A SiC power device packaging reliability detection method and system

Abstract

The present invention relates to the field of semiconductor device testing technology, and specifically to a SiC power device packaging reliability detection method and system. The present invention obtains an on-resistance change curve of a device under test in which the on-resistance changes with a test current; based on a preset on-resistance nominal value and a preset on-resistance tolerance, screens out abnormal on-resistance, and then obtains an abnormal change sub-segment in the on-resistance change curve; obtains an abnormal index of the device under test according to the change trend and change speed of the abnormal change sub-segment, combined with its fluctuation characteristics; performs several tests on the device under test, and obtains a reliability index of the device under test according to the similar characteristics of the abnormal indexes in all test processes and the similar characteristics of all on-resistance change curves, combined with the abnormal index, and then evaluates the packaging reliability of the device under test. The present invention analyzes the dynamic changes of on-resistance under a dynamic test environment to improve the accuracy of packaging reliability detection of silicon carbide power devices.

Description

Technical Field

[0001] The present invention relates to the technical field of semiconductor device testing, and in particular to a method and system for detecting the reliability of SiC power device packaging. Background Art

[0002] Silicon carbide power devices (SiC power devices) are semiconductor power devices with the advantages of higher voltage resistance, high temperature resistance, thermal conductivity and lower conduction loss. They are widely used in power conversion under high voltage, high frequency and harsh working environments. The packaging of SiC power devices is an important part of protecting the devices, providing electrical connections and heat dissipation. The packaging quality directly affects the performance, life and safety of the devices. Therefore, it is very important to test the packaging reliability of SiC power devices.

[0003] The increase in on-resistance is one of the signs of device aging and a potential sign of poor device packaging reliability. In the prior art, the on-resistance of silicon carbide power devices is usually measured under specific current conditions, and the device packaging reliability is evaluated through simple conventional standards or parameters. However, this method can only provide the performance of the device under a single working state, lacks dynamics, and cannot fully reflect the performance of the device under different working conditions. It is easy to ignore early or minor fault signs, which leads to poor packaging reliability detection of silicon carbide power devices. Summary of the invention

[0004] In order to solve the technical problem that the prior art has poor packaging reliability detection effect on silicon carbide power devices, the purpose of the present invention is to provide a SiC power device packaging reliability detection method and system, and the technical solution adopted is as follows:

[0005] A method for detecting reliability of SiC power device packaging, the method comprising:

[0006] Obtaining the on-resistance of the device under test at different test currents, and obtaining an on-resistance variation curve of the on-resistance as the test current varies;

[0007] Based on the preset on-resistance nominal value and the preset on-resistance tolerance, abnormal on-resistance is screened out from all the on-resistances; based on the first occurrence of the abnormal on-resistance, an abnormal change sub-segment in the on-resistance change curve is obtained; based on the change trend and change speed of the abnormal change sub-segment, an abnormal change growth index of the on-resistance of the device under test is obtained; based on the fluctuation characteristics of the abnormal change sub-segment, combined with the abnormal change growth index, an abnormal index of the device under test is obtained;

[0008] The device under test is tested several times to obtain the on-resistance change curve corresponding to each test process and the abnormality index of the device under test in each test process; based on the similar characteristics of all the abnormality indices and the similar characteristics of all the on-resistance change curves, combined with the abnormality index, the reliability index of the device under test is obtained; and the packaging reliability of the device under test is evaluated based on the reliability index.

[0009] Furthermore, the method for obtaining the abnormal on-resistance includes:

[0010] According to the deviation of each on-resistance relative to the preset on-resistance nominal value, the growth parameter of each on-resistance is obtained; the difference between the growth parameter of each on-resistance and the preset on-resistance tolerance is used as the abnormal growth index of each on-resistance; the on-resistance whose abnormal growth index is greater than the preset threshold is used as the abnormal on-resistance.

