Oxide Aperture Heterojunction Bipolar Transistors

Oxide Aperture Heterojunction Bipolar

Transistors

James G. Champlain

Ph.D. DefenseUniversity of California, Santa Barbara

February 22, 2002

PhDDefenseAcknowledgements

Advisor Umesh MishraCommittee Art Gossard, Evelyn Hu, and Jim SpeckMishra Group (old and new)

Ale, Ariel, Can, Dan, Dario, DJ, Gia, Haijiang, Huili, Ilan, Jae, Jason, Jeff, Lee,Likun, Mary, Naiqain, Nguyen, Peter, Prashant, Primit, Rama, Rob C., Rob U.,Sten, Tim, Yifeng, and Yingda

Yee Kwang for all the discussions on HBTs and his excellent simulationsYork Group

Amit, Angelos, Baki, Bruce, Chris, Hongtao, Jim, Joe, Justine, Nadia, Nick, Paolo,Pengcheng, Pete, Troy, and Vicki

MBEJohn English

System B Eric, Prashant, Richard, SheilaFriends

Todd, Dave, and JohnNick, Omer, Cathy, Fay, Rex, Fereshteh, Neil, Heather, and Stefanie

Peter, Lee, and Rama

Eric, Prashant, Richard, Sheila, Borys, Can, Dave, and Max

PhDDefenseOverview

Introduction

Oxide Aperture HBTs

Growth as related to bipolar transistors Growth of GaAs0.49Sb0.51, AlAs0.56Sb0.44, and AlxGa1-xAsySb1-y

n-type doping of arsenide-antimonides

Current density: area versus perimeter Aperture placement Diode Selection Rules

Generation Ø Generation I Generation II RF results

Transistor Design and Growth

Oxide Aperture Diodes

Conclusions and Future Work

PhDDefenseIntroduction

device circuit system


complexity


requiredoperatingfrequency

complexity

The ever-growing demand in communication and radar technology for increased bit-rates and frequency resolution requires systems capable of providing increased bandwidth and clock rates

Requires increased bandwidth from the devices, which are the building block of the system

As complexity increases, the max operating frequency decreases

A given system frequency, a higher device operating frequency is required

Given that these systems are built from circuits, which in turn are built from basic devices and passive elements

PhDDefensefττττ and fmax

fτ The current-gain cutoff frequency

fmax The maximum frequency of oscillation

Defined as the frequency at which the short-circuit current (h21) gain goes to unity

Defined as the frequency at which the unilateral power gain (U) goes to unity

( ) ( )12 BE BC B C E C BC

C

kTf C C R R CqIτη τ τ

π

= + + + + +

transit times

IC

emitter

collector

base

e-

IC

emitter

collector

base

e-

base

collector

emitter

IC IC

e

collector

charging times

8maxB BC

ffR C

τ

π=

PhDDefense

selective regrowth techniques

The Oxide Aperture HBT

8maxB BC

ffR C

τ

π= In order to increase fmax, RB and/or CBC must be reduced

Techniques such as increased base doping and lateral scaling can reduce RB; but an extrinsic portion of CBC still remains, limiting the devices frequency performance

intrinsic deviceE

B

Cextrinsic CBC

Various techniques have addressed the reduction of CBC

E

B

C

B

E

C

E

C

B

Oxide Aperture HBT

selective regrowth techniques, sidewall contacted basesselective regrowth techniques, sidewall contacted bases, undercut collectorsselective regrowth techniques, sidewall contacted bases, undercut collectors, implanted emittersselective regrowth techniques, sidewall contacted bases, undercut collectors, implanted emitters, and the transferred substrate technology

PhDDefenseOverview

Introduction

Oxide Aperture HBTs








Introduction

Oxide Aperture HBTs





PhDDefenseTransistor Design: Materials

The majority of high-speed HBTs produced today have been InP emitter-up transistors

n+ InGaAs emitter contact

n+ AlInAs emitter

n grade from InGaAsto AlInAs

p+ InGaAs base

n- InGaAs collector

n+ InGaAs sub-collector

AlInAs does not readily oxidize and cannot be easily used to form an oxide apertureRox,AlInAs = 2.4 µm/hr @ 520 °C

substrate

emitter

emitt

er

base

base

collector

colle

ctor

PhDDefenseTransistor Design: Materials

5.6 5.7 5.8 5.9 6.0 6.1 6.20.0

0.5

1.0

1.5

2.0

2.5

Band

gap

Ener

gy, E

g (eV

)

