High-frequency noise modelling and the scaling of the noise parameters of polysilicon emitter bipolar junction transistors

1996 ◽  
Vol 74 (S1) ◽  
pp. 195-199 ◽  
Author(s):  
M. Jamal Deen

This paper presents detailed results from modelling the four noise parameters: minimum noise figure (NFMIN), noise resistance (RN), optimal source resistance (RS,OPT), and reactance (XS,OPT) as functions of frequency and collector-biasing current. Compared to previous BJT (bipolar junction transistor) high-frequency noise models, we include the emitter resistance, which results in an increased input device impedance, and a degeneration of the device transconductance. We also give an explicit formula for the noise resistance. We present noise results for polysilicon emitter bipolar transistors as a function of emitter areas to demonstrate how the noise parameters scale with emitter areas over a range of frequencies. However, these results are given only for devices in which the pad impedances are much larger than the device input impedance, so that very little input signal is lost through the pads to ground.

1996 ◽  
Vol 74 (S1) ◽  
pp. 200-204 ◽  
Author(s):  
M. J. Deen ◽  
J. Ilowski

This paper presents detailed experimental results on the input impedance of bonding pads, and a simple electrical model that accurately describes the impedances of these pads as a function of frequency for five different types of pad structures. It also describes the effects of the bonding pads on the minimum high-frequency-noise figure (NFMIN) of polysilicon emitter npn bipolar junction transistors as functions of collector current density and emitter areas. It was found that for devices with larger emitter areas (AE > 48 μm2) and relatively low base resistance (RB < 30 Ω), the effects of the same bonding pads on the noise figure was not as pronounced as for the smaller area devices with larger base resistances. For a device with AE = 3.2 μm2 operating at 1 GHz, biased with a collector density of 0.15 mA μm−2, neglect of the effects of the bonding pads results in too low NFMIN (by 1–2 dB) when calculated values were compared to measurements. Finally, for devices with the same emitter areas, bonding pads with smaller impedance results in a larger NFMIN compared to measurements on similar transistors with pads of larger impedances.


2004 ◽  
Vol 809 ◽  
Author(s):  
J.G Tartarin ◽  
G. Cibiel ◽  
A. Monroy ◽  
V. Le Goascoz ◽  
J. Graffeuil

ABSTRACTThe rapid expansion of SiGe technologies during the last decade essentially due to civil telecommunication's applications have led Si/SiGe based heterojunction bipolar transistors (HBTs) to excellent performance levels, allowing high frequency low noise circuit designs such as linear low noise amplifiers( RF noise) or also low-phase noise oscillators (LF noise). Among these technologies, the SiGe BiCMOS one integrates digital and RF functions on the same chip. Fast improvements of the technological process have been performed thanks to large efforts allowed to characterization and modeling of the devices. We have investigated on the influence of technological parameters such as Germanium profile, doping level and thickness of the base layer (5 different wafers) on the dynamic and high frequency noise performances to converge towards the optimum technological process (now available with the BiCMOS6G processed by ST microelectronics). We made use of scattering parameters [S] measurements on the devices to extract the electrical parameters of our small signal model. The high frequency noise parameters based on the electrical model (with noise sources added to the junction, resistances) are simulated and compared with the measured noise parameters of the devices. The four noise parameters (Fmin, Rn, and complex Γopt) measurements have been performed from 1 GHz to 12 GHz, and the dynamic S parameters measurements have been realized in the 40 MHz-40 GHz range. These models have been used to enable the identification of the limiting parameters on the dynamic performances and on the high frequency noise parameters.


Author(s):  
Asmaa Nur Aqilah Zainal Badri ◽  
Norlaili Mohd Noh ◽  
Shukri Bin Korakkottil Kunhi Mohd ◽  
Asrulnizam Abd Manaf ◽  
Arjuna Marzuki ◽  
...  

