Reliability Evaluation of InGaAsN for PA Handset Applications

An initial assessment of the reliability of InGaP/InGaAsN transistors was investigated in order to evaluate this new material system for cell phone applications. The investigation consisted of single temperature HTOL with a modest case temperature of 200C, Vce of 5.0V, and a current density of 25kA/cm. InGaP/InGaAsN variants and InGaP/GaAs controls were stressed at the same time. The failure mode for both the InGaP/InGaAsN and InGaP/GaAs transistors was identified through IV characterization as Beta degradation due to an increase in Ib. All other device parameters are stable. The Beta degradation data was used to construct a probability plot. The InGaP/InGaAsN transistors had greater or comparable reliability to the control population, thus indicating that this material has sufficient reliability for cell phone applications. INTRODUCTION InGaAsN material has been receiving a great deal of attention for its potential applications in both optical and electronic devices [1]. In particular, the lower band-gap and device turn-on, when used as the base on an HBT, provide additional margin for handset PA designs. However, in current commercial applications, reliability is a primary concern of handset customers The InGaAsN material system grown on GaAs is often a strained material (lattice mis-matched) system. Theoretically, this places a strain limit on how thick this layer can be before dislocations form that will dramatically degrade the device characteristics and reduce the device reliability [2]. Hence, a thorough understanding of device reliability over a range of base thicknesses is critical for both device design and understanding the manufacturing tolerance of this material system. In this feasibility study, we evaluate the reliability of this new technology. This evaluation is designed to provide forward knowledge of the reliability and confidence that the lifetimes are sufficient for cell phone applications. To accomplish this, devices from this new technology with a range of base thicknesses and controls from a qualified technology (InGaP/GaAs) were subjected to single temperature HTOL. By comparing the lifetimes of a sufficiently sized population through the use of a lognormal probability plot, we provide confidence that the technology has sufficient reliability. EXPERIMENTAL Since this material system is strained, there is a limit to how thick the base layer can be grown. Calculations show that this value is theoretically close to 900 A. However, work in other systems show that this theoretical limit can be violated [3]. Also, early tests in our laboratory with this material showed a high degree of instantaneous burn-in. As a result of these two factors, we studied devices with 500 A and 900 A base thickness possessing both high and low burn-in. The Gummel plots for the different variants are shown in Figure 1. While this is not a characterization report we note the lower Vbe turn-on and no abnormalities in series resistance, gain, or leakage. An InGaP/GaAs Gummel plot is shown for reference. 1.0E-12 1.0E-11 1.0E-10 1.0E-09 1.0E-08 1.0E-07 1.0E-06 1.0E-05 1.0E-04 1.0E-03 1.0E-02 1.0E-01