Method for optimizing test time of semiconductor device test items

Abstract

The application provides a method for optimizing test time of a semiconductor device test item, comprising extracting a sub-test file containing a target test item from an initial test file, wherein the sub-test file includes an initial test waiting time of the target test item; performing data fitting on sample test values to obtain a functional relationship between a test result of the target test item and a pre-test waiting time; replacing the initial test waiting time with a final pre-test waiting time based on the original test values and the functional relationship to obtain a new test file; and testing the target test item of the semiconductor device based on the new test file. The initial test waiting time is replaced with the final pre-test waiting time to generate a new test file capable of testing a small current of the semiconductor device, and the small current of the semiconductor device is tested using the new test file, thereby improving the test efficiency of the small current in serial testing while ensuring test accuracy.

Description

Technical Field

[0001] This invention relates to the field of semiconductor device testing technology, and in particular to a method for optimizing the testing time of semiconductor device test items. Background Technology

[0002] Wafer Acceptance Test (WAT) is a crucial step in ensuring that integrated circuit devices on a wafer meet the required performance and quality parameters. It is an indispensable component in the design, manufacturing, and application of integrated circuit products, and has become one of the key technologies for ensuring the reliability of integrated circuit products. Among the target test items measured by WAT testing equipment, low-current testing accounts for a very large proportion, and its testing time accounts for a significant portion of the total product testing time. Therefore, reducing the testing time of low-current testing and ensuring the accuracy of its data plays a vital role in integrated circuit design, manufacturing, and application.

[0003] Currently, the integrated circuit industry mainly uses two testing modes: parallel testing and serial testing. Serial testing accounts for the vast majority of current testing. Therefore, while ensuring the accuracy of low-current data testing with testing equipment, improving the efficiency of serial testing can significantly improve the return on investment. In serial testing, the total testing time for low-current testing includes two parts: pre-test waiting time and the test duration. The pre-test waiting time is the period after the test equipment applies a signal, during which the signal must be allowed to stabilize completely before testing begins. Existing testing equipment, when performing low-current testing, cannot accurately determine the pre-test waiting time. To ensure testing accuracy, redundant waiting times are typically set, resulting in slow testing speed and low testing efficiency.

[0004] In view of this, it is necessary to improve the optimization methods for the test time of semiconductor device test items in the existing technology in order to solve the above problems. Summary of the Invention

[0005] The purpose of this invention is to disclose an optimization method for the test time of semiconductor device test items, which is used to solve many defects in the existing optimization methods for the test time of semiconductor device test items, especially to improve the test efficiency of serial testing of small current while ensuring the accuracy of small current testing.

[0006] To achieve the above objectives, the present invention provides a method for optimizing the test time of semiconductor device test items, comprising:

[0007] Extract a sub-test file containing the target test item from the initial test file, wherein the sub-test file includes the initial test wait time T0 of the target test item;

[0008] Based on the sub-test file, the original test values ​​of the target test items of several semiconductor devices are obtained, and the average value and standard deviation of several original test values ​​are calculated.

[0009] Set several incremental pre-test waiting times for samples, and obtain the corresponding sample test values ​​for the target test item under different pre-test waiting time conditions;

[0010] Data fitting is performed on the sample test values ​​to obtain a functional relationship between the test result of the target test item and the pre-test waiting time;

[0011] Based on the original test values ​​and the functional relationship, the final pre-test waiting time T for the target test item is determined. f ;

[0012] Replace the initial test waiting time T0 with the final test pre-test waiting time T. f To obtain new test files, and to test the target test items of the semiconductor device based on the new test files.

[0013] As a further improvement of the present invention, the step of setting a plurality of incremental pre-test waiting times for samples, and obtaining the corresponding sample test values ​​of the target test item under different pre-test waiting time conditions, includes:

[0014] The starting value T1 of the pre-test waiting time is set, and the value is incremented by a certain time interval Δt to obtain several different pre-test waiting times T. n T n = T1 + (n-1)*Δt, where n is a natural number greater than 1, T1 < T0, and the maximum value T in the waiting time before the sample test. max ≧2*T0;

[0015] Waiting time T before testing the n distinct samples n Under the given conditions, obtain the sample test value S corresponding to the target test item. n .

[0016] As a further improvement of the present invention, the minimum time that the test system can distinguish is used as the starting point value T1 of the waiting time before the sample test.