Lattice Constant, a (Å)5.6 5.7 5.8 5.9 6.0 6.1 6.2

0.0

0.5

1.0

1.5

2.0

2.5

InAs

AlAs

GaAs

Band

gap

Ener

gy, E

g (eV

)


0.0

0.5

1.0

1.5

2.0

2.5

InP

InAs

AlAs

GaAs

Band

gap

Ener

gy, E

g (eV

)


0.0

0.5

1.0

1.5

2.0

2.5

InP

InAs

AlAs

GaAs

Band

gap

Ener

gy, E

g (eV

)


0.0

0.5

1.0

1.5

2.0

2.5

InP

InAs

AlAs

GaAs

Band

gap

Ener

gy, E

g (eV

)


0.0

0.5

1.0

1.5

2.0

2.5

InP

In0.53Ga0.47As

Al0.48In0.52As

InAs

AlAs

GaAs

Band

gap

Ener

gy, E

g (eV

)


0.0

0.5

1.0

1.5

2.0

2.5

InP

In0.53Ga0.47As

Al0.48In0.52AsAlSb

GaSb

InAs

AlAs

GaAs

Band

gap

Ener

gy, E

g (eV

)


0.0

0.5

1.0

1.5

2.0

2.5

InP

In0.53Ga0.47As

Al0.48In0.52AsAlSb

GaSb

InAs

AlAs

GaAs

Band

gap

Ener

gy, E

g (eV

)


0.0

0.5

1.0

1.5

2.0

2.5

InP

In0.53Ga0.47AsGaAs0.49Sb0.51

Al0.48In0.52As

AlAs0.56Sb0.44

AlSb

GaSb

InAs

AlAs

GaAs

Band

gap

Ener

gy, E

g (eV

)

Lattice Constant, a (Å)

PhDDefense

AlAsSb emitter

Transistor Design: MaterialsOxide Aperture HBT

n- InGaAs collector

p+ GaAsSb base

n grade from AlAsSbTo GaAsSb

n+ AlAsSb emitter

n+ grade from GaAsSbTo AlAsSb

n+ GaAsSb sub-emitter

baseemitterAl(Ga)AsSb

grades

collector

substrate

Rox,AlAsSb = 5 µm/hr @ 350 °C

emitt

erba

seco

llect

or

PhDDefenseGrowth of GaAsSb/AlAsSb

Two methods to grow GaAsSb/AlAsSb

Digital Alloy of III-As/III-Sb Analog Alloy

Shutter Cycles

III-As

III-Sb

Sb Sb

III III

xx

Φ∝Φ

Pros: Highly reproducibleCons: Heavy wear & tear

Pros: Greatly reduced wear & tearCons: Requires growth

calibrations, maintenance

Growth of AlxGa1-xAsySb1-y

A fully digital growth (III-As/III-Sb) is undesirable due to the wear & tearA fully analog growth is practically impossible, requiring cals for each compositionTherefore, a digital alloy of the analog alloys GaAsSb/AlAsSb is used to form the

grades between GaAsSb and AlAsSb

PhDDefensen-type Doping of Antimonides

Silicon is the most widely used n-type dopant in III-V semiconductors

group III n-type

group V p-typeAs a result of being a group IV element,

it is also amphoteric by nature

For most III-V semiconductors, Si resides on the group III site as a donorConversely, in Sb-based semiconductors Si incorporates as an acceptor

Therefore, a group VI element, tellurium, is used as an n-type dopant for Sb-based compounds

One drawback encountered with doping of the Sb-based semiconductors is the poor mobility of electrons, especially at higher doping levels

1E16 1E17 1E18

100

1000

AlAs0.56Sb0.44

AlSb

GaAs

Hal

l ele

ctro

n m

obilit

y, µ

(cm

2 /V se

c)

Hall electron concentration, n (cm-3)

PhDDefense

Introduction

Oxide Aperture HBTs








Introduction

Oxide Aperture HBTs






Overview

PhDDefenseOxide Aperture Diodes

For every transistor, there are two diodes

All bipolar transistors are constructed from 2 pn diodesUnder forward-active operation, the base-emitter junction acts as a source

of carriers and the base-collector junction acts as a sink

Therefore, much of the transistors current and current-related characteristics can be related back to the base-emitter junction