[0011] Furthermore, the method for obtaining the abnormal change sub-segment includes:

[0012] The current corresponding to the abnormal on-resistance that appears for the first time is taken as the critical current, and the curve segment after the critical current in the on-resistance change curve is taken as the abnormal change sub-segment.

[0013] Furthermore, the method for obtaining the abnormal change growth index includes:

[0014] In the abnormal change sub-segment, the test current corresponding to the maximum value of the abnormal growth index is taken as the extreme abnormal current; the negative correlation mapping result of the sum of the critical current and the extreme abnormal current is multiplied by the maximum abnormal growth index as the first abnormal growth index; the slope of the abnormal change sub-segment is taken as the second abnormal growth index;

[0015] The first abnormal growth index and the second abnormal growth index are integrated to obtain the abnormal change growth index of the on-resistance of the device under test.

[0016] Furthermore, the method for obtaining the abnormality index includes:

[0017] The number of extreme value points in the abnormal change sub-segment is used as a first fluctuation parameter; and a second fluctuation parameter is obtained according to the similarity of the current difference between the test currents corresponding to the adjacent maximum value points in the abnormal change sub-segment;

[0018] The first fluctuation parameter and the second fluctuation parameter are integrated to obtain the fluctuation abnormality parameter of the on-resistance; and the fluctuation abnormality parameter and the abnormal change growth index are integrated to obtain the abnormality index of the device under test.

[0019] Furthermore, the method for obtaining the reliability index includes:

[0020] According to similar features of all the abnormal indexes and similar features of the critical currents in all the on-resistance change curves, combined with the maximum abnormal index, initial reliability of the device under test is obtained;

[0021] According to the fluctuation similarity characteristics of all the on-resistance change curves, a test reference weight is obtained;

[0022] The initial reliability is weighted using the test reference weight, and the weighted result is used as the reliability index of the device under test.

[0023] Furthermore, the method for obtaining the initial reliability includes:

[0024] In all test processes of the device under test, the differences between the different abnormal indexes are accumulated as the first parameter, the maximum abnormal index is used as the second parameter, and the extreme difference of the critical current in all the on-resistance change curves is used as the third parameter;

[0025] The first parameter, the second parameter and the third parameter are fused, and a negatively correlated normalized result of the fusion result is used as the initial reliability of the device under test.

[0026] Furthermore, the method for obtaining the test reference weight includes:

[0027] Under the same test current, the negatively correlated normalized result of the variance of the on-resistance in all the on-resistance change curves is used as a test reference weight sub-parameter; and the test reference weight sub-parameters under the same test current are combined to obtain a test reference weight.

[0028] Furthermore, the method for evaluating the packaging reliability of the device under test based on the reliability index includes:

[0029] When the reliability index is greater than or equal to a preset index threshold, the package of the device under test is determined to be reliable; when the reliability index is less than the preset index threshold, the package of the device under test is determined to be unreliable.

[0030] A SiC power device package reliability detection system comprises a memory, a processor and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, steps of a SiC power device package reliability detection method are implemented.

[0031] The present invention has the following beneficial effects:

[0032] The present invention obtains an on-resistance variation curve of a device under test as the on-resistance varies with a test current, and then screens out abnormal on-resistance based on a preset on-resistance nominal value and a preset on-resistance tolerance, and then obtains an abnormal change sub-segment in the on-resistance variation curve according to the abnormal on-resistance that appears for the first time, so as to provide an analysis basis for evaluating the packaging reliability of a SiC power device through the change of its on-resistance in the subsequent process; further, according to the change trend and change speed of the abnormal change sub-segment, an abnormal change growth index of the on-resistance of the device under test is obtained, and the abnormal change growth index preliminarily reflects that there may be problems in the packaging of the device under test, so as to facilitate the subsequent evaluation of the abnormal index of the device under test in combination with the fluctuation characteristics of the abnormal change sub-segment; based on the variation characteristics of the on-resistance with the current, the fluctuation characteristics of the abnormal change sub-segment are analyzed, and the abnormal index of the device under test is obtained in combination with the abnormal change growth index; in order to avoid the contingency of a single test, the device under test is tested several times, and the reliability index of the device under test is obtained according to the similar characteristics of the abnormal indexes obtained in all test processes, the similar characteristics of all on-resistance variation curves and the abnormal index; finally, the packaging reliability of the device under test is evaluated based on the reliability index. The present invention improves the accuracy of packaging reliability detection of silicon carbide power devices by analyzing the dynamic change of on-resistance under a dynamic test environment. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0034] Figure 1 A flow chart of a SiC power device packaging reliability detection method provided by one embodiment of the present invention;

