AKM implements tests to fit the purpose of the verification, with reference to industry standards such as JEITA, JEDEC and AEC-Q100.
Also, with reference to JEITA EDR-4708 (Guideline for LSI Reliability Qualification Plan), which was newly established in April 2011, we have added and are placing importance on a test method to suit the failure mode, with the purposes of improving the accuracy of reliability predictions and improving the ability to detect failures for each failure mode.
AKM can consider referencing other standards as needed.
- Confirmation of application and required reliability level depending upon the application and specific customer requirements
- Confirmation of differences (novelty) compared with existing products and decision on the qualification details
- Selection of test method
Method 1 Testing using AKM's standard test conditions: Product level tests
Method 2 Testing with failure mode: Testing at TEG level and at product level
Method 1 Testing using AKM's standard test conditions: The test conditions and number of samples are selected based on related standards
Method 2 Testing with failure mode: The test conditions and number of samples derived from Webull analysis with the failure mode
- Verification that all required elements and parameters identified at the planning stage have been satisfied.
AKM implements the qualification when the necessary requirements decided at the product planning and design start stages are satisfied.
- Process reliability qualification: Reliability qualification for the wafer manufacturing process
- Assembly reliability qualification: Reliability qualification for the package assembly process
- Individual product reliability qualification: Reliability qualification for individual products
- Family qualification: We analyze the differences between the relevant product and existing products for which reliability qualification has already been completed and for which there are sufficient proven results in mass production. This reliability qualification can then be applied if it is possible to view the relevant product as a family device of the existing products.
The distribution of failures is divided roughly into the three areas of early failures, operation period failures and wear-out failures. It is possible to show the relationship between the instantaneous failure rate and the actual usage time as a bathtub curve, as shown in Figure below.
It is also possible to know the failure distribution from the Weibull distribution profile parameter m.
- - - Early failure (m<1), operation period failure (m=1), wear-out failure (m>1)
Accelerated tests are defined in JIS(*) Z 8115 :(*) JIS : Japanese Industrial Standards
“Tests done at stress levels which exceed the specified values of the standard conditions, to reduce the observation time for the item's reaction to stress or to increase the reaction during a given time.
Note : In order to maintain the appropriateness of the testing, the accelerated test must not alter the basic fault modes, the failure mechanism or the relationship between these two.
Accelerated testing is a method that focuses on a specific stress in the usage environment. Testing is performed with stricter conditions to accelerate the failure phenomena physically or chemically, so that the failure rate and life under normal conditions can be estimated in a shorter period of time (or with a reduced number of test samples). However, there are limits to the acceleration possible at the product level, so it is not always possible to recreate the targeted failure mode. We also perform highly accelerated tests at the element level using TEG (Test Element Groups).
The four main models of accelerated testing are as follows:
(1) Temperature acceleration model (Arrhenius model)
(2) Temperature difference acceleration model (Eyring model)
(3) Humidity acceleration model
(4) Voltage acceleration model
The above four models are considered when carrying out the product reliability tests and validation.