A reliability test system and method for chip packaging
By setting different accelerated stress levels in chip packaging testing, calculating characteristic lifetime values and comparison parameters, determining the consistency of failure mechanisms, and identifying the critical point of the degradation curve of individual samples, the problem of lifetime prediction deviation and fixed threshold misjudgment caused by mechanism switching in the prior art is solved, and more accurate reliability testing is achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- JIANGSU MINGXIN ADVANCED TECHNOLOGY CO LTD
- Filing Date
- 2026-04-13
- Publication Date
- 2026-06-12
Smart Images
Figure CN122193876A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of chip packaging reliability testing technology, specifically a reliability testing system and method for chip packaging. Background Technology
[0002] In the field of chip packaging reliability testing, accelerated life testing is the mainstream method for evaluating interconnect reliability. This method places the sample under accelerated stress such as high temperature and temperature cycling, and predicts the life under normal operating conditions by monitoring the change in daisy chain resistance. Existing technologies generally imply two basic assumptions: first, accelerated stress only changes the failure rate but not the failure mechanism; second, all samples have similar degradation trajectory patterns, and failure can be determined using a uniform fixed threshold.
[0003] The above assumptions face fundamental challenges in advanced packaging, such as 3D stacking.
[0004] Packaging involves various materials such as silicon, copper, solder, and organic substrates, and each failure mechanism corresponds to a different activation energy. When the stress level changes, the dominant failure mechanism may switch. For example, at room temperature, thermal fatigue crack propagation is the main factor, while at high temperature, it may switch to excessive growth of intermetallic compounds. If the failure data measured at high temperature is directly extrapolated to the room temperature lifetime, the lifetime prediction will deviate significantly from the true value. This problem further reveals the logical flaws of the fixed threshold method: because the degradation trajectory patterns of different samples vary significantly, using a uniform numerical threshold to cut curves of different shapes makes it impossible to distinguish between "early slow degradation" and "late rapid degradation," resulting in a large number of sub-healthy samples being misjudged as qualified.
[0005] Existing technologies either rely on expensive destructive physical analysis to verify the mechanism afterward, or employ methods such as multi-stage thresholding and machine learning, but neither has fundamentally solved the core question of "whether accelerated stress changes the failure mechanism". When the consistency of the mechanism cannot be guaranteed, all subsequent failure determinations and lifetime extrapolations based on fixed thresholds lose their physical basis. Therefore, there is an urgent need for a reliability test data analysis method that can verify the consistency of the mechanism and, on this basis, realize individualized failure determination.
[0006] To address the above problems, this invention proposes a reliability testing system and method for chip packaging. Summary of the Invention
[0007] The purpose of this invention is to provide a reliability testing system and method for chip packaging to solve the problems raised in the prior art.
[0008] To achieve the above objectives, the present invention provides the following technical solution: A reliability testing method for chip packaging includes the following steps: S1. Obtain the packaged sample to be tested, divide the packaged sample to be tested into a first group and a second group, place the first group under a first accelerated stress level for reliability testing, place the second group under a second accelerated stress level for reliability testing, and continuously collect data on the changes in electrical parameters of the interconnect structure of each sample with test time and number of cycles during the test. S2. Determine a first characteristic lifetime value based on the change data of the first group of samples, determine a second characteristic lifetime value based on the change data of the second group of samples, and obtain a comparison parameter based on the first characteristic lifetime value and the second characteristic lifetime value through algebraic operations. S3. If the comparison parameters meet the preset consistency conditions, it is determined that the failure mechanisms under the first accelerated stress level and the second accelerated stress level are consistent. S4. If the failure mechanism is consistent, perform differential calculation on the change data of each individual sample to obtain the acceleration characteristics of the degradation curve of the sample, identify the critical point where the acceleration characteristics change, and use the test time and number of cycles corresponding to the critical point as the individual failure criterion of the sample. S5. Replace the preset fixed threshold with the individual failure criterion for reliability statistics.