Oxides formed from the wet oxidation of Al-based semiconductors have been shown to have fairly high interface recombination velocities

ee e

ee

emitter base collector

PhDDefenseOxide Aperture Diodes

Oxide aperture diodes were fabricated to understand the potential effect of the oxide aperture on the current-voltage characteristics of the BE junction of the HBT

p+ GaAsSb base


n+ AlAsSb emitter

n+ grade from GaAsSbTo AlAsSb


SI InP substrate

BE grade

emitter grade

tBE

2000 1500 1000 500 0-2.0

-1.6

-1.2

-0.8

-0.4

0.0

0.4

0.8

1.2

emittersub-emitter emittergrade BE grade base

E - E

F (eV

)

Distance from surface (Å)

Layer Structure Band Diagram

Three thickness of BE grade were examined: 100 Å, 500 Å, 1000 Å

PhDDefense

0.00 0.25 0.50 0.751E-16

1E-15

1E-14

1E-13

1E-12

1E-11

1E-10

1E-9

1E-8

1E-7

1E-6

1E-5

1E-4

K J

K (A

/µm

) and

J (A

/µm

2 )

V (V)

Current Density: A versus PIn addition to the ideality factor (η), the geometric dependence of the current

provides information on the nature of current conduction within a pn junction

In actual diodes, the total current can be written as a sum of an area dependent and perimeter dependent current measured j jI J A K P= ⋅ + ⋅

jmeasuredmeasured

j j

PIJ J KA A

= = + ⋅

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.90

5

10

15

20

25

increasing V

Vapp = 0.24 V Vapp = 0.30 V Vapp = 0.36 V

J = 0.15 nA/µm2

K = 3.02 nA/µm

J = 0.49 nA/µm2

K = 8.91 nA/µm

J = 1.87 nA/µm2

K = 28.6 nA/µm

I mea

sure

d / A

(nA

/µm

2 )

P / A (1/µm)0.0 0.1 0.2 0.3 0.4 0.5

0

2

4

ηk η j

Idea

lity

Fact

or

V (V)

For a given bias with Pj / Aj as the independent variable, J corresponds to the y-intercept and K to the slope of the equation

( )1j TV VSJ J e η= −

( )1k TV VSK K e η= −

PhDDefenseOxide Aperture Diodes: Results

tBE = 100 Å

B

E

depletionregion

tBE = 500 Å

B

E

tBE = 1000 ÅB

E

100 250 500 750 10000.8

1.0

1.2

1.4

1.6

1.8

2.0

2.2

2.4

ηunoxidized = 1

Idea

lity

Fact

or, η

Base-Emitter Grade Thickness, tBE (Å)

aperture perimeter aperture area > aperture area

1.36Between mesa & aperture area1000 Å1.01Aperture area500 Å1.88Aperture perimeter100 Å~1Mesa areaunoxidized

ηηηηDominant geometrytBE

PhDDefenseOxide Aperture Diodes: Sims

Simulations performed using ATLAS simulation software generally agree with the experimental results

tBE = 1000 ÅtBE = 500 ÅtBE = 100 Å

Simulations performed without interface recombination

PhDDefenseOxide Diodes: Selection Rules

The Three Bears

(1) The oxide aperture must effectively channel carriers injected from the emitter into the base

(2) The current-voltage characteristic should scale with aperture area

(3) The ideality factor (η) should be as close to ideal (η → 1) as possible

The Three Bears:The Three Bears: ones too shortThe Three Bears: ones too short, ones too longThe Three Bears: ones too short, ones too long, and ones just right

B

E

depletionregion

tBE = 100 Å

B

E

tBE = 500 Å

B

E

tBE = 1000 Å

B

E

depletionregion

tBE = 100 Å

PhDDefenseOverview

Introduction

Oxide Aperture HBTs








Introduction







PhDDefenseOxide Aperture HBTs

Three generations of oxide aperture HBTs were fabricated:

Generation Ø: The first generation of oxide aperture HBTs, fabricated before the oxide aperture diode study

Generation I: The first generation of HBTs to show true transistor action, fabricated after the oxide aperture diode study

Generation II: The first generation of oxide aperture HBTs to produce high-frequency results

-2.0

-1.5

-1.0

-0.5

0.0

0.5

1.0 InGaAs collectorGa(Al)AsSb

emitterGaAsSb

base

E - E

F (eV

)

n- InGaAs collector

p+ GaAsSb base


n+ AlAsSb emittern+ grade from GaAsSb

To AlAsSb


Laye

r Str

uctu

re

Ban

d D

iagr

am

PhDDefenseOxide Aperture HBT: Fabrication

SiO2 dummy collector

1 Deposit dummy collector

2 Etch base mesa

3 Oxidize emitter aperture

4 Etch collector mesa

5 Deposit B/E contacts

6 Etch emitter mesa

7 Deposit AC lines

8 Planarize and deposit posts

9 Deposit interconnects

PhDDefenseOxide Aperture HBT: Fabrication

HBTs of various sizes

Misaligned HBTsTLMs

cell

CPW lines

HBT

cell

HBT

close-up of HBT

emitter

base

collector

PhDDefense

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

2.0

Ga(Al)AsSbgrade

Ga(Al)AsSbemitter InGaAs collectorInGaAs

base

E - E

F (eV

)

