Ultra Low Power (<2mW) Noise Performance of InGaPIGaAs HBT

The noise parameters of Npn InGaPIGaAs HBT devices are reported for an array of bias conditions from 2-18 GHz. A minimum noise figure of 0.8 dB with associated gain of 16 dB at 2 GHz was achieved at a collector bias of 0.83mA and a collector-emitter voltage of 2V. This was achieved by a device with emitter area of 3xlOp.111~. A model is also presented that is used to compare measured and theoretical results The low power consumption, high associated gain, low equivalent noise resistance, and variable optimum noise match coupled with an excellent degree of linearity (I) makes the InGaP/GaAs HBT an ideal candidate for low noise amplifters (2-6GHz) with minimal power consumption. Introduction: Portable, battery operated circuit applications require devices that provide minimum noise figure and maximum linearity with minimal power consumption. In particular, CDMA and PCM CIA communication applications call for these components to be designed for 2.4 and 5.8 GHz. By carefdly selecting the device emitter area and bias condition, the topology of the matching network and associated loss and added noise can be minimized (Z). InGaP/GaAs HBTs have lower number of recombination centers in the emitter-base junction and subsequently exhibit lower l/f noise than comparable AlG&GaAs HBTs. In fact the I/f noise performance is comparable to that of sicon based BJTs (3,4). The high frequency noise performance is improved with the InGaP/GaAs HBT compared to silicon through superior gain at microwave frequencies which is the direct result of a shorter emitter-collector transition time. We believe that the minimum noise figure presented in this work is among the lowest achieved for a bipolar technology in this fiequency band Device Fabrication and Measurement. The devices were fabricated and measured as described in (SI. Lower recombination current in the base-emitter junction fiom enhancements in material growth and lower base resistance through improvements in processing have contributed to minor improvements in overall noise performance over previously reported results All of the devices tested had a current gain of approximately fifty. Devices with 3x10p2, 3x5p2, 3x20m2, 2x3x10pz, 3x40p2, and 8x3x5p2 emitter areas were used for this study. Results are reported for the 3x10~~ device which achieved the best noise performance of the devices tested. Modelinn: The simplest form of the equivalent circuit for the HBT is shown in Figure 1. The shot noise current is simply determined from an examination of each junction within the transistor (6) The relation of the noise currents to the bias currents is given below: I = 2qIc4f