[0035] Figure 2 A flow chart of a method for obtaining a reliability index provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0036] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a detailed description of a SiC power device packaging reliability detection method and system proposed by the present invention, its specific implementation, structure, features and effects, in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0037] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0038] The specific scheme of a SiC power device packaging reliability detection method and system provided by the present invention is described in detail below with reference to the accompanying drawings.

[0039] See also Figure 1 , which shows a flow chart of a SiC power device packaging reliability detection method provided by an embodiment of the present invention, specifically comprising:

[0040] Step S1, obtaining the on-resistance of the device under test under different test currents, and obtaining an on-resistance variation curve of the on-resistance varying with the test current.

[0041] In order to evaluate the packaging reliability of SiC power devices through the change of on-resistance, in one embodiment of the present invention, any SiC power device is used as a device to be tested for analysis and detection; first, the test conditions of the device to be tested are set, for example, if the rated current of the device to be tested is 50-150A, the starting value of the test current is set to the current corresponding to the left end point of the range lower than the rated current, and the end value of the test current is set to the current corresponding to the right end point of the range of the rated current, such as setting the range of the test current to 20-150A;

[0042] Increase the current to be tested at a growth rate of 1A per minute, and obtain the on-resistance at each test current in real time; for example, start from 20A, and obtain the on-resistance of the device to be tested at 20A; after the first minute, increase the test current to 21A, and obtain the on-resistance of the device to be tested at 21A; and so on, obtain the on-resistance at each test current, and use the test current as the horizontal axis parameter and the on-resistance as the vertical axis parameter to map the on-resistance at each test current to the constructed coordinate system, and fit the on-resistance change curve of the on-resistance changing with the test current.

[0043] It should be noted that obtaining the on-resistance of the device under test is an existing technology well known to those skilled in the art and will not be described in detail here. The implementer may also define the test conditions by himself.

[0044] It should be noted that the embodiment of the present invention regards the on-current detection process corresponding to the test current change process as a complete device under test test process.

[0045] Step S2, based on the preset on-resistance nominal value and the preset on-resistance tolerance, screen out abnormal on-resistance from all on-resistances; obtain the abnormal change sub-segment in the on-resistance change curve according to the first occurrence of abnormal on-resistance; obtain the abnormal change growth index of the on-resistance of the device under test according to the change trend and change speed of the abnormal change sub-segment; obtain the abnormal index of the device under test according to the fluctuation characteristics of the abnormal change sub-segment combined with the abnormal change growth index.

[0046] During the test process, as the test current increases, the thermal effect generated when the test current passes through the silicon carbide power device will cause the temperature of the silicon carbide power device to rise, thereby causing a change in the on-resistance, but the increase in on-resistance is usually within a relatively stable range; if the increase in on-resistance exceeds the specified range, it can be preliminarily considered that the packaging of the device to be tested is abnormal, affecting its performance stability and reliability; based on this, the growth of on-resistance can be used to evaluate and screen abnormal on-resistance, in preparation for subsequent analysis of abnormal changes in on-resistance.

[0047] Preferably, in one embodiment of the present invention, if the changed on-resistance deviates from its nominal value and the increase value is greater than the allowable tolerance range, it means that the on-resistance increases abnormally; therefore, the method for obtaining the abnormal on-resistance includes:

[0048] According to the deviation of each on-resistance relative to the preset on-resistance nominal value, the growth parameter of each on-resistance is obtained; the difference between the growth parameter of each on-resistance and the preset on-resistance tolerance is used as the abnormal growth index of each on-resistance; the on-resistance with an abnormal growth index greater than a preset threshold is used as an abnormal on-resistance.