Oxide HBTs: Generation Ø

Key Differences of Gen Ø HBT

InGaAs base Abrupt BE junction 2000 Å collector

η ≈ 1.7 β on the order of 10-6 ~ 10-5

Non-existent IC, low ββββ, and high ηηηη all suggest that the characteristic is recombination dominated

Results prompted the oxide diode study0.0 0.5 1.0 1.5 2.0 2.51E-11

1E-10

1E-9

1E-8

1E-7

1E-6

1E-5

1E-4

1E-3

0.01

0.1

IB IC

I B and

I C (A)

VBE (V)

0

2

4

6

8

10

12

14

β (10-6)

β

PhDDefenseOxide HBTs: Generation I

-2.0

-1.5

-1.0

-0.5

0.0

0.5


emitterGaAsSb

base

E - E

F (eV

)

0.0 0.5 1.0 1.51E-7

1E-6

1E-5

1E-4

1E-3

0.01

I B and

I C (A)

VBE (V)

IB IC

0

25

50

75

100

β

β

0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.750

2

4

6

8

10IB = 0 ~ 90 µA IB, step = 10 µA

I C (mA

)

VCE (V)

Key Differences of Gen I HBT

Lower doping in base and emitter 2000 Å collector

β ≈ 90

η ≈ 2.95

η ≈ 3.34

Fabricated from same material as the oxide diodes

PhDDefenseVoffset and Vknee

Voffset : VCE at which IC = 0

Vknee : VCE at which the device switches between forward-saturation and forward-active operation (VBC ≈ 0)

( ),

,

ln 1bc sat bc bcoffset bc BE B B B B

F be sat be be

IkTV V I R I Rq I

η ηηα η η

+ − − + !

( ) ( ), ,

ln ln1 1

be E R C bc F E Cknee E E C C

be sat F R bc sat F R

kT I I kT I IV I R I Rq I q I

η α η αα α α α

− −− + + − − !

Definitions

Vbkd

VoffsetVknee

VCE

IC

Modified Ebers-Moll Model

RB

RE RC

intrinsic model

PhDDefense

VCE

IC

VCE

IC

VCE

IC

VCE

IC

increasing RE

VCE

IC

The Knee Voltage: Vknee

( ) ( ), ,

ln ln1 1

be E R C bc F E Cknee E E C C

be sat F R bc sat F R

kT I I kT I IV I R I Rq I q I

η α η αα α α α

− −− + + − − !

0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.750

2

4

6

8

10IB = 0 ~ 90 µA IB, step = 10 µA

I C (mA

)

VCE (V)

REVknee

Model Measured

PhDDefense

0.0 0.5 1.0 1.5 2.0 2.50.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

V CE (V

)

IB (mA)

Generation I HBT

Generation I: Emitter Resistance

The emitter resistance (RE) of a bipolar transistor can be measured by floating-collector measurements

RE

RC

RB

IB

VCE

IB

VCE

REIB

0.0 0.5 1.0 1.5 2.0 2.50.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

V CE (V

)

IB (mA)

0

200

400

600

800

1000

RE (O

hms)

Generation I HBT

RE ≈ 85 ΩΩΩΩ

0.0 0.5 1.0 1.51E-11

1E-10

1E-9

1E-8

1E-7

1E-6

1E-5

1E-4

1E-3

0.01

0.1

I (A

)

V (V)

0

200

400

600

800

1000

rd (Ohm

s)

oxide diode

rd ≈ 75 ΩΩΩΩ

At a current level of ~4 mA with RE ≈ 85 ΩΩΩΩ ! Vknee ≥ 0.34 V

PhDDefenseOxide HBTs: Generation II

-2.0

-1.5

-1.0

-0.5

0.0

0.5


emitterGaAsSb

base

E - E

F (eV

)

What have we learned so far

Generation Ø and oxide diodesOxide aperture placement is critical

Generation IResistance in the emitter due to the low electron mobility in AlAs0.56Sb0.44

is a problem (high Vknee)

Generation II

Oxide aperture is located 500 Å from the pn junction (same as Gen I)