[0009] Furthermore, S1 further includes the following: Obtain the packaged sample to be tested, and divide the packaged sample to be tested into a first group and a second group; The first group was subjected to a reliability test under a first accelerated stress, which was a first temperature cycling condition. The first temperature cycling conditions include a first high temperature holding temperature, a first low temperature holding temperature, a first heating / cooling rate, and a first cycle duration; The second group was subjected to a reliability test under a second accelerated stress, which was a second temperature cycling condition. The second temperature cycling conditions include a second high temperature holding temperature, a second low temperature holding temperature, a second heating / cooling rate, and a second cycle duration; The first temperature cycling conditions are different from the second temperature cycling conditions; During the reliability test, the electrical parameters of the interconnect structure of each sample were continuously collected, and the changes of the electrical parameters with the number of temperature cycles were recorded. The electrical parameter is the daisy chain resistance.
[0010] Furthermore, S2 further includes the following: Obtain data on the change of daisy chain resistance of each sample in the first group of samples with the number of temperature cycles during the reliability test process; The first failure time of each sample is determined based on the change data. The first failure time is the number of temperature cycles corresponding to the first time when the daisy chain resistance of the sample exceeds the initial resistance value of the first preset multiple. The first failure time of each sample in the first group of samples is sorted in ascending order to obtain the first failure time sequence. Based on the first failure time series, the median rank method is used to calculate the estimated cumulative failure probability for each failure time. In a double logarithmic coordinate system, the first failure time is used as the horizontal axis and the estimated cumulative failure probability is used as the vertical axis; Linear interpolation is performed on the failure data of the first group of samples to obtain the first characteristic lifetime value corresponding to the cumulative failure probability reaching a preset threshold. The second characteristic lifetime value was determined based on the change data of the second group of samples using the same method. Obtain the first absolute temperature corresponding to the first accelerating stress and the second absolute temperature corresponding to the second accelerating stress; The comparison parameter P is calculated based on the first characteristic lifetime value, the second characteristic lifetime value, the first absolute temperature, and the second absolute temperature. The formula for calculating the comparison parameter P is as follows: P=(lnη1-lnη2) / [(1 / T2)-(1 / T1)]; Where η1 is the first characteristic lifetime value, η2 is the second characteristic lifetime value, T1 is the first absolute temperature, T2 is the second absolute temperature, and ln represents the natural logarithm operation.
[0011] Furthermore, S3 further includes the following: Obtain the first absolute temperature corresponding to the first accelerating stress and the second absolute temperature corresponding to the second accelerating stress; Calculate the comparison parameter P1 under the first temperature combination based on the first absolute temperature and the second absolute temperature; Obtain the third set of samples; The third group of samples was subjected to a reliability test under a third accelerated stress. The third characteristic lifetime value is determined based on the change data of the third group of samples, and the third absolute temperature corresponding to the third accelerating stress is obtained. Calculate the comparison parameter P2 under the second temperature combination based on the first characteristic lifetime value and the third characteristic lifetime value; Calculate the comparison parameter P3 under the third temperature combination based on the second characteristic lifetime value and the third characteristic lifetime value; Calculate the maximum relative difference Δ between the comparison parameters under the first temperature combination, the comparison parameters under the second temperature combination, and the comparison parameters under the third temperature combination; The formula for calculating the maximum relative difference Δ is: Δ={[max(P1,P2,P3)-min(P1,P2,P3)] / [(1 / 3)(P1+P2+P3)]}×100%; Wherein, P1 is the comparison parameter under the first temperature combination, P2 is the comparison parameter under the second temperature combination, P3 is the comparison parameter under the third temperature combination, max indicates taking the maximum value, and min indicates taking the minimum value; The maximum relative difference is compared with a preset allowable deviation range; If the maximum relative difference is less than the upper limit of the preset allowable deviation range, then the failure mechanisms under the first accelerated stress, the second accelerated stress, and the third accelerated stress are determined to be consistent. If the maximum relative difference is greater than or equal to the upper limit of the preset allowable deviation range, it is determined that the failure mechanism has changed, the data extrapolation under the current accelerated stress level is terminated, and a prompt message to adjust the test stress is output.
[0012] Furthermore, S4 further includes the following: Data on the change of daisy chain resistance of each individual sample with test time and number of temperature cycles during the reliability test are obtained, and the change data constitutes the degradation curve of the sample. The degradation curve is differentially calculated to obtain the acceleration characteristic sequence of the sample; Each acceleration feature value in the acceleration feature sequence is used to characterize the degree of change in degradation rate within adjacent test time intervals and the degree of change in degradation rate within adjacent temperature cycle intervals.
[0013] Furthermore, S4 further includes the following: Obtain the acceleration feature sequence of each individual sample and identify the critical point in the acceleration feature sequence where the acceleration feature value first exceeds the preset fluctuation range; The test time and number of temperature cycles corresponding to the critical point are used as the individual failure criteria for the sample.
