EKV3.0 model code & parameter extraction

EKV3.0 model code &parameter extraction

Antonios BazigosNational Technical University of Athens, [email protected]

Matthias BucherTechnical University of Crete, Chania, [email protected]

EKV Users’ Meeting/WorkshopNovember 4-5, 2004, EPFL, Lausanne, Switzerland

Friday, 5 November 2004 EKV3.0 Workshop 2

Outline

! EKV3.0 Verilog-A code! Phenomena covered! Parameter extraction methodology! Illustration of model characteristics! Summary


The Verilog-A code at a glance

! The Verilog-A code of EKV3.0" Contained in one file “ekv3.va”" ~1400 lines (46KB)" ~50 intrinsic model parameters" BSIM4-like junction diode models & effects" Optional source/drain, gate, substrate resistors" Developed using ELDO and SPECTRE" Used as the reference code for all model implementations" ADMS is being used in order to obtain “standard” C code

versions for various simulators.


Verilog-A code vs. C code

! Verilog-A code of EKV3.0 has a brother in C code (ELDO)" Hand-coded for comparison with ADMS-created C

code" C code allows implementation of advanced aspects

(NQS, NQS noise) " Verilog-A and C codes otherwise have the same

functionality" Writing Verilog-A code is simpler but code is less

efficient (2-3 times slower than C).


if (f_x_res==1) beginVS=V(si,b); VD=V(di,b);

endelse begin

VS=V(s,b); VD=V(d,b);end

Choosing to use the internal nodes or not

Few, brief excerpts of the Verilog-A code of EKV3.0 1/2

module ekv3(d,g,s,b);inout d,g,b,s;electrical d,g,s,b;electrical di,si,gi,bi;

Defining external and internal nodes of the module

parameter real p_cox = 0.012 from (0.0:inf);parameter real p_xj = 20.0E-09 from (0.0:inf);parameter real p_vto = 0.3 from (-inf:inf);parameter real p_phi_f = 0.45 from [0.1:inf);parameter real p_gamma_b = 0.3 from (0.0:inf);parameter real p_gamma_g = 4.1 from (0.0:inf);

Defining parameters, default and acceptable values


Few, brief excerpts of the Verilog-A code of EKV3.0 2/2

//RSCELeff_o_LR = Leff/p_lr;one_lr = 1.0 - exp(-Leff_o_LR*Leff_o_LR);dvtrsce = 2.0*p_qlr*one_lr/(cox*Leff_o_LR*UT);

Calculating the ∆VTRSCE

I(b,g) <+ ddt(QB); Describing the AC behaviour of the transistor. QB is the charge related to the substrate node.

if (f_x_res==1) beginV(d ,di) <+ I(d,di)*rlx;V(s ,si) <+ I(s,si)*rlx;I(di,si) <+ IDS;

end else beginI(d ,s ) <+ IDS;V(di,b ) <+ 0.0; V(si,b) <+ 0.0;

end

Describing the external resistors. In case they are not used the internal nodes are grounded to bulk node.


Phenomena covered by EKV3.0 --Associated parameters 1/2

--NQS (AC, noise)

GAMMA2, VR, DVRNon-Uniform Vertical Doping

IBA, IBB, IBNImpact ionization currentKG, XB, UBGate currents (IGS, IGD, IGB)

LOV, GAMMAOV(NOV), VFBOVBias-dependent Overlap Capacitances

KP(U0), E0, E1, ETAZC, THC

Mobility (reduction due to vertical field effect) Covering: Surface Roughness, Phonon Scattering, Coulomb Scattering

COX(TOX), PHIF, GAMMA(NSUB), VTO(VFB), GAMMAG(NGATE)

Physical Modelling of Charges Including Accumulation RegionCovering Polysilicon Depletion and Quantum mechanical effects

Modelled effect Related parameters


Phenomena covered by EKV3.0 --Associated parameters 2/2

various parameters (7)Temperature Effects

Various parameters (DL, WQLR, …)Geometrical effects, Width scaling

LETA, LETA2, WETASource and Drain Charge Sharing

LR, QLR, NLRReverse Short Channel Effect

WR, QWR, NWRInverse Narrow Width Effect

ETAD, SIGMADDrain Induced Barrier Lowering

AF, KFNoise (Flicker / Thermal)

UCRIT(VSAT), LAMBDA,DELTA

Longitudinal Field EffectVelocity Saturation, Channel Length Modulation

TOTAL

Modelled effect Related parameters

~60


Basic parameter extraction methodology

CGG vs VG:COX, VTO,

GAMMA, PHIF, GAMMAG

CGG vs VG:COX, VTO,

GAMMA, PHIF, GAMMAG

gm vs VG (lin):DL, RSX

(fixing RSCE for correct VTO)