Increased doping in the emitter and base

Increased collector width to 3000 Å

The Three Bears: cant be too close, cant be too far, has to be just right

PhDDefense

4 x 10 HBT2 x 10 HBT1 x 10 HBTAE = 3.5 x 9.5 µm2

AC = 4 x 10 µm2

Generation II: Gummel and CE

0.00 0.25 0.50 0.75 1.00 1.25 1.501E-11

1E-10

1E-9

1E-8

1E-7

1E-6

1E-5

1E-4

1E-3

0.01

0.1

IB IC

I B and

I C (A)

VBE (V)

0

1

2

3

4

β

β

0.0 0.2 0.4 0.6 0.8 1.0 1.20

1

2

3

4

IB = 0 ~ 1.5 mA IB, step = 150 µA

I C (mA

)

VCE (V)

0.00 0.25 0.50 0.75 1.00 1.25 1.501E-11

1E-10

1E-9

1E-8

1E-7

1E-6

1E-5

1E-4

1E-3

0.01

0.1

IB IC

I B and

I C (A)

VBE (V)

0

2

4

6

8

10

12

14

16

β

β

0.0 0.2 0.4 0.6 0.8 1.0 1.20

2

4

6

8

10

12

14

16IB = 0 ~ 1.25 mA IB, step = 125 µA

I C (mA

)

VCE (V)

0.00 0.25 0.50 0.75 1.001E-11

1E-10

1E-9

1E-8

1E-7

1E-6

1E-5

1E-4

1E-3

0.01

0.1 IB IC

I B and

I C (A)

VBE (V)

0

10

20

30

40

50

60

β

β

0.0 0.2 0.4 0.6 0.8 1.0 1.20

5

10

15

20

25

30IB = 0 ~ 0.5 mA IB, step = 50 µA

I C (mA

)

VCE (V)

AE = 0.5 x 9.5 µm2

AC = 1 x 10 µm2AE = 1.5 x 9.5 µm2

AC = 2 x 10 µm2

ββββ ≈ 2.6

ηηηη ≈ 1

ββββ ≈ 11

ηηηη ≈ 1

ββββ ≈ 40

ηηηη ≈ 1ηηηη ≈ 1.18ηηηη ≈ 1.18 ηηηη ≈ 1.18

PhDDefense

B

E

C

0.00 0.04 0.08 0.12 0.16 0.20 0.240

10

20

30

40

50

60

70

80

R E (Ohm

s)

1 / AE (1/µm2)

Generation II: Emitter Resistance

0 1 2 3 4 50.14

0.16

0.18

0.20

0.22

0.24

0.26

0.28

0.30

0.32

4 x 10 HBT 2 x 10 HBT 1 x 10 HBT

V CE (V

)

IB (mA)0 1 2 3 4 5

0.14

0.16

0.18

0.20

0.22

0.24

0.26

0.28

0.30

0.32

4 x 10 HBT 2 x 10 HBT 1 x 10 HBT

V CE (V

)

IB (mA)

Floating-collector measurements

RE ≈ 20 Ω

RE ≈ 34 Ω

RE ≈ 79 Ω

0.00 0.04 0.08 0.12 0.16 0.20 0.240

10

20

30

40

50

60

70

80

R E (Ohm

s)

1 / AE (1/µm2)

0.00 0.04 0.08 0.12 0.16 0.20 0.240

10

20

30

40

50

60

70

80

R E (Ohm

s)

1 / AE (1/µm2)

0.00 0.04 0.08 0.12 0.16 0.20 0.240

10

20

30

40

50

60

70

80

R E (Ohm

s)

1 / AE (1/µm2)

decreasing junction

resistance

extrinsic resistnace

RE,int RE,acc

Emitter resistance (RE) has reduced as a direct result of increased doping

RE increases linearly with 1 / AE Suggests internal resistance (RE,int)

dominates of lateral access resistance (RE,acc)

AE

PhDDefense

emitter

base

coll

Generation II: Gain VariationThe reduced gain with decreasing

emitter/collector area is most probably a result of lateral diffusion of carriers

10 15 20 25 30 35 400

10

20

30

40

50

β

AC (µm2)

measured β

10 15 20 25 30 35 400

10

20

30

40

50

β

AC (µm2)

measured β simulated β

Simulations

-6-4

-20

24

6

0.00

0.25

0.50

0.75

1.00

-6

-4

-2

0

2

46

J / JE

y (µm

)

x (µm)