[0014] Furthermore, S5 further includes the following: The individual failure criteria for each sample are determined after the failure mechanism is determined to be consistent. The individual failure criteria include the test time and the number of temperature cycles corresponding to the critical point. The individual failure criterion is used as the failure time data of the sample to replace the failure time data determined by the preset fixed threshold. Based on the failure time data, the reliability lifetime distribution of the first group of samples, the second group of samples, and the third group of samples was fitted to obtain lifetime distribution parameters under each accelerated stress level.
[0015] A reliability testing system for chip packaging includes a test grouping and data acquisition module, a characteristic lifetime and comparison parameter calculation module, a mechanism consistency judgment module, an individual failure critical point identification module, and a reliability statistics module. The test grouping and data acquisition module is used to acquire the packaged sample to be tested, divide the packaged sample to be tested into a first group and a second group, place the first group under a first accelerated stress for reliability testing, place the second group under a second accelerated stress for reliability testing, and continuously collect data on the changes in electrical parameters of the interconnect structure of each sample with test time and temperature cycle number during the test. The feature lifetime and comparison parameter calculation module is used to determine a first feature lifetime value based on the change data of the first group of samples, determine a second feature lifetime value based on the change data of the second group of samples, and obtain comparison parameters based on the first feature lifetime value and the second feature lifetime value through algebraic operations. The mechanism consistency judgment module is used to determine that the failure mechanisms under the first accelerated stress and the second accelerated stress are consistent when the comparison parameters meet the preset consistency conditions. The individual failure critical point identification module is used to perform differential calculation on the change data of each individual sample to obtain the acceleration characteristics of the sample degradation curve when the failure mechanism is consistent, identify the critical point where the acceleration characteristics change, and use the test time and temperature cycle number corresponding to the critical point as the individual failure criterion of the sample. The reliability statistics module is used to replace the preset fixed threshold with the individual failure criteria for reliability statistics.
[0016] Compared with the prior art, the beneficial effects of the present invention are: 1. This invention sets up at least two sets of reliability tests under different accelerated stress levels, performs algebraic calculations based on the characteristic lifetime values and stress level parameters of each set of samples to obtain comparison parameters, and determines whether the failure mechanism has changed based on whether the comparison parameters meet the preset consistency conditions. Thus, the equivalence between accelerated stress and failure mechanism can be quantitatively verified in the data analysis stage, solving the fundamental problem of lifetime extrapolation distortion caused by mechanism switching in the prior art, and avoiding the limitations of relying on expensive destructive physical analysis for post-verification. 2. Under the premise of consistent failure mechanism, this invention obtains acceleration characteristics by differential calculation of the degradation curve of each individual sample, and identifies the critical point where the acceleration characteristics change significantly as an individual failure criterion, replacing the traditional fixed threshold judgment method. This allows the failure judgment to adapt to the individual differences in the degradation trajectory of different samples, solving the problem that the fixed threshold method cannot distinguish between early slow degradation and late rapid degradation, leading to the misjudgment of sub-healthy samples as qualified, and realizing the individualized identification of failure critical points. Attached Figure Description
[0017] Figure 1 This is a flowchart of a reliability testing method for chip packaging according to the present invention. Detailed Implementation
[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0019] Example: Figure 1 As shown, the present invention provides a technical solution. Example
[0020] like Figure 1 As shown, the present invention provides a reliability testing method for chip packaging, including the following steps S1 to S5.
[0021] S1. Sample Grouping and Data Acquisition Acquire test package samples and divide them into two groups. The first group is subjected to a first accelerated stress level for reliability testing, and the second group is subjected to a second accelerated stress level for reliability testing. During the testing process, continuously collect data on the changes in electrical parameters of the interconnect structure of each sample over test time and cycle number. Specifically, the accelerated stress can be achieved through temperature cycling, but it is not limited to this. Other accelerated stresses such as constant high temperature, high humidity, and voltage bias can also be used. This embodiment uses temperature cycling as an example for illustration.
