complete analysis compound fets based transient...

VLSI DESIGN1998, Vol. 8, Nos. (1-4), pp. 313-317Reprints available directly from the publisherPhotocopying permitted by license only

(C) 1998 OPA (Overseas Publishers Association) N.V.Published by license under

the Gordon and Breach SciencePublishers imprint.

Printed in India.

Complete RF Analysis of Compound FETs Basedon Transient Monte Carlo Simulation

S. BABIKER, A. ASENOV *, N. CAMERON, S. P. BEAUMONT and J. R. BARKER

Department of Electronics and Electrical Engineering, Nanoelectronics Research Centre, Glasgow University,Glasgow G12 8LT, Scotland, UK

In this paper we described a complete methodology to extract the RF performance of’real’ compound FETs from time domain Ensemble Monte-Carlo (EMC) simulationswhich can be used for practical device design. The methodology is based on transientfinite element EMC simulation of realistic device geometry. The extraction of theterminal current is based on the Ramo-Shockley theorem. Parasitic elements like thegate and contact resistances are included in the RF analysis at the post-processing stage.Example of the RF analysis of pseudomorphic HEMTs illustrates our approach.

Keywords: Monte-Carlo, RF analysis, compound FETs, simulation

1. INTRODUCTION

The remarkable RF performance of compoundFETs such as GaAs MESFETs and InGaAsHEMTs with channel lengths down to 0.1 lam isdue to well pronounced velocity overshoot, Theuse of simulation for predictive analysis anddesign of such devices require in many cases theemployment of full scale EMC technique [1-3].However, most of the published EMC studies ofcompound FETs consider simplified device geo-metry and focus mainly on the transport physicsand the effect of the enhanced channel velocity onthe DC device characteristics. Far more impor-tant for the proper design of modern short

channel compound FETs is the RF performancewhich is determined not only by the high field

transport but also by the device geometry and thesurface effects. The T- or I-shaped gate, the gaterecess and the passivation in such devicescritically affect the device parasitics and theoverall RF device performance.

In this paper we describe a methodology basedon the EMC simulation to investigate the RFperformance of FETs. The terminal currents areestimated using the Ramo-Shockley theorem. Thedevice parasitics are included through the properfinite-element description of the gate and recessshapes. The external parasitics are included in thepost-processing stage.

* Corresponding author.

313

314 S. BABIKER et al.

2. TRANSIENT CURRENT

The Heterojunction 2D Finite Element simulator

(H2F) and its Monte-Carlo module are describedin detail elsewhere [4, 5]. In the MC simulation thetotal transient terminal current in response to a

step change in the applied voltages required for they-parameter extraction is the sum of the particlecurrent and the displacement current. Accordingto the Ramo-Shockley theorem [6] the instanta-neous transient current in electrode due to Ndiscrete moving charges within the device is givenby the sum of the current I’i(t solely contributedby the movement of the N charged particles withfixed potentials at electrodes and the current I’i’(t)induced due to the time-varying potentials of theelectrodes through the capacitive coupling throughthe electrodes. The current Ii(t) is given by

N

Ii(t) Z qjvj(t) f (1)j--1

where qj. is the charge of the super-particle, v.i(t isthe velocity of the particle andf. is the solution ofthe Laplace equation V.(cg7jl.)=0 with a unitvoltage applied to electrode i, while all otherelectrodes are grounded.The current Ii’(t) associated with the time

varying potential is calculated from the capaci-tance matrix components Cij associated with theelectrodes of the simulated device. The capacitancematrix components are obtained from the solutionof the Laplace equation as Cij-- AQi/AVj whereAQi is the change in the electrode charge inresponse to a change in the potential of a

particular electrode. When a step perturbationA Vj. of the terminal voltage is applied Ii’(t) flowsonly during the first time step At and is given by:

Ii’(t) CiiA

(2)At

The displacement current during the remainingpart of the transient is related to the chargesinduced on the electrode by the moving particles in

the device associated with the redistribution of themobile charge and is accounted for by Eq. (1).

3. RF ANALYSIS

The flow diagram of the complete time domain RFEMC analysis is given in Figure 1. The y-para-meters are calculated by Fourier decomposition ofthe current transients obtained in response to stepperturbations in the terminal voltages [7]. The cut-off frequency of the simulated device j’m isextracted by solving log[Gsim(logf)]-0 where

G. did/dig is the current gain expressed as afunction of y-parameters.

In order to extract the maximum frequency ofoscillation of the simulated device fsim themax

y-parameters are transformed into s-parametersSsim The maximum frequency of oscillation/’simmax

FIGURE Flow diagram of the RF EMC analysis.