1 x 10

-6-4

-20

24

6

0.00

0.25

0.50

0.75

1.00

-6

-4

-2

0

2

46

J / JE

y (µm

)

x (µm)

2 x 10

-6-4

-20

24

6

0.00

0.25

0.50

0.75

1.00

-6

-4

-2

0

2

46

J / JE

y (µm

)

x (µm)

4 x 10

PhDDefense

Impact Ionization

Generation II: Output Conductance

0.0 0.5 1.0 1.50

5

10

15

20

25

30

I C (mA

)

VCE (V)

The output conductance (go) of the oxide aperture HBT is much higher than expected

The common-emitter breakdown voltage (Vbkd) is lower than expected

Early Effect

Early Effect

The Early effect (AKA base-width modulation) describes the increase in output current due to a change in the minority carrier profile in the base as a result of depletion at the BC junction

Only transistors with fairly low doped bases suffer from the Early Effect

-0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.00.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

I C (A)

VCE (V)-0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

I C (A)

VCE (V)-0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

I C (A)

VCE (V)-0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

I C (A)

VCE (V)

-VA

1E-4 1E-3

1

10

4 x 10 HBT 2 x 10 HBT 1 x 10 HBT

Extra

cted

VA (V

)

IB (A)

2 4 6 8 10101

102

103

104

105

GaAs InP InGaAs

α n (cm

-1)

Inverse electric field, 1/E (10-6 cm/V)

-0.8

-0.6

-0.4

-0.2

0.0

0.2

0.4

0.6

0.8

1.0

1.2

E - E

F (eV

)

e-

0.31 eVCE∆ !e- e-

h+

Impact Ionization

PhDDefense

E

B

C

E

B

C

Generation II: Misaligned HBTsIn conventional HBTs, the active collector area is larger than the active

emitter area

In oxide aperture HBTs, the active collector area and emitter area are roughly the same size

E

C

B

E

C

B

Aligned Misaligned

Aligned Misaligned

PhDDefense

+ 2 µµµµm

0.0 0.2 0.4 0.6 0.8 1.0 1.20.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

IB = 0 ~ 0.5 mA IB, step = 50 µA

I C (mA

)

VCE (V)

+ 1 µµµµm

0.0 0.2 0.4 0.6 0.8 1.0 1.20.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

IB = 0 ~ 0.5 mA IB, step = 50 µA

I C (mA

)

VCE (V)

Generation II: Misaligned HBTs

+ 0 µµµµm

0.00 0.25 0.50 0.75 1.001E-11

1E-10

1E-9

1E-8

1E-7

1E-6

1E-5

1E-4

1E-3

0.01

0.1 IB IC

I B and

I C (A)

VBE (V)

0

10

20

30

40

50

60

β

β

ββββ ≈ 400.00 0.25 0.50 0.75 1.00 1.25 1.50

1E-11

1E-10

1E-9

1E-8

1E-7

1E-6

1E-5

1E-4

1E-3

0.01

0.1 IB IC

I B and

I C (A)

VBE (V)

0.0

0.5

1.0

1.5

2.0

β

β

ββββ ≈ 1.250.00 0.25 0.50 0.75 1.00 1.25

1E-11

1E-10

1E-9

1E-8

1E-7

1E-6

1E-5

1E-4

1E-3

0.01

0.1 IB IC

I B and

I C (A)

VBE (V)

0.0

0.1

0.2

0.3

0.4

0.5

β

β

ββββ ≈ 0.25

0.0 0.2 0.4 0.6 0.8 1.0 1.20

5

10

15

20

25

30IB = 0 ~ 0.5 mA IB, step = 50 µA

I C (mA

)

VCE (V)0.0 0.2 0.4 0.6 0.8 1.0 1.2

0

5

10

15

20

25

30IB = 0 ~ 0.5 mA IB, step = 50 µA

I C (mA

)

VCE (V)

0.0 0.2 0.4 0.6 0.8 1.0 1.20.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

IB = 0 ~ 0.5 mA IB, step = 50 µA

I C (mA

)

VCE (V)0.0 0.2 0.4 0.6 0.8 1.0 1.2

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

IB = 0 ~ 0.5 mA IB, step = 50 µA

I C (mA

)

VCE (V)

AC = 4 x 10 µm2 and AE = 3.5 x 10 µm2 for all HBTsΙΒ = 0 ~ 0.5 mA at 50 µA steps

PhDDefenseRF Measurements

Scattering parameter (S-parameter) measurements were performed on the Generation II HBTs in order to evaluate the transistor's frequency performance

a1 a2

b1 b2

DUT

From S-parameters, characteristics like the short-circuit current gain (h21) and unilateral power gain (U) can be extracted