[0017] As a further improvement of the present invention, the step of performing data fitting on the sample test values ​​includes:

[0018] Remove outliers from the sample test values, and then perform data fitting on the sample test values ​​after removing outliers.

[0019] As a further improvement of the present invention, the data fitting includes a mathematical fitting method of linear regression.

[0020] As a further improvement of the present invention, the functional relationship includes a monotonically decreasing function or a constant function.

[0021] As a further improvement of the present invention, the final pre-test waiting time T of the target test item is determined based on the original test value and the functional relationship. f ,include:

[0022] If the functional relationship is a monotonically decreasing function, obtain the optimized theoretical value S of the target test item, S = μ + m * σ, where μ is the average value of the original test values ​​of the plurality of items, σ is the standard deviation of the original test values ​​of the plurality of items, and m is the optimization coefficient and 0 < m < 1.

[0023] Based on the optimized theoretical value S of the target test item and the functional relationship, the final pre-test waiting time T of the target test item is obtained. f The final pre-test waiting time T for the target test item f The optimized theoretical value S of the target test item corresponds to the pre-test waiting time value in the functional relationship.

[0024] As a further improvement of the present invention, the value range of the optimization coefficient m is: 0.4 < m < 0.8.

[0025] As a further improvement of the present invention, the final pre-test waiting time T of the target test item is determined based on the original test value and the functional relationship. f ,include:

[0026] If the functional relationship is a constant function, then the minimum time that the test system can distinguish is taken as the final pre-test waiting time T of the target test item. f .

[0027] As a further improvement of the present invention, the target test item is a small current test, and the value of the small current is less than 1μA.

[0028] Compared with the prior art, the beneficial effects of the present invention are:

[0029] This invention extracts a sub-test file containing only the target test item from an initial test file. The sub-test file includes the initial test waiting time for the target test item. Based on the sub-test file, the original test values ​​of the target test item for several semiconductor devices are obtained, and the average and standard deviation of the original test values ​​are calculated. Under different sample pre-test waiting time conditions, corresponding sample test values ​​of the target test item are obtained, and data fitting is performed on the sample test values ​​to obtain the functional relationship between the test result of the target test item and the pre-test waiting time. Based on the original test values ​​and the functional relationship, the final pre-test waiting time of the target test item is determined, thus accurately knowing the pre-test waiting time. By replacing the initial test waiting time with the final pre-test waiting time, a new test file capable of testing small currents of semiconductor devices is generated. Using the new test file to test small currents of semiconductor devices can improve the testing efficiency of serial testing of small currents while ensuring testing accuracy. Attached Figure Description

[0030] Figure 1 This is a flowchart of a method for optimizing the test time of semiconductor device test items according to the present invention;

[0031] Figure 2 A flowchart for setting several incremental pre-test waiting times for samples and obtaining the corresponding sample test values ​​for the target test item under different pre-test waiting time conditions;

[0032] Figure 3 This is a flowchart for determining the final pre-test waiting time for the target test item based on the original test values ​​and function relationships.

[0033] Figure 4 The diagram illustrates a scenario where the functional relationship is a monotonically decreasing function, where the functional relationship is: y = a + bx + c × e dx a = 0.57427084, b = -0.17853597, c = 2.49884853, d = -9.57339664, x is in s, y is in pA;

[0034] Figure 5 The graph shows a function relationship as an example of a constant function relationship, where the function relationship is: y = f, f = 13.6713, and the unit of y is nA. Detailed Implementation

[0035] The present invention will now be described in detail with reference to the embodiments shown in the accompanying drawings. However, it should be noted that these embodiments are not intended to limit the present invention. Equivalent changes or substitutions in function, method, or structure made by those skilled in the art based on these embodiments are all within the scope of protection of the present invention.

[0036] The optimization method for the test time of semiconductor device test items disclosed in the present invention aims to obtain a reasonable pre-test waiting time for target test items on the premise of ensuring high-precision testing, and avoid the influence of insufficient waiting time on test results or the low test efficiency caused by too long waiting time. A new test file is established based on the optimized pre-test waiting time of the small current test item, and the new test file is used to test the target test item of the semiconductor device, which can not only ensure high-precision test results, but also achieve the purpose of improving the test efficiency of the target test item.

[0037] Refer to Figures 1 to 5 As shown, this embodiment discloses an optimization method for the test time of semiconductor device test items (hereinafter referred to as "method"), which includes the following steps S1 to S6.

