Device Research Conference 2007

Erik Lind, Adam M. Crook, Zach Griffith, and Mark J.W. RodwellDepartment of Electrical and Computer EngineeringUniversity of California, Santa Barbara, CA, 93106-9560, USA

Xiao-Ming Fang, Dmitri Loubychev, Ying Wu, Joel M. Fastenau, and Amy LiuIQE Inc. 119 Bethlehem, PA 18015, USA

lind@ece.ucsb.edu, 805-893-3273, 805-893-3262 fax

560 GHz ft, fmax InGaAs/InP DHBT in a novel dry-etched emitter process

Latest UCSB DHBT w/ refractory emitter metal

Scalable below 128 nm width

New refractory metal emitter technology 250 nm emitter width

0

5

10

15

20

25

30

109 1010 1011 1012

109

1010

1011

1012

Gai

ns (d

B)

Frequency (Hz)

U

H21

MAG/MSG

Ie=22 mA, V

ce=1.45V

Je=16 mA/m2, V

cb=0.4V

ft, f

max=560 GHz

First RF resultsSimultaneously 560 GHz ft & fmax

BVceo = 3.3V

Standard figures of merit / Effects of Scaling

Small signal current gain cut-off frequency (from H21)

cbcexcbjec

Bcb CRRCC

qITnk

f

21

effcbCRffeffbb ,,

max 8

VIC

c

cb

Charging time for digital logic

•Power gain cut-off frequency (from U)

Thinning epitaxial layers (vertical scaling) reduces base and collector transit times… But increases capacitances

Reduce Rbb and Ccb, through lateral scaling

More efficient heat transfer

e

e

InP

ceee

WL

KVWJT ln

emitter 500 250 125 nm width 16 9 4 m2 access

base 300 150 75 width, 20 10 5 m2 contact

collector 150 100 75 nm thick, 5 10 20 mA/m2 current density5 3.5 3 V, breakdown

f 400 500 700 GHzfmax 500 700 1000 GHz

power amplifiers 250 350 500 GHz digital clock rate 160 230 330 GHz(static dividers)

What parameters are needed for THz HBTs?

✓✓

✓✓

✓

✓✓ c<1 m2 U. Singisetti

DRC 2007

Ohmic contacts and epitaxial scaling good for 64nm HBT node!

Develop technology suitable for aggressive lateral scaling

TiW emitter dry etch formation

TiW

Cr

• Lift-off no good at <300 nm!• Dry etching – very high aspect ratio• Optical Lithography

O2 plasmaICP Cl2O2 etch

Mask Plate

SiO2

ICP SF6/Ar etch

•Refractory Ti0.1W0.9 is thermally stable

• c < 1m2 possible – no degradation of contact resistance for anneals up to 400C

TiW dry etching results• 250-300 nm emitters routinely formed• Demonstrated scalability down to 150 nm • ~ 50-100 nm tapering during etch sets scaling limit

•Emitter metal height ~ 500-600 nm

150 nm

Wet etch does not scale

{111}A planes – slow{101}A planes – fast

InGaAs etch ~ 30-40nmInP etch ~ 40 nm

Minimum emitter width ~ 150 nm!

Collector

Base

baseInP emitter

InGaAs emitter

TiW metal

Hybrid dry/wet etch• Anisotropic dry etch InGaAs, part of InP emitter, Cl2N2

• Formidable problem – InClx formation• Extensive UCSB know how on dry etching• Solution – low power ICP etch @ 200C, low Cl2 concentration• Short InP wet etch

TiW

InGaAs

InP

InGaAs Base

Current UCSB TiW emitter process

SiO2

TiW

InGaAs n++

InGaAs p++ Base

InP n

CrSF6/Ar ICP Cl2/N2 ICPSiNx sidewall

Wet EtchHCl:H3PO4

BHF

Ti

a b c d e

InGaAs p++ Base InGaAs p++ Base InGaAs p++ Base InGaAs p++ Base

InP nInGaAs n++ InGaAs n++

• 5 nm Ti layer for improved adhesion• 25 nm SiNx sidewalls protects Ti/TiW during Cl2 and BHF etch, improves adhesion• Standard triple mesa• BCB pasivation

Emitter prior to InP wet etch

Layer structure -- 70 nm collector DHBTThickness

(nm) Material Doping cm-3 Description

30 In0.53Ga0.47As 51019 : Si Emitter cap

10 In0.53Ga0.47As 41019 : Si Emitter

60 InP 31019 : Si Emitter

10 InP 1.21019 : Si Emitter

20 InP 1.01018 : Si Emitter

22 InGaAs 5-91019 : C Base

5.0 In0.53Ga0.47 As 21017 : Si Setback

11 InGaAs / InAlAs 21017 : Si B-C Grade

3 InP 6.2 1018 : Si Pulse doping

51 InP 21017 : Si Collector

5 InP 11019 : Si Sub Collector

5 In0.53Ga0.47 As 21019 : Si Sub Collector

300 InP 21019 : Si Sub Collector

Substrate SI : InP

Objective:• Thin collector and base for decreased electron transit time• Collector doping designed for high Kirk threshold, designed for 128 nm node• Setback and grade thinned for improved breakdown