Subsequently, the current-gain cutoff frequency (fτ) and maximum frequency of oscillation (fmax) can be determined

1E8 1E9 1E100

2

4

6

8

10

12

14

16

18

20

22

h 21 a

nd U

(dB)

f (Hz)1E8 1E9 1E10

0

2

4

6

8

10

12

14

16

18

20

22

h 21 a

nd U

(dB)

f (Hz)

fττττ ≈ 8.3 GHzfmax ≈ 17.76 GHz

h21

U

PhDDefense

0 20 40 60 80 100

2

4

6

8

10

12

14

1 x 10 HBT 2 x 10 HBT 4 x 10 HBT

f τ and

f max

(GH

z)

JE (kA/cm2)

Measured fττττ and fmax

fmax

fττττ

The measured values of fτ and fmaxfor the oxide aperture HBTs were lower than expected

Measured fττττ and fmax

fττττ = 8.3 GHzfmax = 17.76 GHz

Maximum measured fmax

for a 2 x 10 HBT (IC ≈ 9 mA, VCE = 2 V)

Calculated fττττ and fmax

Calculations using the measured RE, RB, and assuming a fully depleted collector suggest: fττττ ≈ 25 GHz fmax ≈ 35 GHz

fτ and fmax were calculated using the classical hybrid-πmodel for a bipolar transistor

( ) ( )1

12

TBE BC B C E C BC

C

Vf C C R R CIτ

η τ τπ

−

= + + + + + 8max

B BC

ffR C

τ

π=

PhDDefense

B

E

CRCRB

RE

CBC

CDCBErπ

intrinsic

standard

gmvπ

Parasitic Base-Emitter Capacitance

Re-examining the of the oxide aperture HBT structure suggests the presence of a parasitic capacitance in the base-emitter resulting from the collector-up design

E

C

B

“intrinsic” deviceparasitic capacitance

Modified modelBy using MoTC, the parasitic capacitance leads to an additional delay time

within fτ 1

, ,1

2T B

x std E xC o

V Rf R CIτ τ

ητπ β

−

= + + +

, ,, 8 8

std xmax x

B BC B BC

f ff

R C R Cτ τ

π π= ≠

,,

12std

stdfττ

τπ

=Cx

standard

PhDDefense

C

B

E

( ) ( )1

12

TBE BC B C E C BC

C

Vf C C R R CIτ

η τ τπ

−

= + + + + +

( )1

12

T E CBE E BC C B C BC C

C C E C

V r rf c A c A c AJ A A Aτη τ τ

π

−

= + + + + +

fττττ ScalingIn a transistor where AC ≈ AE, fτ should be essentially independent of emitter size

( ) ( )1

12

TBE BC B C E C BC

C

Vf c c r r cJτ

η τ τπ

−

= + + + + +

In the oxide aperture HBT, this is not true1

, ,1

2T B

x std E xC o

V Rf R CIτ τ

ητπ β

−

= + + +

0 5 10 15 20 25 30 352

4

6

8

10

12

measured

f t (GH

z)

AE (µm2)

Cx

wE

lE

0 5 10 15 20 25 30 352

4

6

8

10

12

measured modeled

f t (GH

z)

AE (µm2)

By applying the above equation, fτ,std = 1/2πττ,std ≈ 26.3 GHz and Cx ≈ 345 fF

PhDDefenseRF Device Modeling

Circuit designers require device models in order to design accurately and model circuits

Device models are useful to device engineers because they can lend insight into the operation of the device, in addition to verifying proposed phenomenon

Proper high-frequency model extraction requires:

(1) A well-behaved, stable device

(2) Y-parameters that accurately represent the device

Model Extraction

Y-parameters that accurately represent the transistor require prior knowledge of the model ??

PhDDefenseDevice Model

Using measured characteristics (RB, RE, βo, gm) and initial estimates of other characteristics (rBC, CBC, Cx) a simple model was generated

rBC = 6.5 kΩ

RB = 83 Ω

RE = 34 Ω

Cx = 250 fF

rπ = 36.5 Ω CBE,D = 2 pF gm = 0.35e-jω(4.4 psec)

B C

E

CBC = 25fF

S11

S22

S21 / 2

S12 x 5

1E8 1E9 1E100

4

8

12

16

20

24

U

h21

measured

h 21 a

nd U

(dB)

f (Hz)1E8 1E9 1E10

0

4

8

12

16

20

24

U

h21

measured modeled

h 21 a

nd U

(dB)

f (Hz)