[0038] Step S1: Extract a sub-test file containing the target test item from the initial test file. The sub-test file includes the initial test waiting time T0 of the target test item.

[0039] Specifically, refer to Figure 1 As shown, obtain the initial test file of the semiconductor device preset in the test system. The initial test file contains test condition information of some electrical performance parameters of the semiconductor device, such as small current and threshold voltage. A sub-test file containing only the target test item is extracted from the initial test file through a computer program.

[0040] It should be noted that the target test item includes current or voltage test. For the convenience of description, in this embodiment, the small current test of the semiconductor device is used as the target test item for description. Among them, the small current refers to a current with a value less than 1 μA. In the initial test file, since the pre-test waiting time of the small current cannot be accurately known, in order to improve the test accuracy, the set redundant initial test waiting time T0 of the small current is included in the sub-test file.

[0041] Step S2: Based on the sub-test file, obtain the original test values of the target test items of several semiconductor devices, and calculate the average value and standard deviation of several original test values.

[0042] Specifically, based on the sub-test file, obtain the original test values X1, ……, X n . Through the average value formula: Calculate the average value μ of several original test values. And through the standard deviation formula: Calculate the standard deviation σ of several original test values.

[0043] Step S3: Set several increasing sample pre-test waiting times, and obtain the corresponding sample test values of the target test item under different sample pre-test waiting time conditions.

[0044] Preferably, the reference Figure 2 As shown, several incremental pre-test waiting times are set, and the corresponding sample test values ​​of the target test item are obtained under different pre-test waiting time conditions, including the following steps S201 to S202:

[0045] S201. Set the starting value T1 for the pre-test waiting time of the samples, and increment it by a certain time interval Δt to obtain n different pre-test waiting times T. n T n = T1 + (n-1)*Δt, where n is a natural number greater than 1, T1 < T0, and the maximum value T in the waiting time before the sample test. max ≧2*T0;

[0046] For example, the reference Figure 2 As shown, the initial test waiting time T0, which is less than the target test item's initial test waiting time, is selected as the starting point T1 of the sample test waiting time, i.e., T1 < T0. The time is increased and fluctuated at certain time intervals Δt to obtain n different sample test waiting times T. n : T1, T1+Δt, ..., T1+(n-1)*Δt, which is T n =T1 + (n-1)*Δt, where n is a natural number greater than 1, T n ≥2*T0. Here's an example of how to obtain the pre-test waiting time for a small current test item: Assume the test system measures the pre-test waiting time for the small current test item to be 20ms. Set T1 to 1ms and increment it at Δt intervals of 10ms to obtain five different sample pre-test waiting times: 1ms, 11ms, 21ms, 31ms, 41ms, and 51ms. The last sample pre-test waiting time is the maximum value T of the six different sample pre-test waiting times. max =51ms is greater than twice the initial test waiting time T0 of the small current test item. In this embodiment, only the waiting time before testing six different samples is used as an example. However, in actual operation, based on different starting values ​​T1 and different time intervals Δt, different numbers of waiting times before testing the small current test items can be obtained. It is sufficient that the maximum value of the waiting time before testing is greater than twice the initial test waiting time T0 of the small current test item. This invention does not limit the number of sample waiting times before testing. Preferably, the starting value T1 of the sample waiting time before testing can be set as the minimum time that the test system can distinguish.

[0047] S202, Waiting time T before testing n distinct samples n Obtain the corresponding sample test value S of the target test item under the given conditions. n .

[0048] Specifically, refer to Figure 2 As shown, based on the sample pre-test waiting time obtained in step S201, the waiting time T before different sample tests is measured by the testing system. n Test value S of small current n For example, the testing system measures the small currents S1, S2, S3, S4, S5, and S6 of a semiconductor device during pre-test waiting times of 1ms, 11ms, 21ms, 31ms, 41ms, and 51ms. The small currents S1-S6 can be the average of the small current values ​​of several similar devices under test during the corresponding pre-test waiting times, thus ensuring the accuracy of subsequent data fitting through sufficient sample data.