<p>This study reviews related studies on the impact of the layout dependent effects on high frequency and RF noise parameter performances, carried out over the past decade. It specifically focuses on the doughnut and multi- finger layouts. The doughnut style involves the polygonal and the 4- sided techniques, while the multi-finger involving the narrow-oxide diffusion (OD) and multi-OD. The polygonal versus 4-sided doughnut, and the narrow-OD with multi-fingers versus multi-OD with multi- fingers are reviewed in this study. The high frequency parameters, which are of concern in this study, are the cut- off frequency (f<sub>T</sub>) and the maximum frequency (f<sub>MAX</sub>), whereas the noise parameters involved are noise resistance (R<sub>N</sub>) and the minimum noise figure (NF<sub>min</sub>). In addition, MOSFET parameters, which are affected by the layout style that in turn may contribute to the changes in these high frequency, and noise parameters are also detailed. Such parameters include transconductance (G<sub>m</sub>); gate resistance (R<sub>g</sub>); effective mobility (μ<sub>eff</sub>); and parasitic capacitances (c<sub>gg</sub> and c<sub>gd</sub>). Investigation by others has revealed that the polygonal doughnut may have a larger total area in comparison with the 4- sided doughnut. It is also found by means of this review that the multi-finger layout style with narrow-OD and high number of fingers may have the best performance in f<sub>T</sub> and f<sub>MAX</sub>, owing partly to the improvement in G<sub>m</sub>, μ<sub>eff</sub>, c<sub>gg</sub>, c<sub>gd</sub> and low frequency noise (LFN). A multi-OD with a lower number of fingers may lead to a lower performance in f<sub>T</sub> due to a lower G<sub>m</sub>. Upon comparing the doughnut and the multi-finger layout styles, the doughnuts appeared to perform better than a standard multi-finger layout for f<sub>T</sub>, f<sub>MAX</sub>, G<sub>m</sub> and μ<sub>eff</sub> but are poorer in terms of LFN. It can then be concluded that the narrow-OD multi-finger may cause the increase of c<sub>gg</sub> as the transistor becomes narrower, whereas a multi-OD multi-finger may have high R<sub>g</sub> and therefore may lead to the increase of f<sub>T</sub> and f<sub>MAX</sub> as the transistor becomes narrower. Besides, the doughnut layout style has a higher G<sub>m</sub> and f<sub>T</sub>, leading to larger μ<sub>eff</sub> from the elimination of shallow trench isolation (STI) stress.</p>


2007 ◽  
Vol 42 (5) ◽  
pp. 1034-1043 ◽  
Author(s):  
Saman Asgaran ◽  
M. Jamal Deen ◽  
Chih-Hung Chen ◽  
G. Ali Rezvani ◽  
Yasmin Kamali ◽  
...  

2010 ◽  
Vol 25 (10) ◽  
pp. 105011 ◽  
Author(s):  
Juan M López-González ◽  
Paulius Sakalas ◽  
Michael Schröter

Micromachines ◽  
2020 ◽  
Vol 12 (1) ◽  
pp. 7
Author(s):  
Yu-Shyan Lin ◽  
Shin-Fu Lin

This study proposes AlGaN/GaN/silicon high-electron mobility transistors (HEMTs) grown by a metallorganic chemical vapor deposition (MOCVD) system. The large-signal linearity and high-frequency noise of HEMTs without and with different passivation layers are compared. The experimental data show that the addition of a TiO2 passivation layer to undoped AlGaN/GaN HEMT’s increases the value of the third-order intercept point (OIP3) by up to 70% at 2.4 GHz. Furthermore, the minimum noise figure (NFmin) of the HEMT with TiO2 passivation is significantly reduced.


2010 ◽  
Vol E93-C (5) ◽  
pp. 678-684 ◽  
Author(s):  
Hiroshi SHIMOMURA ◽  
Kuniyuki KAKUSHIMA ◽  
Hiroshi IWAI

Sign in / Sign up

Export Citation Format

Share Document