[0022] The first group of samples was tested under the first temperature cycling condition, which included a first high temperature holding temperature, a first low temperature holding temperature, a first heating and cooling rate, and a first cycle duration. The second group of samples was tested under the second temperature cycle conditions, which included the second high temperature holding temperature, the second low temperature holding temperature, the second heating and cooling rate, and the second cycle duration. The first temperature cycle conditions are different from the second temperature cycle conditions. For example, the high temperature holding temperature of the first temperature cycle conditions is higher than the high temperature holding temperature of the second temperature cycle conditions, or the low temperature holding temperature is lower, or the heating and cooling rates are faster, etc.
[0023] During the test, the daisy chain resistance value of each sample was continuously collected, and the resistance reading at the end of each temperature cycle was recorded. Daisy chain resistance is a commonly used electrical parameter characterizing the health of interconnect structures, and its changes can reflect the degradation of interconnect points.
[0024] By setting up at least two sets of tests with different stress levels, a data foundation was provided for subsequent verification of the consistency of the mechanism. This step does not require any changes to the existing test hardware; it only requires adding a stress level to the test plan.
[0025] S2. Calculation of characteristic lifetime value and comparison parameters The first characteristic lifetime value is determined based on the change data of the first group of samples, and the second characteristic lifetime value is determined based on the change data of the second group of samples. The comparison parameters are obtained by algebraic operations based on the first characteristic lifetime value and the second characteristic lifetime value.
[0026] Specifically, for the first group of samples, data on the change of daisy chain resistance of each sample with the number of temperature cycles were obtained during the reliability test. The first failure time for each sample is determined based on the change data. The first failure time is the number of temperature cycles corresponding to the first time the daisy chain resistance of the sample exceeds the initial resistance value of the first preset multiple. The first preset multiple is usually 1.1, which means that a resistance exceeding 10% of the initial value is considered a failure. However, this multiple can be preset according to the specific package type and test standards.
[0027] The first failure time of each sample in the first group of samples is sorted in ascending order to obtain the first failure time sequence.
[0028] The median rank method is used to calculate the cumulative failure probability estimate corresponding to each failure time based on the first failure time series. The median rank method formula is: F(i)=(i-0.3) / (n+0.4); Where i is the sequence number of the failure time in the sequence, n is the total number of samples in the group, and F(i) is the estimated cumulative failure probability corresponding to the i-th failure time.
[0029] In a double logarithmic coordinate system, with the first failure time as the horizontal axis and the cumulative failure probability estimate as the vertical axis, linear interpolation is performed on the failure data of the first group of samples to obtain the first characteristic lifetime value corresponding to the cumulative failure probability reaching the preset threshold. The preset threshold is usually set to 63.2% because the characteristic lifetime in the Weibull distribution is defined as the lifetime corresponding to a cumulative failure probability of 63.2%, but other values can also be used. Linear interpolation method: Connect the data points on a log-log graph to find the x-coordinate value corresponding to the threshold value on the y-axis.
[0030] Using the same method, the second characteristic lifetime value was determined based on the change data of the second group of samples.
[0031] Obtain the first absolute temperature T1 corresponding to the first accelerating stress and the second absolute temperature T2 corresponding to the second accelerating stress; The absolute temperature is obtained by adding 273.15 to the Celsius temperature.
[0032] The comparison parameter P is calculated based on the first characteristic lifetime value η1, the second characteristic lifetime value η2, the first absolute temperature T1, and the second absolute temperature T2. The calculation formula is as follows: P=(lnη1-lnη2) / [(1 / T2)-(1 / T1)]; This comparison parameter is essentially a simplified representation of the apparent activation energy, which reflects the dependence of lifetime on temperature. It can be obtained through algebraic operations without the need for complex modeling; this step transforms the original test data into quantitatively comparable feature values, laying the foundation for judging the consistency of the mechanism.
[0033] S3. Failure Mechanism Consistency Determination If the comparison parameters meet the preset consistency conditions, the failure mechanisms under the first accelerated stress level and the second accelerated stress level are determined to be consistent.
[0034] To make a more reliable judgment, three or more stress levels can be set, and multiple sets of comparison parameters can be calculated; this embodiment uses three sets as an example for illustration: Obtain the first absolute temperature corresponding to the first accelerating stress and the second absolute temperature corresponding to the second accelerating stress, and calculate the comparison parameter P1 under the first temperature combination based on the first absolute temperature and the second absolute temperature.
[0035] A third set of samples was obtained and subjected to a third accelerated stress for reliability testing. This third accelerated stress was a third temperature cycling condition, different from the first and second temperature cycling conditions. Based on the variation data of the third set of samples, the third characteristic lifetime value η3 was determined, and the third absolute temperature T3 corresponding to the third accelerated stress was obtained.