[0049] As an example, the preset on-resistance nominal value is set to 50mΩ. Since the allowable range of the device on-resistance increase value is 15%-20%, the preset on-resistance tolerance is set to an intermediate value of 17.5%, i.e., 0.175; the preset threshold is set to 0, and after obtaining the abnormal growth index using the calculation formula of the abnormal growth index, the on-resistance with an abnormal growth index greater than 0 is taken as the abnormal on-resistance; it should be noted that the implementer can also set the preset on-resistance nominal value and the preset on-resistance tolerance based on the relevant parameters of the device to be tested; other preset thresholds can also be set, such as 0.1 or other values ​​close to 0.

[0050] The calculation formula of abnormal growth index is: ; Where i is the serial number of the on-resistance; is the abnormal growth index of the i-th on-resistance; is the value of the i-th on-resistance; is the preset nominal value of on-resistance; is the preset on-resistance tolerance; is the growth parameter of the i-th on-resistance.

[0051] In the formula, when the deviation of the on-resistance relative to the preset on-resistance nominal value is positive and the larger it is, the larger the growth parameter of the on-resistance is; then when the growth parameter is greater than the preset on-resistance tolerance, it means that the on-resistance not only deviates from the preset on-resistance nominal value, but also exceeds the preset on-resistance tolerance corresponding to the allowable increase range, and its abnormal growth index is also larger.

[0052] Considering that under normal circumstances, the on-resistance of the silicon carbide power device will change as the current gradually increases; when the test current is small, the power consumption and temperature rise of the silicon carbide power device are not obvious, and the on-resistance basically remains stable and close to its nominal value; as the test current increases, the power consumption and temperature of the silicon carbide power device begin to gradually increase, and the on-resistance will slowly increase accordingly; when the test current further increases and approaches the rated current of the silicon carbide power device, its power consumption and temperature rise significantly, and the on-resistance increases rapidly; therefore, considering that when the on-resistance deviates from the preset on-resistance nominal value for the first time and exceeds the preset on-resistance tolerance, it means that in the subsequent test process, it may always deviate from the nominal value and exceed the tolerance;

[0053] Therefore, in a preferred embodiment of the present invention, the current corresponding to the first occurrence of abnormal on-resistance is taken as the critical current, and the curve segment after the critical current in the on-resistance change curve is taken as the abnormal change sub-segment; in preparation for the subsequent analysis of the on-resistance change characteristics and the evaluation of the abnormal index of the device under test.

[0054] Considering that within the abnormal change sub-segment, if the on-resistance changes faster with the test current and the change trend continues to show an increasing trend, it means that the device under test is facing a heat accumulation effect, which indirectly indicates that there is a problem with the packaging or heat dissipation system of the device under test, and the heat cannot be effectively diffused, resulting in a sharp rise in temperature in certain areas of the device under test, thereby causing an increase in the on-resistance; therefore, the embodiment of the present invention will obtain the abnormal change growth index of the on-resistance of the device under test according to the change trend and change speed of the abnormal change sub-segment; the abnormal change growth index preliminarily reflects that there may be problems with the packaging of the device under test, which is convenient for further combining the fluctuation characteristics of the abnormal change sub-segment to evaluate the abnormal index of the device under test.

[0055] Preferably, in one embodiment of the present invention, considering the first appearance time of the abnormal on-resistance and the appearance time of the abnormal on-resistance with the maximum abnormal growth index, the earlier the abnormal growth index appears, and the larger the corresponding abnormal growth index is, the faster the abnormal change speed of the device under test is, and the current changes with time, then the abnormal change speed can be evaluated based on the size of the corresponding test current; considering that the slope of the abnormal change sub-segment can more intuitively reflect the abnormal change speed; based on this, the method for obtaining the abnormal change growth index includes:

[0056] In the abnormal change sub-segment, the test current corresponding to the maximum value of the abnormal growth index is taken as the extreme abnormal current; the negative correlation mapping result of the sum of the critical current and the extreme abnormal current is multiplied by the maximum abnormal growth index as the first abnormal growth index; the slope of the abnormal change sub-segment is taken as the second abnormal growth index; the first abnormal growth index and the second abnormal growth index are combined to obtain the abnormal change growth index of the on-resistance of the device under test.