(fixing RSCE for correct VTO)

ID vs VG (lin):LETA, [ETAD]


VTO vs L:LR, QLR, NLR

RSCE,[LETA, LETA2,

ETAD]


RSCE,[LETA, LETA2,

ETAD]

CGG vs VG:LOV, GAMMAOV,[VFBOV], DLC


Wide Long CVWide Long CV

Wide Long IVWide Long IV

Wide Short IVWide Short IV Wide All Lengths IV

Wide All Lengths IV

Wide Short CVWide Short CV

Narrow channel similar procedure


Narrow shortcombined effects

[fine tuning]


[fine tuning]

ΤΕΛΟΣ -- THE END -- FIN

gm vs VG (lin):KP, E0, E1,

[ETA]

gm vs VG (lin):KP, E0, E1,

[ETA]Width scaling:

All lengths w.r.t. width

Width scaling:All lengths w.r.t.

width

ID vs VG (sat):ETAD, [LETA]


Id, gds vs VD [strong inversion]:

UCRIT,LAMBDA, DELTA

[weak inversion]:ETAD


UCRIT,LAMBDA, DELTA


Temperature analysis



Refined parameter extraction methodology

CGG vs VG:COX, VTO,

GAMMA, PHIF, GAMMAG,[AQM]

CGG vs VG:COX, VTO,

GAMMA, PHIF, GAMMAG,[AQM]






RSCE,[LETA, LETA2,

ETAD]


RSCE,[LETA, LETA2,

ETAD]



Wide Long CVWide Long CV

Wide Long IVWide Long IV

Wide Short IVWide Short IV Wide All Lengths IV

Wide All Lengths IV

Wide Short CVWide Short CV




[fine tuning]


[fine tuning]

gm vs VG (lin):KP, E0, E1,[ETA]gm vs VG (lin):

KP, E0, E1,[ETA]


width


width




UCRIT,LAMBDA, DELTA



UCRIT,LAMBDA, DELTA


gm vs VG (sat):ZC, THC

gm vs VG (sat):ZC, THC

VP, n vs VG (lin):LETA

VP, n vs VG (lin):LETA

VP, n vs VG (lin):[GAMMA], [VTO]

VP, n vs VG (lin):[GAMMA], [VTO]

n0 vs L:LETA, LETA2

n0 vs L:LETA, LETA2

ΤΕΛΟΣ -- THE END -- FIN




Example parameter set for 0.12µm CMOS

+ LDW = 1.5n+ SIGMAD = 1+ WDL = 0+ ETAD = 0.41+ DW = -32n*** DIBL+ DL = -23n

+ VFBOV = 250m+ WETA = 450m*** Geometrical+ GAMMAOV = 2.5+ LETA2 = 0+ DELTA = 1.65+ LOV = 30.0n+ LETA = 385m+ LAMBDA = 0.2+ AQM = 0.5*** Overlap Caps*** Charge Sharing+ UCRIT = 3.6MEG***QM [fine tuning]+ TE1EX = 1.5+ NWR = 5.5m*** VSAT / CLM+ N0 = 1.04+ TE0EX = -15+ QWR = 470u+ THC = 0+ GAMMAG = 14+ TLAMBDA = 0+ WR = 80n+ ZC = 1.0+ GAMMA = 0.25+ UCEX = 0.8*** INWE+ ETA = 1.8+ PHIF = 0.5+ TETA = 6.25m+ NLR = 150m+ E1 = 380.0MEG+ VTO = 225m+ BEX = -1.8+ QLR = 2.8m+ E0 = 10.0G+ XJ = 20.0n+ TCV = 579u+ LR = 55n+ KP = 524u+ COX = 12.25m*** Temp params*** RSCE*** Mobility***General


0

0.2

0.4

0.6

0.8

1

-3 -2 -1 0 1 2 3VG [V]

CG

G /

(W*L

*CO

X) [-

]

Polysilicon depletion effectGAMMAG

GAMMAGGAMMAG 10 VGAMMAG 6 V

nmosW=10uL=10u


Quantum mechanical effectAQM

0

0.2

0.4

0.6

0.8

1

-3 -2 -1 0 1 2 3VG [V]

CG

G /

(W*L

*CO

X) [-

]

QM OFF;GAMMAGQM ON; GAMMAGQM ON; GAMMAG 6 V

nmosW=10uL=10u


Overlap CapacitanceGAMMAOV

0.0E+00

2.0E-08

4.0E-08

6.0E-08

8.0E-08

-3 -2 -1 0 1 2 3VG [V]