[0049] Step S4: Test the sample value S n Perform data fitting to obtain a functional relationship between the test results of the target test item and the waiting time before the test;

[0050] Specifically, the following explanation uses the hypothetical test system described above to measure the small currents S1, S2, S3, S4, S5, and S6 of a semiconductor device at pre-test waiting times of 1ms, 11ms, 21ms, 31ms, 41ms, and 51ms as examples. Before data fitting of the sample test values ​​S1, S2, S3, S4, S5, and S6, outliers in the data are removed using the 3σ principle, and then data fitting is performed on the sample test values ​​after outlier removal. The test system measures the small current values ​​S1, S2, S3, S4, S5, and S6 at different sample pre-test waiting times. n Abnormalities may occur. For example, several small current test values ​​S measured by the test system may be abnormal. n If all values ​​are positive, but one or two small current test values ​​are negative, then the negative small current data is considered an outlier. Negative small current test values ​​need to be removed from the sample test values ​​before data fitting is performed. Since the 3σ principle for outlier removal is a current technology, it will not be elaborated upon here.

[0051] Specifically, data fitting includes mathematical fitting methods such as linear regression. For example, the least squares method is used to perform mathematical curve fitting on small current test values ​​after outlier removal. Since using the least squares method for mathematical curve fitting of data is an existing technique, it will not be elaborated upon here.

[0052] After fitting the data, it was found that the functional relationships included monotonically decreasing functions or constant functions.

[0053] Step S5: Based on the original test values ​​and functional relationships, determine the final pre-test waiting time T for the target test item. f ;

[0054] Specifically, refer to Figure 3 As shown, based on the original test values ​​and functional relationships, the final pre-test waiting time T for the target test item is determined. f This includes steps S51 to S53:

[0055] Step S51: Determine the type of the function.

[0056] Specifically, refer to Figure 4 , Figure 5 As shown, after data fitting, it was found that the functional relationship includes either a monotonically decreasing function or a constant function. When the functional relationship between the test result of the small current test item and the waiting time before the test is a monotonically decreasing function, the functional relationship can be expressed as: y = a + bx + c × e dx Parameters a, b, c, and d are constants. When the functional relationship between the test result of the small current test item and the waiting time before the test is a constant function, the functional relationship can be expressed as: y = I, where I is a constant greater than 0.

[0057] Step S52: If the function is a monotonically decreasing function, determine the final pre-test waiting time T for the target test item. f Including steps S521 to S522:

[0058] S521. Obtain the optimized theoretical value S of the target test item, S = μ + m * σ, where μ is the average value of several original test values, σ is the standard deviation of several original test values, and m is the optimization coefficient and 0 < m < 1.

[0059] Specifically, the theoretical optimized value of the small current test item is S: S = μ + m * σ, where μ is the average value of several original test values, σ is the standard deviation of several original test values, and the optimization coefficient m is an empirical value. The range of the optimization coefficient m is 0.4 < m < 0.8, and it can be 0.5, 0.6, 0.8, etc. For the convenience of calculation, the optimization coefficient m is generally taken as 0.5.

[0060] S522. Based on the optimized theoretical value S of the target test item and the functional relationship, obtain the final pre-test waiting time T of the target test item. f The final pre-test wait time T for the target test item f Let S be the pre-test waiting time value corresponding to the optimized theoretical value S of the target test item in the functional relationship.

[0061] Specifically, using an example of a monotonically decreasing function, we will illustrate how to determine the final pre-test waiting time T for a small current test item. f Data fitting yielded the following results: Figure 4One of the monotonically decreasing functions shown is: y = 0.57427084 - 0.17853597x + 2.49884853 × e -9.57339664x Where y corresponds to the unit pA and x corresponds to the unit s. The optimized theoretical value S is expressed as the function y = 0.57427084 - 0.17853597x + 2.49884853 × e -9.57339664x The corresponding pre-test wait time value is the final pre-test wait time T. f .

[0062] Step S53: If the functional relationship is a constant function, determine the final pre-test waiting time T for the target test item. f Including step S531:

[0063] S531. The minimum time that the test system can resolve is taken as the final pre-test waiting time T for the low-current test item. f .

[0064] For example, data fitting yields results such as Figure 5 An example of the constant function shown is: y = 13.6713, where the unit of y is nA. In this case, the minimum resolvable time of the test system is used as the final pre-test waiting time T for the small current test item. f .

[0065] Step S6: Replace the initial test waiting time T0 with the final test waiting time T. f This is to obtain new test files and to test the target test items of the semiconductor device based on the new test files.