RF ANALYSIS 315

is then extracted by solving log[MAGsim(logf)] =0where MAGsim is the maximum available gain.For typical MESFET and HEMT simulation

domains, we have adopted the equivalent circuitmodel presented in Figure 2. The gate resistance,the contact resistances and any external inductivecomponents are excluded from the EMC simula-tion. The source Rsl and the drain Rdl resistancesin Figure 2 represent the resistance of the regionsbetween the gate and the source and drain ohmiccontacts respectively. To extract accurately thesmall signal equivalent circuit the y-parameters ofthe simulated device ysimare transformed into z-

parameters Zsim. The estimated source and drainresistances Rsl and Rdl are subtracted from Zsim toobtain the z-parameters Zint of the ’intrinsic’device. Finally Zint are transformed back into y-parameters yint from which the components of the’intrinsic’ small signal circuit can be analyticallyextracted [7]:

In order to evaluate the cut-off frequency f,}ealand the maximum frequency of oscillationsfeaxl ofthe ’real’ device the gate resistance Rg, the contactresistances R and eventually the inductive com-

ponents Lg, L and Ld first have to be incorporatedin the z-parameters of the real device Zreal whichare then transformed into yral. From yreal andSral the figures of merit of the ’real’ device we canestimate feal and frealmax. The intrinsic minimumnoise figure is also evaluated from the two-port y-parameters and the current traces.

GLg Rg

Simulation

Ls

R Ld

Real

oD

4. RESULTS

We apply the described RF analysis in thesimulation of a 120nm T-gate InGaAs channelpHEMT with 22 nm gate to channel separation [8]The y-parameters extracted from the Fourierdecomposition of the transients are shown inFigures 3 a,b. Table I summarises the small signal

.--.--.-.-.-.--.-.-.-.--,--.---o--- Re(Y11)

o Im(Y11)---o--- Re(Y21 ....

$- Im(Y21 ..o.-

io 4’o do ioFrequency (GHz)

0.04

0.03

-0.01-

----m-- Re(Y12o Im(Y12)

---o--- Re(Y22a- Im(Y22

,’" "0""""0-..0,......Ii20 40 60 80 100 120

Frequency (GHz)

FIGURE 2 Small signal equivalent circuit of a real FET.Figure also shows the ’intrinsic’ and ’simulated’ deviceequivalent circuit.

FIGURE 3 Y-parameters as a function of frequency extractedfrom the Fourier decomposition of the current transients atVg 0.2 and Va 1.5 V.

316 S. BABIKER et al.

TABLE Small signal equivalent circuit components calculated from the EMC simulations, Vg =-0.2 and Va 1.5 V. Given alsoare the experimental values, fT and fmax are given at Rg Rc 5 f

Cgs(fF) Cgd(fF) Cds(fF) gmo(mS) gds(fF) Ri(f) r(pS) fT(GHz) fmax(GHz)MC 75.5 14.2 15.1 87.6 11.2 2.3 0.17 142 164Exp. 77.8 8.73 14 67 8.65 3.75 0.32 114 150

700

600

500-

400

3oo-

200

100

00

Rc=0 f2

\ Rc=la\ Rc=2f,,.,. \ Rc=3,.",,,,. Rc=4D,",3,,\ a:Q"’"’-._’N’,, " Experiment

10Gate Resistance (f)

FIGURE 4 Effect of gate and contact resistances on fmax.

0.4-

0.3

0.2

0.1

00 20 40 60 80 100 120

Frequency (GHz)

FIGURE 5 Intrinsic noise figure for the 120nm pHEMTextracted from the current traces at Vg-- 0.2 and Va 1.5 V.

equivalent circuit parameters extracted from theEMC simulations, Vg=-0.2 and Vd 1.5V, to-gether with the experimental values. The highersimulation gmo andfr values is due to the fact thatthe MC simulations overestimate the velocityovershoot in the channel. The significant effectthe gate and contact resistances have on themaximum frequency of oscillation is depicted inFigure 4. The intrinsic noise figure extracted fromthe transients is shown in Figure 5.

5. CONCLUSION

In this paper we have presented a comprehensivemethodology for the RF analysis of FETs. Themethodology is based on the transient EMCsimulations of the intrinsic device followed by a

post-processing stage during which the parasiticelements are included. It allows realistic estimationof the RF performance from MC simulations. Thecapabilities of the scheme are illustrated inexample simulations for the 120nm gate lengthpHEMT.

References

[1] Park, D. H. and Brennan, K. F. (1990). "Monte Carlosimulation of 0.35-1am gate-length GaAs and InGaAsHEMT’s", IEEE Trans. Electron Devices, 37, 618-628,

[2] Kizilyally, I. C., Artaki, M., Shah, N. J. and Chandra, A.(1991). "Scaling properties and short-channel effects insubmicrometer A1GaAs/GaAs MODFET’s: A MonteCarlo study", IEEE Trans. Electron Devices, 40, 234-249.

[3] Kim, K. W., Tian, H. and Litteljohn, M. A. (1991)."Analysis of delta doped and uniformly doped A1GaAs/GaAs HEMT’s by ensemble Monte Carlo simulation",IEEE Trans. Electron Devices, 38, 1737-1742.

[4] Babiker, S., Asenov, A., Barker, J. R. and Beaumont, S. P.(1996). "Finite element Monte Carlo simulation of recessgate compound FETs", Solid State Electronics, 39, 629-635.