[0057] As an example, the calculation formula for the abnormal change growth index is:

[0058] ; Where C is the abnormal change growth index of the abnormal change sub-segment; is the largest abnormal growth index; is an exponential function with the natural constant e as base; is the critical current; It is an extremely abnormal current; is the slope of the abnormal change subsegment; is a linear normalization function; is the first abnormal growth index; It is the second abnormal growth index.

[0059] In the above formula, the sum of critical current and extreme abnormal current is taken as the negative correlation exponential function The x in the equation is used to adjust the logic so that the smaller the corresponding sum value is, the earlier the abnormal on-resistance appears, and the larger the corresponding abnormal growth index is, the faster the abnormal change speed is, and the larger the first abnormal growth index is; the slope is linearly normalized so that the larger the slope is and the value is greater than 0, the larger the corresponding normalized value is, the faster the abnormal change speed is, and the larger the second abnormal growth index is; finally, the two are multiplied and combined to obtain the abnormal change growth index of the on-resistance of the device under test.

[0060] It should be noted that the acquisition and normalization of the slope are already existing technologies and will not be described in detail. In other examples, the implementer may also adopt other negative correlation mapping methods such as taking the inverse, other normalization methods, or fusion methods such as addition or weighted summation, which will not be described in detail.

[0061] Considering that if there is a problem with the packaging of the silicon carbide power device, when the test current increases, the on-resistance will show an unstable change trend, and may increase significantly and the change is relatively disordered, and no longer follows the relatively smooth law of moderate increase with increasing current that the silicon carbide power device itself should have; therefore, the embodiment of the present invention further obtains the abnormal index of the device to be tested based on the fluctuation characteristics of the abnormal change sub-segment combined with the abnormal change growth index.

[0062] Preferably, in one embodiment of the present invention, considering that the number of extreme value points can indirectly reflect the fluctuation of the abnormal change sub-segment, and the current difference corresponding to the adjacent maximum or minimum values ​​can be regarded as the duration of a fluctuation, if the difference in the duration of the fluctuation is greater, it further indicates that the fluctuation of the abnormal change sub-segment is more violent and disordered; therefore, the method for obtaining the abnormal index includes:

[0063] The number of extreme value points in the abnormal change sub-segment is used as the first fluctuation parameter; and the second fluctuation parameter is obtained according to the similarity of the current difference between the test currents corresponding to the adjacent maximum value points in the abnormal change sub-segment;

[0064] The first fluctuation parameter and the second fluctuation parameter are integrated to obtain the fluctuation abnormality parameter of the on-resistance; the fluctuation abnormality parameter and the abnormal change growth index are integrated to obtain the abnormality index of the device under test.

[0065] As an example, the calculation formula for the anomaly index is:

[0066] ;in, is the abnormal index of the device under test; C is the abnormal change growth index of the abnormal change sub-segment; is the number of extreme points in the abnormal change sub-segment, also the first fluctuation parameter; is the group number of adjacent maximum value points in the abnormal change sub-segment; is the total number of groups of adjacent maximum value points in the abnormal change sub-segment; is the absolute value of the current difference between the test currents corresponding to the kth group of adjacent maximum value points in the abnormal change sub-segment; is the absolute value of the current difference between the test currents corresponding to the k+1th group of adjacent maximum value points in the abnormal change sub-segment; is the second fluctuation parameter; It is the abnormal fluctuation parameter of on-resistance.