Vbe = 1.0 V, Vcb = 0.0 V

Je = 0, 14, 21 mA/m2-3.0

-2.5

-2.0

-1.5

-1.0

-0.5

0.0

0.5

80 100 120 140 160 180 200 220 240

Ener

gy (e

V)

distance (nm)

emitterbase

collector

valence band

conduction band

Je = 0mA/m2, 15mA/m2, 30mA/m2

RF & DC data –70 nm collector, 22 nm base InP Type-I DHBT

0

5

10

15

20

25

30

109 1010 1011 1012

109

1010

1011

1012

Gai

ns (d

B)

Frequency (Hz)

U

H21

MAG/MSG

Ie=22 mA, V

ce=1.45V

Je=16 mA/m2, V

cb=0.4V

ft, f

max=560 GHz

0

5

10

15

20

0 0.5 1 1.5 2 2.5 3 3.5 4

BCDHJLBBBB

Vce

(V)

25 mW/m2

Peak fmax

Peak ft

J e (mA

/um

2 )

Aje= 0.25 x 4.5 m2

Ajc= 0.6 x 5 m2

BVceo

~ 3.3 V

Emitter width ~ 250 nm

First reported device with ft, fmax > 500 GHz

BVCEO ~ 3.3 V, BVCBO = 3.9 V (Je,c = 15kA/cm2)

Emitter contact (from RF extraction), Rcont < 5 m2

Base: Rsheet = 780 /sq, Rcont ~ 15 m2

Collector: Rsheet = 11.1 /sq, Rcont ~ 10.1 m2

10-11

10-9

10-7

10-5

10-3

10-1

0

5

10

15

20

25

0.0 0.2 0.4 0.6 0.8 1.0

I c, Ib (A

)

Vce

(V)

Current G

ain

Aje= 0.25 x 4.5m2, A

jc= 0.6 x 5m2

nc= 1.05

nb=2.01

Ic

Ib

RF data –70 nm collector, 22 nm base InP Type-I DHBT

4

5

6

7

8

9

10

3

4

5

6

7

5 10 15 20 25C

cb (f

F)

Ccb /A

je (fF/um2)

Je (mA/m2)

Vcb

= 0.4 V

0.0 V

Aje=0.25 x 5.5 m2 A

jc=0.6 x 6 m2

Figure 2. Common

-emitter I

-V characteristics

Vcb

= -0.1 V

0.1 V

0.2 V

0.3 V

450

475

500

525

550

575

600

10 12.5 15 17.5 20 22.5J

e (mA/m2)

f (GH

z)

Vcb

= 0.4V

0.0V

0.1V

-0.1V

• No detectable increase in Ccb for high Je

• Indicates that Kirk threshold is not reached• Device performance limited by self heating?• Further scaling needed for better thermal performance!

cb

f

c

cb

cbcb

cb

cc

cb

VIQ

VVQ

IIC

Kirk effect increases Ccb

...even in DHBTs

Ccb / Je

Breakdown performance of InP HBTs

1

10

100 1000

InP/GaAsSbInP/InGaAsInP SHBTThis work

Brea

kdow

n volt

age,

BVce

o (V)

f (GHz)

?

Type I DHBT and Type II DHBT – similar BVceo

Type II DHBTs has low fmax due to base resistance

SHBT have substantially lower breakdown

Ec

Ev

Ec

Ev

Type I DHBT

Type II DHBT

0

100

200

300

400

500

600

700

800

0 100 200 300 400 500 600 700 800

TeledyneUIUC DHBT NTTFujitsu HEMT

SFU/ETHzUIUC SHBT UCSB NGST Pohang SHBTHRLIBM SiGe

Vitesse

f max

(GH

z)

ft (GHz)

500 GHz400 GHz300 GHz200 GHz maxff

Updated July 2007

600 GHz

Current status of fast transistors

)( ),/( ),/( ),/(

hence , :digital

gain, associated ,F :amplifiers noise low

mW/ gain, associated PAE,

:amplifierspower

)11(

2/) (alone or

min

1max

max

max

max

cb

cbb

cex

ccb

clock

ττVIRVIRIVC

f

m

ff

ff

ffff

:metrics better much

:metrics popular

250-300nm

600nm

300-400nm

Emitter Process will scale to 128 nm & below

• Emitter-base junction has been scaled down to 64 nm• Pure W emitter- no tapering

61 nm emitter-base junction!

Tungsten

Emitter metal peeled off during cross section cleave..

Conclusion

• HBTs @ 128 nm node→ 700 GHz ft, xx GHz fmax feasblechallenges: contact resistivity, robust deep submicron processes

• Dry etch based DHBT emitter process → sub 100 nm junctions feasible

• First RF results: 250nm junctions: simultaneous ft,fmax ~ 560 GHz• 125 nm devices should come soon !

• … THz before next DRC?

This work was supported by DARPA SWIFT program and a Swedish Research Council grant.