PhDDefense

1E8 1E9 1E100

4

8

12

16

20

24

U

h21

measured

h 21 a

nd U

(dB)

f (Hz)1E8 1E9 1E10

0

4

8

12

16

20

24

U

h21

measured modeled

h 21 a

nd U

(dB)

f (Hz)

Modified Device Model

S11

S22

S21 / 2

S12 x 5

rBC = 6.5 kΩ

RB = 83 Ω

RE,int = 30.5 Ω

Cx = 280 fF

rπ = 36.5 Ω CBE,D = 2 pF gm = 0.35e-jω(4.4 psec)

B C

E

CBC = 25fF

RE,acc = 3.5 Ω

A model which includes the lateral emitter access resistance (RE,acc)results in a slightly more accurate model

PhDDefenseModel without Cx

Removal of the parasitic capacitance radically affects the S-parameters, resulting in a marked increase in fτ with a minimal perturbation in fmax

S21 / 2

S12 x 5

S11 S22

1E8 1E9 1E100

4

8

12

16

20

24

U

h21 measured

h 21 a

nd U

(dB

)

f (Hz)1E8 1E9 1E10

0

4

8

12

16

20

24

U

h21 measured modeled

h 21 a

nd U

(dB

)

f (Hz)

PhDDefenseOverview

Introduction

Oxide Aperture HBTs








Introduction

Oxide Aperture HBTs







PhDDefenseConclusions

The validity and application of the oxide aperture HBT design for high maximum frequency of oscillation (fmax) was examined

The work culminated with the fabrication of an oxide aperture HBT with an fτ on the order of 8 ~ 10 GHz and fmax of 12 ~ 17 GHz, with the maximum measured fmax being 17.76 GHz

Two critical aspects concerning the design and fabrication of the oxide aperture HBT became apparent:

(1) The impact of the oxide aperture on the DC operation of the BE junction and thereby the HBT

(2) The effect of the parasitic BE capacitance on the high-frequency performance of the HBT

The Three Bears

1

, ,1

2T B

x std E xC o

V Rf R CIτ τ

ητπ β

−

= + + +

PhDDefenseFuture Work

The research presented here represents a foundation

Suggested areas of research:

for reduced base-collector capacitance and increased fmax

sub-micron scaling

for reduced base transit times and increased fτbandgap graded base layers

Specific areas that must be examined in future research:

(1) collector designfor reduced output conductance and increased common-emitter breakdown

(2) base-emitter junction optimizationfor reduced emitter resistance and improved performance

(3) self-limiting oxidationfor sub-micron scaling and reduced parasitic base-emitter capacitance


collector designThe high output conductance in the devices in this

work were attributed to impact ionization in the InGaAs collector

-1.2

-0.9

-0.6

-0.3

0.0

0.3

0.6

0.9

1.2

E - E

F (eV

)

-1.2

-0.9

-0.6

-0.3

0.0

0.3

0.6

0.9

1.2

E - E

F (eV

)A simple solution is to use an InP collector

Experiments by other groups have shown devices fabricated with this BC combination have increased output resistances and breakdown voltages

base-emitter junction designThe BE junction in this work is not the

optimum design

High RE due to low mobility AlAsSb in the emitterAlternatives:

InP-based emitterAlternative oxide source material: AlInAs

-2.0

-1.6

-1.2

-0.8

-0.4

0.0

0.4

0.8

1.2 Ga(Al)AsSbgrades

GaAsSbsub-emitter

GaAsSbbase

E - E

F (eV

)

-2.0

-1.6

-1.2

-0.8

-0.4

0.0

0.4

??

emitter GaAsSbbaseInPInP AlAsSb

E - E

F (eV

)


self-limiting oxidationThe only means to control oxidation depth currently

is time and temperature

0 10 20 30 40 50 60 70 80 900

5

10

15

20

25

30

35

40

45 325 °C 350 °C 360 °C 375 °C

Oxi

datio

n de

pth,

d (µ

m)

Oxidation time, t (min)

This required the base mesa to be similar to and centered on the collector mesa

As a result, the oxide depth is equidistant from the outside edge of the oxidation mesa

oxide aperture HBT conventional HBT Possible Solutions

dox

dox

Stress/Strain Engineering:By inducing strain selectively under

the collector, it may be possible to alter the oxidation conditions

Selective Disordering:By generating disorder only outside

of the collector, it may be possible to increase the oxidation outside of the collector

Documents

Oxide Aperture Heterojunction Bipolar Transistors