batch number, the model of the wafer, the coordinates of the component under test on the wafer, the component number to be tested, and the parameters of the test component and the test results are sent to a test data analysis station for analysis. The normal distribution map of various collected data is used to calculate the difference number and standard deviation, as a quality control monitoring method or test program stability analysis.

The wafer is divided into a plurality of components to be tested. When the components to be tested are tested for electrical characteristics and performance by the wafer, the batch number of the wafer, the type of the wafer, and the component to be tested are collected after wafer testing by a collecting unit on the wafer. The coordinates of the coordinates on the wafer, the component number to be tested, and the parameters of the test component and the test results are sent to a test data analysis station for analysis, so that the analysis of the component under test can be re-adjusted according to the customer's needs. Parameter, which is a relative value or other difference value instead of an absolute value, and can achieve the effect of re-rating.

The wafer is divided into a plurality of components to be tested. When the components to be tested are tested for electrical characteristics and performance by the wafer, the batch number of the wafer, the type of the wafer, and the component to be tested are collected after wafer testing by a collecting unit on the wafer. Coordinates on the wafer, The measured component number and the parameters of the tested component and the test results are sent to a test data analysis station for analysis. Through the perspective analysis of multiple wafers, the reliability and stability of the product can be observed to help the process. And design.

Among them, if the finished product requires a large amount of original data, the data can be published in the cloud.

Wherein, the normal distribution map is represented by a three-dimensional space coordinate, and includes an X coordinate axis representing a first parameter, a Y coordinate axis representing a second parameter, and a Z coordinate axis representing a third parameter, wherein the plurality of coordinate points X, Y, Z forms a normal distribution map of the three-dimensional figure.

The above and other objects and advantages of the present invention will be readily understood from

Of course, the invention may be varied on certain components, or in the arrangement of the components, but the selected embodiments are described in detail in the specification and their construction is shown in the drawings.

1‧‧‧ wafer

11‧‧‧Measured components

2‧‧‧Collection unit

3‧‧‧Test data analysis station

4‧‧‧ computer

5‧‧‧ computer

6‧‧‧Tester main control computer

7, 8‧‧ ‧ classification group

9‧‧‧Internal network

Fig. 1 is a diagram showing the structure of a wafer used in the quality control method of the present invention.

Figure 2 is a flow chart of the quality control method of the present invention.

Figure 3 is a schematic illustration of the apparatus for testing data of the quality control method of the present invention.

Figure 4 is a diagram showing a normal distribution of the wafer of the present invention when it is a batch of data.

Figure 5 is another normal distribution diagram of the wafer of the present invention when it is a batch of data.

Figure 6 is a diagram showing another normal distribution of the wafer of the present invention when it is a batch of data.

Figure 7 is a diagram showing a further normal distribution of the wafer of the present invention when it is a batch of data.

Figure 8 is a normal distribution diagram of the frequency at which the device under test of the present invention is generated during electrical testing.

Figure 9 is a normal distribution diagram of the frequency at which the device under test of the present invention is generated during the electrical test after calibration.

Figure 10 is a normal distribution diagram of the current generated by the device under test in the electrical test as a parameter.

Figure 11 is a method of quality control of conventional wafers.

Referring to Figures 1 through 10, the embodiments shown in the figures are selected for use in the present invention and are for illustrative purposes only and are not limited by such embodiments.

This embodiment provides a quality control method using test data, which includes: as shown in FIG. 1, a wafer 1 is divided into a plurality of device under test (DUT), each of which is tested. After the component 11 is tested by the wafer, the unqualified device under test 11 is inked, and after baking, the tested wafer 1 is processed through the wafer test whole goods station, and then The out-going quality control (OQC) of the items to be shipped is sent to the warehouse until shipment.

As shown in Figure 2, a collection unit 2 (test result data Collection) After the wafer 1 is subjected to circuit probing (CP), the lot number of the wafer 1 (lot no), the type (type no), and the coordinates of the device under test 11 on the wafer 1 (including the X coordinate and The Y coordinate), the measured component number, the result of the binary measurement of the device under test, and the data of the test component test and the test result (data) are sent to a test data analysis station 3 (test data analysis station). analysis.

As shown in FIG. 3, the collection unit 1 is executed by a computer 4, and the test data analysis station 3 is executed by a computer 5, and a test machine main control computer 6 is connected to the two test classification groups 7, 8 through the interior. The connection of the network 9 reaches the transmission of the test data.