[0067] In the above formula, the absolute value of the difference is used to measure the absolute value of the current difference between adjacent groups. The smaller the absolute value of the current difference, the smaller the similarity of the fluctuation duration corresponding to different groups, which means that the fluctuation is more violent and chaotic; the more extreme points in the abnormal change sub-segment, the more violent and chaotic the fluctuation is, and then the two are combined to obtain the fluctuation abnormality parameters; when the change fluctuation of the on-resistance is more chaotic and the growth rate is faster, it further indicates that it deviates from its original change characteristics, and the possibility of packaging abnormality is greater, so it is multiplied by the abnormal change growth index to obtain the abnormal index.

[0068] It should be noted that the acquisition of extreme points is already an existing technology and will not be described in detail; in other examples, the implementer may also adopt fusion methods such as addition or weighted summation, which will not be described in detail.

[0069] Step S3, perform several tests on the device to be tested, obtain the on-resistance change curve corresponding to each test process, and the abnormality index of the device to be tested in each test process; obtain the reliability index of the device to be tested based on the similar characteristics of all abnormal indices and the similar characteristics of all on-resistance change curves, combined with the abnormal index; evaluate the packaging reliability of the device to be tested based on the reliability index.

[0070] Taking into account that a test result may be random, an embodiment of the present invention performs several tests on the device to be tested, for example, at least 3 times, and then obtains the on-resistance change curve corresponding to each test process based on the method described in step S1, and obtains the abnormality index of the device to be tested in each test process based on the method described in step S2, which will not be repeated here.

[0071] Considering that when the abnormal indexes of the devices under test evaluated in different test processes are more similar and the corresponding on-resistance change curves are more similar, it means that in several test processes of the devices under test, the randomness of the evaluation results is lower and the confidence level is higher; thus, the abnormal index of the devices under test can be combined to accurately evaluate their packaging reliability.

[0072] Preferably, in one embodiment of the present invention, the method for obtaining the reliability index includes:

[0073] See also Figure 2 , which shows a flow chart of a method for obtaining a reliability index provided by an embodiment of the present invention, specifically comprising:

[0074] Step S201 , obtaining the initial reliability of the device under test according to similar features of all abnormal indexes and similar features of critical currents in all on-resistance variation curves in combination with the maximum abnormal index.

[0075] Considering that the critical current in the on-resistance change curve is an important characteristic node reflecting the beginning of abnormal change of on-resistance, the test confidence can be preliminarily evaluated based on the similar characteristics of the critical current and the similar characteristics of all abnormal indexes, and then the initial reliability of the device under test can be evaluated in combination with the largest abnormal index; considering that the smaller the difference between the two abnormal indexes, the greater the similarity; the smaller the range in the critical current, the more similar the two extremes are, which also indirectly reflects that the critical current is more similar; at the same time, if the maximum abnormal index of the device under test is smaller in all test processes, it further indicates that it is more reliable;

[0076] Therefore, in a preferred embodiment of the present invention, the method for obtaining the initial reliability includes:

[0077] During all test processes of the device under test, the differences between different abnormal indexes are accumulated as the first parameter, the maximum abnormal index is taken as the second parameter, and the extreme difference of the critical current in all on-resistance change curves is taken as the third parameter; the first parameter, the second parameter and the third parameter are fused, and the negatively correlated normalized result of the fusion result is taken as the initial reliability of the device under test.

[0078] As an example, the initial reliability is calculated as:

[0079] ; Wherein, F is the initial reliability of the device under test; is an exponential function with the natural constant e as base; is the maximum abnormal index of the device under test during all tests, which is also the second parameter; q is the cumulative difference between different abnormal indexes of the device under test during all tests, which is also the first parameter; l is the extreme difference of the critical current in all on-resistance change curves, which is also the third parameter.