[0036] The comparison parameter P2 under the second temperature combination is calculated based on the first characteristic lifetime value η1 and the third characteristic lifetime value η3, and the comparison parameter P3 under the third temperature combination is calculated based on the second characteristic lifetime value η2 and the third characteristic lifetime value η3.
[0037] The formula for calculating the maximum relative difference Δ among P1, P2, and P3 is: Δ={[max(P1,P2,P3)-min(P1,P2,P3)] / [(1 / 3)(P1+P2+P3)]}×100%; The maximum relative difference Δ is compared with a preset allowable deviation range. The allowable deviation range can be set based on engineering experience, for example, 5%.
[0038] If Δ is less than the upper limit of this range, then the failure mechanism under the three accelerated stresses is determined to be consistent. If Δ is greater than or equal to the upper limit, the failure mechanism is determined to have changed. In this case, the data extrapolation under the current accelerated stress level should be terminated, and a prompt message to adjust the test stress should be output, such as suggesting that the accelerated stress be reduced or a segmented model be used.
[0039] This step, through simple statistical comparison, completes the quantitative verification of the consistency of the mechanism during the data analysis stage, avoiding the limitations of traditional methods that rely on expensive destructive physical analysis or ex-post judgment.
[0040] S4. Individual Failure Critical Point Identification When the failure mechanism is consistent, the acceleration characteristics of the degradation curve of each individual sample are obtained by differential calculation of the change data. The critical point where the acceleration characteristics change is identified, and the test time and number of cycles corresponding to the critical point are used as the individual failure criteria of the sample.
[0041] Specifically, the daisy-chain resistance of each individual sample during the reliability test is obtained as a function of test time and temperature cycles, and this change data constitutes the degradation curve of the sample.
[0042] The degradation curve is differentially calculated to obtain the acceleration characteristic sequence of the sample. Each acceleration characteristic value in the acceleration characteristic sequence is used to characterize the degree of change in degradation rate within adjacent test time intervals, as well as the degree of change in degradation rate within adjacent temperature cycle intervals.
[0043] Difference calculations can use first-order and second-order differences. First, calculate the first-order difference (velocity): vi = y{i+1} - yi; Where yi is the resistance value of the i-th cycle; Then, calculate the second-order difference (acceleration): ai = v{i+1} - vi.
[0044] Acquire the acceleration feature sequence of each individual sample and identify the critical point in the sequence where the acceleration feature value first exceeds the preset fluctuation range; The preset fluctuation range can be obtained by multiplying the standard deviation of the acceleration value in the early stage of the test (such as the first 10% of the cycle) by 3, which is the 3σ criterion.
[0045] The test time and number of temperature cycles corresponding to the critical point are used as the individual failure criteria for this sample; This criterion reflects the turning point from stable degradation to accelerated degradation of the sample, and is closer to the critical state of physical damage than the traditional fixed threshold.
[0046] This step involves only basic difference operations and requires no prior models or machine learning to achieve individualized failure criteria, thus solving the problem that the fixed threshold method cannot adapt to the differences in degradation trajectories of different samples.
[0047] S5, Reliability Statistics Individual failure criteria are used to replace preset fixed thresholds for reliability statistics.
[0048] The individual failure criteria for each sample are determined after the failure mechanism is determined to be consistent. The individual failure criteria include the test time and the number of temperature cycles corresponding to the critical point.
[0049] The individual failure criterion is used as the failure time data of the sample to replace the failure time data determined by the preset fixed threshold.
[0050] Based on these failure time data, reliability lifetime distribution fitting was performed on the first group of samples, the second group of samples, and the third group of samples. For example, Weibull distribution fitting was used to obtain lifetime distribution parameters (shape parameter β and scale parameter η) under each accelerated stress level. Then, the characteristic lifetimes under various stresses can be extrapolated to normal operating conditions using accelerated models (such as the Arrhenius model) to obtain the lifetime distribution under normal operating conditions.
[0051] This step allows reliability assessments to be based on more realistic physical failure thresholds, avoiding errors introduced by fixed thresholds and improving the accuracy of lifetime predictions.
[0052] Through the above steps, this invention solves two underlying logic problems in the prior art—the inability to verify mechanism consistency and the unreasonable fixed threshold failure criterion—without increasing hardware costs or relying on complex algorithms, thus significantly improving the accuracy and reliability of chip packaging reliability testing.