To further illustrate the invention, it is described with the aid of Annexes I through II. Please refer to Annex 1 for a brief description of the program executed by the test data analysis station 3 through the computer 5. As shown in page 1-13 of Annex 1, the window for performing test data analysis for the program can be selected as shown in page 1 of Annex 1. The filter path (Filter Dir) specifies the conditions of the analysis. For example, the second page of Annex I and select the analysis data path (Logfile Dir) to specify the data to be analyzed, and the output path of the result after analysis is selected on page 3 of Annex I ( Output Dir), which can be used to analyze the data results as shown in pages 14 to 18 of Annex 1.

When you specify the filter path (Filter Dir) in Annex 1, you can see the main options including Testing Information Head, Probe card information, Bin Description Selection Filter, Dut Selection Filter, and Category Selection Filter on page 4 of Attachment 1. Further options in each main option There are sub-options to choose from, such as the main option in Testing Information Head, as shown on page 5 of Attachment 1. It further includes CUSTON_NAME, CUSTOM_Lot_No, CUSTOM_Type_No, etc., and as shown in page 7 of Attachment 1, further options such as Bin1, Bin2, Bin3, etc. are included in the main option of Bin Description Selection Filter. Therefore, the filter conditions can be selected by the filter path (Filter Dir) to achieve the data output suitable for the condition.

In addition, the data analyzed by the program shown in Annex I can be presented in the form of a form as shown in pages 14 to 18 of Annex 1. The data can also be represented graphically with the aid of graphical software.

As shown in Annex 2, after the above data is listed by microsoft excel, the data is graphically represented by its drawing function. The figure shown in Annex 2 is the data result of a single wafer.

As shown in Figures 4 to 9, the normal distribution map of the three-dimensional coordinates can be obtained by calculation of the drawing software. The normal distribution map includes the X coordinate axis representing the first parameter and represents the second in the three-dimensional coordinates. The Y coordinate axis of the parameter and the Z coordinate axis representing the third parameter, the plurality of coordinate points X, Y, and Z in the three-dimensional space coordinate form a normal distribution map of the three-dimensional figure, and since it is the graphical data of the three-dimensional coordinate, the graphic can be rotated , or the cross-section of the graph, to achieve the effect of perspective analysis, and can further understand the results of the batch of wafers in the data.

As shown in Figures 4 to 6, the normal distribution map is drawn in the form of a peak when the wafer is a batch of data. The data in the figure is drawn as a continuous line segment, wherein the first parameter represented by the X coordinate axis is The frequency generated during the electrical test, the second parameter represented by the Y coordinate axis is the number of wafers, and the Z coordinate axis The third parameter represented is the number of components to be tested. As shown in Fig. 7, it is also a normal distribution diagram in which the wafer is in the form of a batch of data, and the difference between the normal distribution diagrams shown in Figs. 4 to 6 is the data in this figure. It is presented in a histogram, by which the difference in data of each columnar segment is clearly presented in the X, Y, and Z coordinates.

As shown in FIG. 8 , the single wafer 1 is used as a test object. It can be seen that each of the measured components 11 of the wafer 1 is defined by a punctuation point X and Y, and in the normal distribution map, the parameter For example, the frequency generated by each of the devices under test 11 during the electrical test is taken as an example, and the data of the frequency generated by each of the devices under test 11 during the electrical test is the coordinate point Z. From the normal distribution diagram shown in FIG. 8, it can be quickly and clearly understood that the frequency of each of the tested components 11 of the tested wafer 1 during the electrical test has a high and low drop, and there is a need for correction in the process. Another example is the normal distribution map shown in FIG. 9 , after the process calibration, the data of the frequency generated by each of the tested components 11 during the electrical test can also be quickly and clearly known, and the test wafer 1 is The frequency at which each of the devices under test 11 is generated during the electrical test is quite close.

As shown in FIG. 10, it also uses a single wafer as a test object. It can be seen that each of the measured components 11 of the wafer 1 is also defined as a punctuation point X, Y, and in this normal distribution diagram, the parameter is Taking the current generated by each of the tested components during the electrical test as an example, the data of the current generated by each of the tested components during the electrical test is the coordinate point Z.

In addition to the above-mentioned frequency generated during the electrical test or the current generated during the electrical test, the parameter may also be the classification of the device under test 11, passing the test, failing the test or grouping according to different test probes.

In the normal distribution diagrams of the foregoing 4th to 10th, the gradation of the device under test 11 can be set as a parameter, or the group of different test probes can be set as a parameter, and the monolithic pattern can be displayed through the normal distribution map. The difference between the components 11 to be tested on the plurality of wafers 1, for example, the difference between the components 11 to be tested belonging to the group of probes, or the component to be tested 11 belonging to the group of different probes. The difference between the measured components 11 on a single or multiple batches of wafers 1 after a unified analysis, such as the overall analysis of all wafers 1 of the same specification, or the same machine A unified analysis of all wafers 1 produced by the station, or a unified analysis of all wafers 1 produced during a specific period.