CG

I [F]

GAMMAOV 1.5 VGAMMAOV 1.0 VGAMMAOV 0.7 V

Dashed lines: Overlap capacitance

nmosW=10uL=120n


0.0E+00

2.0E-08

4.0E-08

6.0E-08

8.0E-08

1.0E-07

-3 -2 -1 0 1 2 3VG [V]

CG

X [F

]

Short-channel CV

CGG

CGB

CGD, CGS

nmosW=10uL=120n


0.0E+00

1.0E-06

2.0E-06

3.0E-06

4.0E-06

5.0E-06

6.0E-06

7.0E-06

8.0E-06

9.0E-06

0 0.5 1 1.5VG [V]

gm [m

ho]

0.0E+00

1.0E-06

2.0E-06

3.0E-06

4.0E-06

5.0E-06

6.0E-06

7.0E-06

8.0E-06

9.0E-06

0 0.5 1 1.5VG [V]

ID [V

]Transconductance factor

KP

nmosW=10uL=10u

V DS 50mV

KP 200uKP 150uKP 100u


0.0E+00

5.0E-06

1.0E-05

1.5E-05

2.0E-05

2.5E-05

0 0.5 1 1.5VG [V]

gm [m

ho]

0.0E+00

5.0E-05

1.0E-04

1.5E-04

2.0E-04

2.5E-04

3.0E-04

3.5E-04

4.0E-04

0 0.5 1 1.5VG [V]

gm [m

ho]

Mobility reduction due to vertical field E0, E1, [ETA]

E0 1G; E1 600ME0 600M;E1 600 ME0 1G; E1 400ME0 600M;E1 400Mnmos

W=10uL=10u OX

bi

CQQE

EE

EE

η

µµ

+=

++

=

⊥

⊥⊥

2

10

0

1

V DS 50mV V DS 1.5V


-0.4

-0.2

0

0.2

0.4

0.6

0.8

1

1.2

-0.2 0.3 0.8 1.3

VG [V]

VP [V

]

1

1.05

1.1

1.15

1.2

1.25

1.3

-0.2 0.3 0.8 1.3

VG [V]

n [-]

Pinch-off voltage & slope factor GAMMA

VTO

1−

=

G

P

dVdVn

GAMMA 0.50 VGAMMA 0.35 VGAMMA 0.20 V

nmosW=10uL=10u

VP vs VG measurement setup.*IB biases the device to

moderate inversion


0.0E+00

2.0E-04

4.0E-04

6.0E-04

8.0E-04

1.0E-03

1.2E-03

1.4E-03

1.6E-03

1.8E-03

0 0.5 1 1.5VG [V]

gm [m

ho]

0.0E+00

2.0E-04

4.0E-04

6.0E-04

8.0E-04

1.0E-03

1.2E-03

1.4E-03

1.6E-03

1.8E-03

0 0.5 1 1.5VG [V]

gm [m

ho]

Leff & series resistance effectsDL, RS

DL 30nDL 20nDL 10n

RS 00.0µΩmRS 30.0µΩmRS 60.0µΩm

DLLL Drawneff +=

nmosW=10uL=120n

V DS 50mV


Velocity saturation & CLM effectsLAMBDA, UCRIT

0.0E+00

5.0E-04

1.0E-03

1.5E-03

2.0E-03

2.5E-03

3.0E-03

3.5E-03

4.0E-03

4.5E-03

0 0.2 0.4 0.6 0.8 1VD [V]

ID [A

]

1.0E-05

1.0E-04

1.0E-03

1.0E-02

1.0E-01

0 0.2 0.4 0.6 0.8 1VD [V]

gds

[mho

]

LAMBDA 0.5LAMBDA 0.3LAMBDA 0.1

0.0E+00

5.0E-04

1.0E-03

1.5E-03

2.0E-03

2.5E-03

3.0E-03

3.5E-03

4.0E-03

4.5E-03

0 0.2 0.4 0.6 0.8 1VD [V]

ID [A

]

1.0E-05

1.0E-04

1.0E-03

1.0E-02

1.0E-01

0 0.2 0.4 0.6 0.8 1VD [V]

gds

[mho

]

UCRIT 5.0MUCRIT 4.0 MUCRIT 3.0M

nmosW=10uL=120n


Scaling due velocity saturationDELTA, UCRIT, LAMBDA

0.4

0.5

0.6

0.7

0.8

0.9

1

1.00E-07 1.00E-06 1.00E-05 1.00E-04L [m]