[5] Asenov, A., Reid, D., Barker, J. R., Cameron, N. andBeaumont, S. P. (1993). "Finite element simulation of

RF ANALYSIS 317

[6]

[7]

[8]

recess gate MESFETs and HEMTs. The Simulator H2F",in Simulation of Semiconductor devices and processes, S.Selberherr, H. Stippel, E. Strasser, eds., Wien: SpringerVerlag, 5, 265-268.Kim, H., Min, H., Tang, T. and Park, Y. (1992). "AnExtended Proof of the Ramo-Shockley Theorem", SolidState Electronics, 34(11), 1251 1253.Gonzfilez, T. and Pardo, D. (1990). "Monte Carlodetermination of the intrinsic small-signal equivalent circuitof MESFET’s," IEEE Trans. Electron Devices, 42, 605-611.Cameron, N., Taylor, M. R. S., Mclelland, H., Holland,M., Thayne, I., Elgaid, K. and Beaumont, S. P. (1995). "Ahigh performance, high yield, dry etched pseudomorphicHEMT for W-band use", IEEE Trans. Microwave Theoryand Techniques Symposium Digest, Orlando, FL, 435-438.

Authors’ Biographies

Sharief Babiker has studied Electrical Engineeringin Khartoum University Sudan (1979-1984).During the period 1985-90 he worked at theElectrical Engineering Department of KhartoumUniversity. In 1994 he was awarded the Ph.D.degree for work on the theory and modelling ofsingle-electronic devices and systems. His currentresearch interests concentrate on the simulationand modelling of FETs with emphasis on sub-micron gate length InGaAs channel pHEMTs forRF applications.Asen Asenov had 10 years industrial experience

as a head of the Process and Device ModellingGroup in IME-Sofia, developing one of the firstintegrated process and device CMOS simulatorsIMPEDANCE. He was visiting professor at thePhysics Department of TU Munich, and iscurrently a Reader in the Department of Electro-nics and Electrical Engineering, Glasgow Univer-sity. As a leader of the Device Modelling Group hehas contributed to the development of 2D and 3Ddevice simulators and their application in thedesign of FETs, SiGe MOSFETs and IGBTSs.

Nigel Cameron graduated from the University ofBath in 1984 and joined British Telecom’s researchlaboratories to work on plasma process researchand development for silicon fabrication. In March1988 he moved to the Nanoelectronics ResearchCentre at the University of Glasgow where his

responsibilities include the development of ultra-fast transistors in compound semiconductors andtheir integration into manufacturable processes formillimetre-wave integrated circuits (MMICs). Hisinterests extend across a broad range of semicon-ductor technologies encompassing materials anddevice physics; circuit design and test; manufactur-ing; and dc and high frequency characterisation.Nigel Cameron is the author or co-author of morethan 35 technical papers.

Steven Beaumont was educated at the Universityof Cambridge and has been with the Departmentof Electronics and Electrical Engineering at theUniversity of Glasgow since 1978. He becameHead of Department in 1995 and he convenes theNanoelectronics Research Centre’s managementcommittee. His research interests lie in the field ofnanometre-scale fabrication and its application toelectronic and optoelectronic devices. He has over100 publications on electron beam nanolithogra-phy, dry etching, short-gate III-V based transis-tors, quantum transport devices, the opticalproperties of quantum dots, and single electrondevices. Latterly he has become involved with theissue of manufacturability of mm-wave circuitsand the use of nanometre-scale fabrication techni-ques coupled with technology-based device simu-lations to forecast performance and yield with theminimum of process iterations.John Barker is Professor of Electronics in the

Department of Electronics and Electrical Engi-neering. He has a long standing interest incomputational methods device modelling andtransport theory. From 1970-85 he was a memberof the Theory group in the Dept. of Physics,University of Warwick, aside from 1978-79 whenhe worked at IBM T. J. Watson Laboratory,North Texas State University and Colorado StateUniversity. From 1987-89 he was academicdirector of the IBM UK/Glasgow UniversityKelvin Project on Numerically Intensive ParallelComputing. He is academic director of the ParallelProcessing Centre at the University of Glasgow.

International Journal of

AerospaceEngineeringHindawi Publishing Corporationhttp://www.hindawi.com Volume 2010

RoboticsJournal of

Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014


Active and Passive Electronic Components

Control Scienceand Engineering

Journal of



RotatingMachinery


Hindawi Publishing Corporation http://www.hindawi.com

Journal ofEngineeringVolume 2014

Submit your manuscripts athttp://www.hindawi.com

VLSI Design



Shock and Vibration


Civil EngineeringAdvances in

Acoustics and VibrationAdvances in



Electrical and Computer Engineering

Journal of

Advances inOptoElectronics

Hindawi Publishing Corporation http://www.hindawi.com

Volume 2014

The Scientific World JournalHindawi Publishing Corporation http://www.hindawi.com Volume 2014

SensorsJournal of


Modelling & Simulation in EngineeringHindawi Publishing Corporation http://www.hindawi.com Volume 2014


Chemical EngineeringInternational Journal of Antennas and

Propagation




Navigation and Observation



DistributedSensor Networks


complete analysis compound fets based transient...

Documents