The Research On The Long-term Storage Reliability Of Transistors In The Marine Environment

In the marine environment, the temperature, humidity, salt spray and otherfactors will accelerate the aging and corrosion of electronic components, resulting insuch failure modes as the electrical performance degradation, structural materialfracture, and mechanical strength decline and so on. All these above influence thestorage reliability of electronic components.At present, there are still less storage reliability studies on the navy electroniccomponents, while the relevant researches have a series of problems, including thelow test level, unknown rules of performance parameter variation, unclear failuremode and mechanism, inadequate assessment and prediction methods. Thus, how thestorage reliability of the navy electronic components should be evaluated has becomean important issue that needs solving as soon as possible.For the military electronic equipment that is stored for a long-term but used onlyonce (the naval vessel electronic equipment and components), the long-term storagetest in the marine environment can provide reference, basis and basic data for itsstorage reliability assessment, which is significant for improving the reliability ofelectronic equipment.The paper selected the storage environment condition of typical surface-vesselsmachine compartment in the naval equipment system as the application background,and the medium and small power transistors as the study object that are used widelyand representatively. By combining the accelerated storage test and on-the-spot one inthe marine environment, the storage reliability of transistors was evaluated.The major works and innovation of the paper contain the following parts:First, the paper studied the basic situations of marine environments and climate,determined the specific test program to evaluate the storage reliability of transistors inthe marine environment. The environment and climate features of the Chinese fourseas were discussed in detail. Based on this, a new assessment method was proposed,which combines the accelerated storage test in the simulated marine environmentstress with the on-the-spot one in the actual marine environment.Secondly, the laboratory accelerated storage tests were conducted in thetemperature, humidity, salt spray and other marine environment stress conditions,including the constant-temperature-humidity accelerated test, salt spray one and three-stress(temperature, humidity and salt spray)comprehensive one developed byourselves. The paper compared the changes of performance parameters, shell-surfacecorrosion areas and relevant parameters before and after the test, analyzing thepossible performance degradation mechanism and failure mechanism.Thirdly, the on-the-spot storage tests in the actual marine environment wereconducted, including the marine equipped test, coastal sea warehouse storage one and warehouse storage one in the terrestrial natural environment. The samples’ parameterdegradations in the marine environment were analyzed and compared with the data bythe laboratory accelerated storage test. Besides, the paper testified the transistorsamples that have been stored for many years, comparing them with the adnominalvalues set in the production. And the samples whose performance parameters exceedtheir standards were found, which were identified as the failure samples. And theiractual storage lifetime data were got in coastal sea warehouse storage condition, takenas the reference result of prediction effect of accelerated storage test.Fourthly, the performance parameters degradations in different accelerated testswere all analyzed. It is found that the degradation rules of reverse leakage current,current gain, ideality factor and barrier height of transistors respectively conform tothe exponential, logistic, exponential growth and exponential function.Fifthly, three methods were used to predict the transistors’ storage lifetime in themarine environment:(1) A new method was proposed to predict the storage lifetime, using the idealityfactor and barrier height as the degradation sensitive parameter degradation, based onthe performance parameter. Adata processing program was developed by us, based onthe performance parameters degradation. Combing this program, C-V features andbarrier height changes, the average storage lifetime in three common distributionswere extrapolated from the degradation data of reverse leakage current and idealityfactors. With Hallberg-Peck model, the storage lifetime in the natural storagecondition was predicted. Meanwhile, the accelerated factors in the test were alsogiven.(2) Based on the Wiener Process of multi-performance parameter degradation, anew prediction model of storage lifetime was established. In this model, using theWiener process simulates the mono-degradation process of each feature degradation;using the Frank Copula function describes the correlation among these featuredegradations. With the maximum likelihood estimation, the parameters in the Wienerprocess can be got; introducing the kendall’s tau, those in the Frank Copula functioncan be estimated. By conducting the regression analyses of the estimated values ofparameters in each stress, their corresponding converted equations can be shown.According to the data by the storage test, the new model was testified. And the resultverified the model can better predict the long-term storage lifetime of transistors.(3) The Support Vector Machine was introduced to evaluate the transistors’storage lifetime. By combining the results of both accelerated test and on-the-spotstorage test in the marine environment, the training and regression model wasestablished to predict the long-term storage lifetime.The results shows that the three prediction methods are all feasible: the methodusing the ideality factor and barrier height as the sensitive parameter, the model based on the Wiener Process of multi-performance parameter degradation and the methodbased on the Support Vector Machine. These methods provide the new ideas for thereliability assessment of electronic components.Finally, the paper discussed in detail the impacts of temperature, humidity, saltspray and other marine environmental stresses on the storage reliability of electroniccomponents, the measures were proposed to improve the reliability of long-termstorage reliability of transistors in the marine environment.The study provides the necessary research basis and methods for the storagereliability assessment of transistors in the marine environment, and also certainreference for the theoretical research of accelerated storage test and statistical analysesof reliability data. And the key technology and related methods can also be extendedand applied in other storage reliability studies of electronic components.