[0080] In the above formula, the three are multiplied together and then used as a negative correlation index function x in , thereby adjusting the logic so that the smaller the corresponding product is, the greater the initial reliability is; in other examples, the implementer may also add or weight the first parameter and the third parameter together, and then use the combined result as the weight of the second parameter, and then perform other negative correlation mappings such as the inverse of the weighted result to obtain the initial reliability; the implementer may also use the variance of the anomaly index to evaluate the first parameter, and may also use the variance of the critical current to evaluate the third parameter, which are both commonly used mathematical operations and will not be described in detail.

[0081] Step S202: obtaining a test reference weight according to the fluctuation similarity characteristics of all on-resistance variation curves.

[0082] Considering that the test conditions of the DUT are similar, if the on-resistance change curve also has certain fluctuation similarity characteristics, it means that the reliability of the multiple test results of the DUT is higher, and the confidence of the subsequent evaluation results is also higher;

[0083] Therefore, in a preferred embodiment of the present invention, the method for obtaining the test reference weight includes:

[0084] Under the same test current, the negatively correlated normalized result of the variance of the on-resistance in all on-resistance change curves is used as the test reference weight sub-parameter; and the test reference weight sub-parameters under the same test current are combined to obtain the test reference weight.

[0085] As an example, consider the variance as a negative correlation index function x in , thereby adjusting the logic so that the smaller the variance, the more similar the on-resistance in all on-resistance change curves under the same test current, and the larger the test reference weight sub-parameter; further, the average of all the test reference weight sub-parameters under the same test current is calculated, and the average is used as the test reference weight. In other examples, the implementer may also use other negative correlation mapping methods such as reciprocal, which will not be described here.

[0086] Step S203: weighting the initial reliability using the test reference weight, and using the weighted result as the reliability index of the device under test.

[0087] As an example, the test reference weight is multiplied by the initial reliability to obtain the reliability index of the device under test.

[0088] After obtaining the reliability index of the device under test, the packaging reliability of the device under test can be evaluated based on the reliability index.

[0089] Preferably, in one embodiment of the present invention, when the reliability index is greater than or equal to a preset index threshold, the package of the device under test is determined to be reliable; when the reliability index is less than the preset index threshold, the package of the device under test is determined to be unreliable. Since the test reference weight and the initial reliability are both normalized, their value range is from 0 to 1, so the preset index threshold is set to 0.6, and the implementer can also set it by himself.

[0090] The present invention also proposes a SiC power device package reliability detection system, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of a SiC power device package reliability detection method are implemented.

[0091] In summary, the present invention obtains an on-resistance variation curve of the device under test as the on-resistance changes with the test current; based on a preset on-resistance nominal value and a preset on-resistance tolerance, screens out abnormal on-resistance, and then obtains an abnormal change sub-segment in the on-resistance variation curve; obtains an abnormal index of the device under test according to the change trend and change speed of the abnormal change sub-segment, combined with its fluctuation characteristics; performs several tests on the device under test, and obtains a reliability index of the device under test according to the similar characteristics of the abnormal index in all test processes and the similar characteristics of all on-resistance variation curves, combined with the abnormal index, and then evaluates the packaging reliability of the device under test. The present invention analyzes the dynamic changes of on-resistance under a dynamic test environment, and improves the accuracy of packaging reliability detection of silicon carbide power devices.

[0092] It should be noted that the sequence of the above embodiments of the present invention is only for description and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0093] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

Claims

1. A SiC power device packaging reliability detection method, characterized in that: The method comprises: Obtaining the on-resistance of the device under test at different test currents, and obtaining an on-resistance variation curve of the on-resistance as the test current varies; Based on the preset on-resistance nominal value and the preset on-resistance tolerance, abnormal on-resistance is screened out from all the on-resistances; based on the first occurrence of the abnormal on-resistance, an abnormal change sub-segment in the on-resistance change curve is obtained; based on the change trend and change speed of the abnormal change sub-segment, an abnormal change growth index of the on-resistance of the device under test is obtained; based on the fluctuation characteristics of the abnormal change sub-segment, combined with the abnormal change growth index, an abnormal index of the device under test is obtained; The device under test is tested several times to obtain the on-resistance change curve corresponding to each test process and the abnormality index of the device under test in each test process; based on the similar characteristics of all the abnormality indices and the similar characteristics of all the on-resistance change curves, combined with the abnormality index, the reliability index of the device under test is obtained; and the packaging reliability of the device under test is evaluated based on the reliability index.