[0053] Example 2 The method of the present invention will be described in detail below using a certain type of chip packaging product as an example.
[0054] This packaged product uses a 3D stacked structure and contains internal microbump interconnects, requiring temperature cycling reliability testing. A total of 45 samples are to be tested.
[0055] First, perform step S1. Randomly divide the 45 samples into three groups: Group 1, Group 2, and Group 3, with 15 samples in each group.
[0056] The first group was tested under the first temperature cycle conditions: high temperature holding temperature 125℃, low temperature holding temperature -40℃, heating and cooling rate 15℃ / min, and cycle duration 60 minutes.
[0057] The second group was tested under the second temperature cycle conditions: high temperature holding temperature 150℃, low temperature holding temperature -55℃, heating and cooling rate 15℃ / min, and cycle duration 60 minutes.
[0058] The third group was tested under the third temperature cycle conditions: high temperature holding temperature 100℃, low temperature holding temperature -25℃, heating and cooling rate 15℃ / min, and cycle duration 60 minutes.
[0059] During the test, the daisy chain resistance value of each sample was continuously collected, and the resistance reading was recorded at the end of each temperature cycle.
[0060] Next, proceed to step S2. For each sample in the first group, when the daisy chain resistance first exceeds 1.1 times the initial resistance value, record the number of temperature cycles at this point as the first failure time of that sample.
[0061] The first failure times of the 15 samples in the first group were sorted in ascending order to obtain a failure time series. The median rank method was used to calculate the estimated cumulative failure probability for each failure time. In a double logarithmic coordinate system, with the failure time as the x-axis and the estimated cumulative failure probability as the y-axis, the number of temperature cycles corresponding to a cumulative failure probability of 63.2% was obtained through linear interpolation, which is the first characteristic lifetime value η1. Similarly, the second set of characteristic lifetime values η2 and the third set of characteristic lifetime values η3 were obtained. In this embodiment, η1 = 1850 cycles, η2 = 620 cycles, and η3 = 3250 cycles were measured.
[0062] Obtain the absolute temperature corresponding to each stress: Group 1 T1 = 125 + 273.15 = 398.15 K, Group 2 T2 = 150 + 273.15 = 423.15 K, Group 3 T3 = 100 + 273.15 = 373.15 K.
[0063] Calculate the alignment parameters: P1=(ln1850-ln620) / [(1 / 423.15)-(1 / 398.15)]≈7.52; P2=(ln1850-ln3250) / [(1 / 373.15)-(1 / 398.15)]≈7.48; P3=(ln620-ln3250) / [(1 / 373.15)-(1 / 423.15)]≈7.55; Then proceed to step S3. Calculate the maximum relative difference: maximum value P3 = 7.55, minimum value P2 = 7.48, average value approximately 7.52, Δ = (7.55 - 7.48) / 7.52 × 100% ≈ 0.93%.
[0064] The upper limit of the preset allowable deviation range is 5%, and 0.93% < 5%. Therefore, it is determined that the failure mechanism is consistent under the three temperature cycling conditions, and further analysis can continue.
[0065] Next, proceed to step S4. Taking a sample from the first group as an example, obtain the data on its daisy chain resistance as a function of the number of cycles to construct a degradation curve; Calculate the first-order difference (velocity) and the second-order difference (acceleration). In the first 200 cycles, the acceleration value fluctuates within ±0.005 Ω / cycle². At the 215th cycle, the acceleration value reaches 0.018 Ω / cycle², exceeding the preset fluctuation range of 0.015 Ω / cycle² (based on three times the standard deviation of acceleration in the first 10% of cycles) for the first time. Therefore, the 215th cycle is identified as the critical point, and the test time (215 cycles) corresponding to this critical point is used as the individual failure criterion for that sample. Repeat this process for all samples to obtain the individual failure criterion for each sample.
[0066] Finally, step S5 is executed. The individual failure criterion (i.e., the number of critical point cycles) for each sample is used as the failure time data, replacing the failure time determined by the traditional fixed threshold method.
[0067] Based on the failure time data of all samples in the first, second, and third groups, a Weibull distribution was fitted to obtain the lifetime distribution parameters under each stress level, which were then extrapolated to the package lifetime under normal operating conditions. For example, the characteristic lifetime at high temperature was extrapolated to room temperature (25°C) using the Arrhenius model to obtain a more accurate lifetime prediction.