IDSA

T / I

DSA

T0 [-

]

DELTA 2.0;UCRIT 3.1MEG ;LAMBDA 0.25DELTA 1.7;UCRIT 3.5 MEG;LAMBDA 0.20DELTA 1.0;UCRIT 5.0 MEG;LAMBDA 0.12


-0.5

-0.3

-0.1

0.1

0.3

0.5

0.7

0.9

-0.2 0.3 0.8 1.3

VG [V]

VP [V

]

1

1.05

1.1

1.15

1.2

1.25

1.3

1.35

1.4

-0.2 0.3 0.8 1.3

VG [V]

n [-]

Pinch-off voltage & slope factorLETA

1−

=

G

P

dVdVn

nmosW=10uL=120n

LETA 0.0LETA 0.5LETA 1.0


0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

1.0E-06 1.0E-04 1.0E-02 1.0E+00 1.0E+02ID/ISPEC [V]

gms*

UT/

ID [-

]

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

0 0.5 1 1.5

VG [V]

ID [A

]

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

1.0E-06 1.0E-04 1.0E-02 1.0E+00 1.0E+02ID/ISPEC [V]

gmg*

UT/

ID [-

]

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

-0.2 0 0.2 0.4 0.6 0.8 1

VS [V]

ID [A

]

DIBL effectETAD, [SIGMAD]

V GB 0.25VV GB 0.50VV GB 0.75VV GB 1.00V

V SB 0.0VV SB 0.2VV SB 0.4VV SB 0.6V

V DB 1.5V

nmosW=10uL=120n

ETAD=0 (Dashed lines)ETAD=1 (Solid lines)


DIBL effect on gdsETAD, [SIGMAD]

1.0E-08

1.0E-07

1.0E-06

1.0E-05

1.0E-04

1.0E-03

1.0E-02

1.0E-01

0 0.2 0.4 0.6 0.8 1VD [V]

gds

[mho

]

ETAD=0 (Dashed lines)ETAD=1 (Solid lines)

nmosW=10uL=120n

V GB 1.00VV GB 0.75VV GB 0.50VV GB 0.25V

0.0E+00

2.0E-03

4.0E-03

6.0E-03

8.0E-03

1.0E-02

1.2E-02

1.00E-11 1.00E-09 1.00E-07 1.00E-05 1.00E-03

ID [A]

gds*

UT/

ID [-

]ETAD 0.75ETAD 0.70ETAD 0.65ETAD 0.60


0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

1.00E-07 1.00E-06 1.00E-05L [m]

VT [V

]Reverse short channel effect

QLR

QLR 0.8e-3QLR 0.6e-3QLR 0.4e-3

VSB= 0.0V

VSB= 0.2V

VSB= 0.4V

DIBL, CS RSCE

LRL

C

eQLR

Veff

OX

LRL

T

eff

⋅

−⋅⋅

=∆

−

2

12


0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

1.00E-07 1.00E-06 1.00E-05L [m]

VT [V

]Reverse short channel effect

NLR

NLR 30.0e-3NLR 20.0e-3NLR 10.0e-3

1

1.05

1.1

1.15

1.2

1.25

1.3

1.00E-07 1.00E-06 1.00E-05L [m]

n0

VSB= 0.0V

VSB= 0.2V

VSB= 0.4V

LRL

C

eNLR

effOX

LRL

BB

eff

⋅

−⋅⋅

+⋅Γ=′Γ

−

2

12

1


Reverse short channel effect LR

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

1.00E-07 1.00E-06 1.00E-05L [m]

VT [V

]

LR 300nmLR 200nmLR 100nm

VSB= 0.0V

VSB= 0.2V

VSB= 0.4V


0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

1.00E-07 1.00E-06 1.00E-05L [m]

VT [V

]Charge sharing

LETA

LETA 0.3LETA 0.6LETA 0.9

1

1.02

1.04

1.06

1.08

1.1

1.12

1.14

1.16

1.18

1.2

1.00E-07 1.00E-06 1.00E-05L [m]

n0

VSB= 0.0V

VSB= 0.2V

VSB= 0.4V


Threshold voltage vs. W & L


Slope factor vs. W & L


Summary

! EKV3.0 code is developed in Verilog-A" Verilog-A is the main platform of code development" Code tested in various simulators, among which ELDO &

Spectre.! A basic methodology of extracting EKV3.0

parameters has been presented" Methodology is similar over many CMOS generations" Effects of main EKV3.0 parameters on DC and CV

characteristics have been illustrated.

Thank you very much for your time and attention

Documents

EKV3.0 model code & parameter extraction