Research On The Interconnect Reliability And Temperature Reliability For RF Power Amplifier

A radio frequency power amplifier(RF PA)is the most critical module in the RF front end of wireless transmitters,and its reliability and performance usually affect the communication quality of the whole communication system directly.In recent years,with the continuous improvement of the integration,the feature size and the interconnect width for the RF IC decrease constantly,which can result in the increase of current density in interconnect sharply.Such a high current density may cause electromigration directly and even bring the problem of the fatal failure such as short-circuit or open-circuit in RF PAs.Therefore,the interconnect reliability of PA is the top concern for the designers.Meanwhile,the performance degradation and failure caused by the temperature have become a major factor for PA reliability.All these problems pose severe challenges for the reliability of RF PAs.Aimed at the two problems of interconnect failure and performance failure for RF PAs,this dissertation makes a further research on the interconnect reliability for the integrated PA and the temperature reliability for the discrete PA.Firstly,in view of the truth that the existing research only realized the interconnect reliability analysis at the DC condition for IC,here both of DC and AC interconnect reliability prediction and analysis for a Ga As MMIC PA are conducted.On the base of this work,due to the disadvantages of being time-consuming,resources-consuming and inconvenient for model modification using the traditional finite element analysis,by combining the artificial neural network and related technologies with the finite element analysis,to accelerate the interconnect reliability analysis for this PA is achieved.Finally,due to the truth that temperature is the most principal factor for RF PA's performance and reliability,a series of reliability experiments are conducted to study the temperature reliability and index failure for the GaN PA here.In particular,the mainly innovation achievements of this dissertation are as follows:1.In order to solve the problem that the finite element analysis is time-consuming,resources-consuming and inconvenient for model modification,this dissertation proposes to combine the finite element analysis with the artificial neural network and related technologies to speed up the interconnect reliability analysis for a GaAs MMIC P A.Furthermore,using the ANSYS parametric design language and the automatic model generating algorithm of the artificial neural network,the automatic construction of the modeling and the quick construction of the mapping relationship can be achieved.It is convenient to modify the model and impro ve the efficiency of modeling and analysis greatly.This can provide new idea for the automatic modeling of the complex IC and rapid interconnect reliability prediction and analysis.2.In order to analyze the relationship between the temperature and the performance parameters of the RF PA quantitatively,a series of reliability experiments are designed and conducted to study the specific influence of temperature variation on the reliability and performance for two different kinds of GaN PAs.This can provide some important guidance for the reliability design of RF PAs.3.On account of the continuous degradation of the RF PA's performance,a method of taking the performance index of RF PA as the degradation data and using the accelerated degradation test to study the mechanism of index failure for RF PAs is proposed here.By testing the performance degradation and failure under the different temperature stress,the actual acceptable working temperature range and the times of temperature shock for the given GaN PA can be obtained.All these can lay substantial foundation for the research of index failure for other electronic products.