[0068] As can be seen from this embodiment, the present invention can verify the consistency of failure mechanisms and identify individualized failure critical points simply through logical analysis and simple algebraic operations on test data, which significantly improves the scientificity and accuracy of chip packaging reliability testing.
[0069] 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.
Claims
1. A reliability testing method for chip packaging, characterized in that: Includes the following steps: S1. Obtain the packaged sample to be tested, divide the packaged sample to be tested into a first group and a second group, place the first group under a first accelerated stress level for reliability testing, place the second group under a second accelerated stress level for reliability testing, and continuously collect data on the changes in electrical parameters of the interconnect structure of each sample with test time and number of cycles during the test. S2. Determine a first characteristic lifetime value based on the change data of the first group of samples, determine a second characteristic lifetime value based on the change data of the second group of samples, and obtain a comparison parameter based on the first characteristic lifetime value and the second characteristic lifetime value through algebraic operations. S3. If the comparison parameters meet the preset consistency conditions, it is determined that the failure mechanisms under the first accelerated stress level and the second accelerated stress level are consistent. S4. If the failure mechanism is consistent, perform differential calculation on the change data of each individual sample to obtain the acceleration characteristics of the degradation curve of the sample, identify the critical point where the acceleration characteristics change, and use the test time and number of cycles corresponding to the critical point as the individual failure criterion of the sample. S5. Replace the preset fixed threshold with the individual failure criterion for reliability statistics.
2. The reliability testing method for chip packaging according to claim 1, characterized in that: S1 further includes the following: Obtain the packaged sample to be tested, and divide the packaged sample to be tested into a first group and a second group; The first group was subjected to a reliability test under a first accelerated stress, which was a first temperature cycling condition. The first temperature cycling conditions include a first high temperature holding temperature, a first low temperature holding temperature, a first heating / cooling rate, and a first cycle duration; The second group was subjected to a reliability test under a second accelerated stress, which was a second temperature cycling condition. The second temperature cycling conditions include a second high temperature holding temperature, a second low temperature holding temperature, a second heating / cooling rate, and a second cycle duration; The first temperature cycling conditions are different from the second temperature cycling conditions; During the reliability test, the electrical parameters of the interconnect structure of each sample were continuously collected, and the changes of the electrical parameters with the number of temperature cycles were recorded. The electrical parameter is the daisy chain resistance.
3. The reliability testing method for chip packaging according to claim 1, characterized in that: S2 further includes the following: Obtain data on the change of daisy chain resistance of each sample in the first group of samples with the number of temperature cycles during the reliability test process; The first failure time of each sample is determined based on the change data. The first failure time is the number of temperature cycles corresponding to the first time when the daisy chain resistance of the sample exceeds the initial resistance value of the first preset multiple. The first failure time of each sample in the first group of samples is sorted in ascending order to obtain the first failure time sequence. Based on the first failure time series, the median rank method is used to calculate the estimated cumulative failure probability for each failure time. In a double logarithmic coordinate system, the first failure time is used as the horizontal axis and the estimated cumulative failure probability is used as the vertical axis; Linear interpolation is performed on the failure data of the first group of samples to obtain the first characteristic lifetime value corresponding to the cumulative failure probability reaching a preset threshold. The second characteristic lifetime value was determined based on the change data of the second group of samples using the same method. Obtain the first absolute temperature corresponding to the first accelerating stress and the second absolute temperature corresponding to the second accelerating stress; The comparison parameter P is calculated based on the first characteristic lifetime value, the second characteristic lifetime value, the first absolute temperature, and the second absolute temperature. The formula for calculating the comparison parameter P is as follows: P=(lnη1-lnη2) / [(1 / T2)-(1 / T1)]; Where η1 is the first characteristic lifetime value, η2 is the second characteristic lifetime value, T1 is the first absolute temperature, T2 is the second absolute temperature, and ln represents the natural logarithm operation.