2. A SiC power device package reliability detection method according to claim 1, characterized in that: The method for obtaining the abnormal on-resistance includes: According to the deviation of each on-resistance relative to the preset on-resistance nominal value, the growth parameter of each on-resistance is obtained; the difference between the growth parameter of each on-resistance and the preset on-resistance tolerance is used as the abnormal growth index of each on-resistance; the on-resistance whose abnormal growth index is greater than the preset threshold is used as the abnormal on-resistance.

3. A SiC power device packaging reliability detection method according to claim 2, characterized in that: The method for obtaining the abnormal change sub-segment includes: The current corresponding to the abnormal on-resistance that appears for the first time is taken as the critical current, and the curve segment after the critical current in the on-resistance change curve is taken as the abnormal change sub-segment.

4. A SiC power device packaging reliability detection method according to claim 3, characterized in that: The method for obtaining the abnormal change growth index includes: In the abnormal change sub-segment, the test current corresponding to the maximum value of the abnormal growth index is taken as the extreme abnormal current; the negative correlation mapping result of the sum of the critical current and the extreme abnormal current is multiplied by the maximum abnormal growth index as the first abnormal growth index; the slope of the abnormal change sub-segment is taken as the second abnormal growth index; The first abnormal growth index and the second abnormal growth index are integrated to obtain the abnormal change growth index of the on-resistance of the device under test.

5. A SiC power device package reliability detection method according to claim 1, characterized in that: The method for obtaining the abnormality index includes: The number of extreme value points in the abnormal change sub-segment is used as a first fluctuation parameter; and a second fluctuation parameter is obtained according to the similarity of the current difference between the test currents corresponding to the adjacent maximum value points in the abnormal change sub-segment; The first fluctuation parameter and the second fluctuation parameter are integrated to obtain the fluctuation abnormality parameter of the on-resistance; and the fluctuation abnormality parameter and the abnormal change growth index are integrated to obtain the abnormality index of the device under test.

6. A SiC power device package reliability detection method according to claim 3, characterized in that: The method for obtaining the reliability index includes: According to similar features of all the abnormal indexes and similar features of the critical currents in all the on-resistance change curves, combined with the maximum abnormal index, initial reliability of the device under test is obtained; According to the fluctuation similarity characteristics of all the on-resistance change curves, a test reference weight is obtained; The initial reliability is weighted using the test reference weight, and the weighted result is used as the reliability index of the device under test.

7. A SiC power device package reliability detection method according to claim 6, characterized in that: The method for obtaining the initial reliability includes: In all test processes of the device under test, the differences between the different abnormal indexes are accumulated as the first parameter, the maximum abnormal index is used as the second parameter, and the extreme difference of the critical current in all the on-resistance change curves is used as the third parameter; The first parameter, the second parameter and the third parameter are fused, and a negatively correlated normalized result of the fusion result is used as the initial reliability of the device under test.

8. A SiC power device package reliability detection method according to claim 6, characterized in that: The method for obtaining the test reference weight includes: Under the same test current, the negatively correlated normalized result of the variance of the on-resistance in all the on-resistance change curves is used as a test reference weight sub-parameter; and the test reference weight sub-parameters under the same test current are combined to obtain a test reference weight.

9. A SiC power device package reliability detection method according to claim 1, characterized in that: The method for evaluating the packaging reliability of the device under test based on the reliability index includes: When the reliability index is greater than or equal to a preset index threshold, the package of the device under test is determined to be reliable; when the reliability index is less than the preset index threshold, the package of the device under test is determined to be unreliable.

10. A SiC power device packaging reliability detection system, characterized in that: The invention comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the steps of a SiC power device packaging reliability detection method as claimed in any one of claims 1 to 9 are implemented.

Images