[0066] Specifically, refer to Figure 1 As shown, the initial test waiting time T0 of the small current test item in the initial test file is replaced with the optimized final pre-test waiting time T. f The goal is to obtain new test files and use them to test small currents in semiconductor devices, thereby improving the efficiency of serial testing for small currents while ensuring test accuracy.

[0067] The detailed descriptions listed above are merely specific descriptions of feasible embodiments of the present invention, and are not intended to limit the scope of protection of the present invention. All equivalent embodiments or modifications made without departing from the spirit of the present invention should be included within the scope of protection of the present invention.

[0068] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

[0069] Furthermore, it should be understood that although this specification describes embodiments, not every embodiment contains only one independent technical solution. This narrative style is merely for clarity. Those skilled in the art should consider the specification as a whole, and the technical solutions in each embodiment can also be appropriately combined to form other embodiments that can be understood by those skilled in the art.

Claims

1. A method for optimizing the test time of semiconductor device test items, characterized in that, include: Extract a sub-test file containing the target test item from the initial test file. The sub-test file includes the initial test wait time for the target test item. T 0 ; Based on the sub-test file, the original test values ​​of the target test items of several semiconductor devices are obtained, and the average value and standard deviation of several original test values ​​are calculated. Set several incremental pre-test waiting times for samples, and obtain the corresponding sample test values ​​for the target test item under different pre-test waiting time conditions; Data fitting is performed on the sample test values ​​to obtain a functional relationship between the test result of the target test item and the pre-test waiting time; Based on the original test values ​​and the functional relationship, the final pre-test waiting time for the target test item is determined. T f ; The initial test waiting time T 0 Replace with the waiting time before the final test. T f To obtain new test files, and to test the target test items of the semiconductor device based on the new test files.

2. The method for optimizing the test time of semiconductor device test items according to claim 1, characterized in that, The step of setting several incremental pre-test waiting times for samples and obtaining corresponding sample test values ​​for the target test item under different pre-test waiting time conditions includes: Set the starting value of the waiting time before the sample test. T 1 and at time interval Δ t Incremental float to obtain the pre-test wait time for several different samples. T n , ,in, n For natural numbers greater than 1, T 1 < T 0 And the maximum value of the waiting time before the sample test. Waiting time before testing n distinct samples T n Under the given conditions, obtain the corresponding sample test value Sn for the target test item.

3. The method for optimizing the test time of semiconductor device test items according to claim 2, characterized in that, The minimum time that the testing system can resolve is used as the starting point for the pre-test waiting time of the sample. T 1 .

4. The method for optimizing the test time of semiconductor device test items according to claim 1 or 2, characterized in that, The data fitting of the sample test values ​​includes: Remove outliers from the sample test values, and then perform data fitting on the sample test values ​​after removing outliers.

5. The method for optimizing the test time of semiconductor device test items according to claim 4, characterized in that, The data fitting includes the mathematical fitting method of linear regression.

6. The method for optimizing the test time of semiconductor device test items according to claim 4, characterized in that, The functional relationship includes monotonically decreasing functions or constant functions.

7. The method for optimizing the test time of semiconductor device test items according to claim 4, characterized in that, The final pre-test waiting time for the target test item is determined based on the original test values ​​and the functional relationship. T f ,include: If the functional relationship is a monotonically decreasing function, obtain the optimized theoretical value S of the target test item. Where μ is the average of the original test values, σ is the standard deviation of the original test values, and m is the optimization coefficient and 0 < m < 1; Based on the optimized theoretical value S of the target test item and the functional relationship, the final pre-test waiting time of the target test item is obtained. T f The final pre-test waiting time for the target test item T f The optimized theoretical value S of the target test item corresponds to the pre-test waiting time value in the functional relationship.

8. The method for optimizing the test time of semiconductor device test items according to claim 7, characterized in that, The optimization coefficient m The range of values ​​for is: 0.4 < m <0.

8.

9. The method for optimizing the test time of semiconductor device test items according to claim 4, characterized in that, The final pre-test waiting time for the target test item is determined based on the original test values ​​and the functional relationship. T f ,include: If the functional relationship is a constant function, then the minimum time that the test system can distinguish is taken as the final pre-test waiting time for the target test item. T f .

10. The method for optimizing the test time of semiconductor device test items according to claim 1, characterized in that, The target test item is a low-current test, and the value of the low current is less than 1μA.

Images