4. The reliability testing method for chip packaging according to claim 1, characterized in that: S3 further includes the following: Obtain the first absolute temperature corresponding to the first accelerating stress and the second absolute temperature corresponding to the second accelerating stress; Calculate the comparison parameter P1 under the first temperature combination based on the first absolute temperature and the second absolute temperature; Obtain the third set of samples; The third group of samples was subjected to a reliability test under a third accelerated stress. The third characteristic lifetime value is determined based on the change data of the third group of samples, and the third absolute temperature corresponding to the third accelerating stress is obtained. Calculate the comparison parameter P2 under the second temperature combination based on the first characteristic lifetime value and the third characteristic lifetime value; Calculate the comparison parameter P3 under the third temperature combination based on the second characteristic lifetime value and the third characteristic lifetime value; Calculate the maximum relative difference Δ between the comparison parameters under the first temperature combination, the comparison parameters under the second temperature combination, and the comparison parameters under the third temperature combination; The formula for calculating the maximum relative difference Δ is: Δ={[max(P1,P2,P3)-min(P1,P2,P3)] / [(1 / 3)(P1+P2+P3)]}×100%; Wherein, P1 is the comparison parameter under the first temperature combination, P2 is the comparison parameter under the second temperature combination, P3 is the comparison parameter under the third temperature combination, max indicates taking the maximum value, and min indicates taking the minimum value; The maximum relative difference is compared with a preset allowable deviation range; If the maximum relative difference is less than the upper limit of the preset allowable deviation range, then the failure mechanisms under the first accelerated stress, the second accelerated stress, and the third accelerated stress are determined to be consistent. If the maximum relative difference is greater than or equal to the upper limit of the preset allowable deviation range, it is determined that the failure mechanism has changed, the data extrapolation under the current accelerated stress level is terminated, and a prompt message to adjust the test stress is output.
5. The reliability testing method for chip packaging according to claim 1, characterized in that: S4 further includes the following: Data on the change of daisy chain resistance of each individual sample with test time and number of temperature cycles during the reliability test are obtained, and the change data constitutes the degradation curve of the sample. The degradation curve is differentially calculated to obtain the acceleration characteristic sequence of the sample; Each acceleration feature value in the acceleration feature sequence is used to characterize the degree of change in degradation rate within adjacent test time intervals and the degree of change in degradation rate within adjacent temperature cycle intervals.
6. The reliability testing method for chip packaging according to claim 5, characterized in that: S4 further includes the following: Obtain the acceleration feature sequence of each individual sample and identify the critical point in the acceleration feature sequence where the acceleration feature value first exceeds the preset fluctuation range; The test time and number of temperature cycles corresponding to the critical point are used as the individual failure criteria for the sample.
7. The reliability testing method for chip packaging according to claim 1, characterized in that: S5 further includes the following: The individual failure criteria for each sample are determined after the failure mechanism is determined to be consistent. The individual failure criteria include the test time and the number of temperature cycles corresponding to the critical point. The individual failure criterion is used as the failure time data of the sample to replace the failure time data determined by the preset fixed threshold. Based on the failure time data, the reliability lifetime distribution of the first group of samples, the second group of samples, and the third group of samples was fitted to obtain lifetime distribution parameters under each accelerated stress level.
8. A reliability testing system for chip packaging, applied to the reliability testing method for chip packaging as described in any one of claims 1-7, characterized in that: It includes a test grouping and data acquisition module, a characteristic lifetime and comparison parameter calculation module, a mechanism consistency judgment module, an individual failure critical point identification module, and a reliability statistics module; The test grouping and data acquisition module is used to acquire the packaged sample to be tested, divide the packaged sample to be tested into a first group and a second group, place the first group under a first accelerated stress for reliability testing, place the second group under a second accelerated stress for reliability testing, and continuously collect data on the changes in electrical parameters of the interconnect structure of each sample with test time and temperature cycle number during the test. The feature lifetime and comparison parameter calculation module is used to determine a first feature lifetime value based on the change data of the first group of samples, determine a second feature lifetime value based on the change data of the second group of samples, and obtain comparison parameters based on the first feature lifetime value and the second feature lifetime value through algebraic operations. The mechanism consistency judgment module is used to determine that the failure mechanisms under the first accelerated stress and the second accelerated stress are consistent when the comparison parameters meet the preset consistency conditions. The individual failure critical point identification module is used to perform differential calculation on the change data of each individual sample to obtain the acceleration characteristics of the sample degradation curve when the failure mechanism is consistent, identify the critical point where the acceleration characteristics change, and use the test time and temperature cycle number corresponding to the critical point as the individual failure criterion of the sample. The reliability statistics module is used to replace the preset fixed threshold with the individual failure criteria for reliability statistics.