By the quality control method of the present invention, the following effects are achieved:

1. Collect the data collected by the unit, use this data to make a normal distribution map of various collected data, calculate the difference number and standard deviation, and use it as a quality control monitoring method or test program stability analysis.

2. The data collected by the collection unit can be re-adjusted according to the customer's needs. The parameter is the relative value or other difference value instead of the absolute value, and the re-grading effect can be achieved.

3. The data collected by the collection unit can make the data appear in three-dimensional space by displaying the normal distribution map in three-dimensional coordinates, thereby achieving the convenience of the viewer of the data at a glance.

4. The data collected by the collection unit 2, if the finished product requires a large amount of original data, the data can be published in the cloud.

5. Through the perspective analysis of multiple batches of wafers 1, the reliability and stability of the product can be observed to help process and design.

6. When the plurality of devices under test 11 are simultaneously tested, by collecting the data and comparing the test differences between the plurality of devices under test 11, the cause of the failure of the device under test 11 can be judged as a problem of the test machine or the test fixture.

The above description of the embodiments is intended to be illustrative of the invention and is not intended to limit the scope of the invention.

From the above detailed description, it will be apparent to those skilled in the art that the present invention can achieve the foregoing objects and is in accordance with the provisions of the Patent Law.

2‧‧‧Collection unit

3‧‧‧Test data analysis station

Claims (6)

A quality control method using test data, the wafer is divided into a plurality of measured components, and each tested component is subjected to wafer testing for electrical characteristics and performance, and a wafer is tested by a collecting unit on the wafer after wafer testing. The batch number, the model of the wafer, the coordinates of the component under test on the wafer, the component number to be tested, and the parameters of the test component and the test results are sent to a test data analysis station for analysis. Collecting the normal distribution map of the data, calculating the difference number and the standard deviation, as the quality control monitoring means or the test program stability analysis, the normal distribution map is presented in a three-dimensional space coordinate, including the X coordinate axis representing the first parameter, representing the second The Y coordinate axis of the parameter and the Z coordinate axis representing the third parameter, the plurality of coordinate points X, Y, and Z in the three-dimensional space coordinate form a normal distribution diagram of the three-dimensional figure. A quality control method using test data, the wafer is divided into a plurality of measured components, and each tested component is subjected to wafer testing for electrical characteristics and performance, and a wafer is tested by a collecting unit on the wafer after wafer testing. The batch number, the model of the wafer, the coordinates of the component under test on the wafer, the component number to be tested, and the parameters of the test component and the test results are sent to a test data analysis station for analysis. Collect the normal distribution map of the data, and use this analysis to re-adjust the parameters of the measured component grading according to the customer's needs. This parameter is a relative value or other difference value instead of an absolute value, and can achieve the effect of re-grading, the normal state The distribution map is presented in a three-dimensional space coordinate, including an X coordinate axis representing the first parameter, a Y coordinate axis representing the second parameter, and Representing the Z coordinate axis of the third parameter, the plurality of coordinate points X, Y, and Z in the three-dimensional space coordinate form a normal distribution map of the three-dimensional figure. A quality control method using test data, the wafer is divided into a plurality of measured components, and each tested component is subjected to wafer testing for electrical characteristics and performance, and a wafer is tested by a collecting unit on the wafer after wafer testing. The batch number, the model of the wafer, the coordinates of the component under test on the wafer, the component number to be tested, and the parameters of the test component and the test results are sent to a test data analysis station for analysis. Collecting the normal distribution map of the data, through the perspective analysis of multiple batches of wafers, the reliability and stability of the product can be observed to help the process and design. The normal distribution map is presented in three-dimensional coordinates, including the X representing the first parameter. The coordinate axis, the Y coordinate axis representing the second parameter, and the Z coordinate axis representing the third parameter, the plurality of coordinate points X, Y, and Z in the three-dimensional space coordinate form a normal distribution map of the three-dimensional figure. According to the request item 1, 2 or 3, the test data is used for the quality control method, wherein if the finished product requires a large amount of original data, the data can be published in the cloud. The quality control method according to claim 1, 2 or 3, wherein the X, Y coordinate axis is an abscissa, the Z coordinate axis is an ordinate, and each of the measured components is defined by coordinate points X and Y. The position on the wafer, and the data of a specific parameter is defined by the coordinate point Z, and a three-dimensional column-like normal distribution map is generated in the three-dimensional space coordinate according to the coordinate points X, Y, and Z corresponding to the respective plurality of measured components. Applying the test data to the quality control method according to claim 5, wherein This parameter is the grading of the device under test, the pass test, the fail test, the group grouped by different test probes, the frequency generated during the electrical test, or the current generated during the electrical test.