1

Harvey N. Rutt, Suresh UppalChong Yew Lee

Optoelectronics Research CenterUniversity of Southampton,Southampton SO17 1BJ

Design of an Electrically Operated mid-infrared solid-state

modulator

AITA7th September 2005

2

Outline

• Background• Choice of Material• The physics• Electrical modulator• Design criteria • Simulation results• Outlook

3

The need for a modulator

• Built in calibration, ‘thermal referencing’

• Variable IR transmission in situations where mechanical solutions are undesirable

• Image from pyroelectric detectors requires modulation

• IR Spatial Light Modulators?

4

Desiderata!

– Polarisation insensitive– Copes with low F number beams, to F#1– High on state transmission (>95%)– Low power consumption– High off state attenuation– Large aperture (to 1cm2)– Full 8-14μm band– Fast – but slow OK for many applications.

Difficult to satisfy simultaneously!

5

For a semiconductor, carrier based device we need:

• A stronger absorption mechanism than the free-carrier Drude-Zener

• A way to maintain large carrier densities at low power cost

• Hence a long lifetime

• Infrared transparency

• Availability!

6

Choice of Material

• Why Ge– Indirect bandgap – long intrinsic lifetime– IR transparent 1.8—18 um– Available in high purity (use in Nuc. Det.)– High carrier mobility– Availability + cost + ease of fabrication– p-type Ge has lh-hh interband transitions in

required spectral range

7

The principle

Band diagram of Ge showing heavy and light hole band

8

Absorption spectra

Wavelength (m)

20.0

10.06.7

5.0

4.0

3.3

2.9

2.5

2.2

2.0

1.8

1.7

Tra

nsm

issi

on (

%)

0

10

20

30

40

50

heavy hole - spin orbit

light hole -spin orbit

heavy hole -light hole

multi phonon band

transitions

Absorption spectra of intrinsic and p type Ge

i

9

Optical modulator

10

Optical Modulator in action

Off state

ON state

11

Basic structure

Optimise:• Doping levels• Width of p, i, and n regions etc•Processing

Note:The IR has to traverseThe doped regions.

12

Design Criteria

OPTICAL ON state (diode electrically off)Ton=To exp(-A) = T0 exp -(Np*xp*σh)

OPTICAL OFF state (diode electrically on)Toff=To exp(-A) = T0 exp (np*σh)

[np=0∫t

cp(x)dx]

Where To=1 for 100% transmission A is the absorption σh is hole absorption cross-section (5.33x10-16cm2) Np is the doping density for holes (/cm3) np is the area carrier density when forward biased (p+i region) xp is the thickness of the p layer

13

Design Criteria

• ON state: max Ton minimum Np*xp

• OFF state: min Toff maximum np

Uniform current injection requires high doping density (Np)

ABSORPTION vs UNIFORMITY trade-off!

14

Lifetime requirement

• We had σh =5.33x10-16cm2

• And Toff=T0 exp (np* σh) [np=0∫t

cp(x)dx]

• Lets assume a uniform hole density• And a 1mm thick device

• And require Toff=0.01 (ie, 1%)

• Then we need np=8.64*1015/cm2

• Or Np=8.64*1016/cm3

15

Design Criteria: required lifetimeRough analytical estimates:

0 0.5 1 1.5 2 2.5 3 3.5 40

10

20

30

40

50

60

70

80

90

100

Foward current density (A/cm2)

To

ff (O

FF

sta

te tr

ans

mis

sio

n, %

)

100 e-6500 e-61000e-6

OFF state (i layer) : np=Τa*J/q (Ta=Tp+Tn)

16

1013

1014

1015

1016

0

10

20

30

40

50

60

70

80

90

100

p-type doping area density (cm-2)

To

n (o

n st

ate

tra

nsim

issi

on,

%)

Design Criteria: Doping

(Np*xp)

ON state: 95% transmission requires an area dopingdensity of ~1x1014 /cm2

17

P

N

i

-V, Cathode

+V, Anode

DeviceCentre

Voltage drop due tosideways current flow

Lateral Current Flow Problem

Very little injection occurs near the

centre

The problem also occurs on the N side, but is less severe

18

Carrier Mobility in Ge

We require high mobility for uniform current distribution!Carrier-carrier scattering degrades mobility above ~ 1e15 /cm3

1014 1019Doping concentration (cm-3)

Mob

ility

(cm

2 /V

s)

19

Drive voltage issues – intrinsic layer thickness

0.1 0.5 1 5 1010

-3

10-2

10-1

100

101

Normalised intrinsic region length (2d/La)

Dro

p a

cro

ss in

trin

sic

reg

ion,

Vm

(V

)

2d < LaCarrier recombinationin doped region, J decreases

2d > LaLarge drop across i region,Increased heat dissipation

20

Simulations: ‘standard’ Ge

• 10μS lifetime, typical of ‘standard’ material

• 1018 /cm3 acceptors P side, 1014/cm2

• 5*1018 /cm3 donors N side, 5*1014/cm2

• Gaussian profile

• Axisymmetric

• 50μm wide metal ring electrodes

21

Hole concentration at 0.6 Volts

REMINDER!

We needed ~

Np=8.64*1016/cm3

Oh dear……..

22

Central Hole Area Concentration as a fuction of Voltage(10us Tn and Tp)

Voltage (V)

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8

Ho

le A

rea

Co

nce

ntr

atio

n

(1e

15/c

m^2

)

0.028

0.030

0.032

0.034

0.036

0.038

0.040

0.042

0.044

Hole Area Concentration

23

Central Transmission through PiN Diode(10us Tn and Tp)

Voltage (V)

0.0 0.2 0.4 0.6

Tra

nsm

issi

on

(%)

97.6

97.8

98.0

98.2

98.4

Po

wer

Dis

sip

ated

(W

)

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

TransmissionPower Dissipated

24

Transmission through Diode vs Horizontal Distance (10us Tn and Tp)

Horizontal Distance (um)

0 500 1000 1500 2000 2500

Tra

nsm

issi

on

(%

)

80

82

84

86

88

90

92

94

96

98

100

0.4 V0.5 V0.6 V

25

Avoid Copper!

Copper has an EXTREMELY high SRH cross section!

Copper diffuses very rapidly through germanium

Tiny concentrations, ~1010/cm3, severely shorten the lifetime.

Nuclear Detector is satisfactory, with lifetimes ~5mS

So we assume 2.5mS, allowing some processing loss

Avoidance of copper contamination is CRUCIAL

26

Hole concentration at 0.4 Volts

REMINDER!

We needed ~

Np=8.64*1016/cm3

Hopeful…….

27

Hole concentration at 0.5 Volts

REMINDER!

We needed ~

Np=8.64*1016/cm3

Interesting……….

28

Hole concentration at 0.6 Volts

REMINDER!

We needed ~

Np=8.64*1016/cm3

Whoopeeeeee…..

29

Voltage (V)

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7

Hol

e A

rea

Con

cent

ratio

n (1

e15/

cm^2

)

0

2

4

6

8

10

12

14

16

Hole Area Concentration

Central Hole Area Concentration as a function of Voltage(2.5 ms Tn and Tp)

30

Central Transmission through PiN Diode(2.5 ms Tn and Tp)

Voltage (V)

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7

Tra

nsm

issi

on (

%)

0

20

40

60

80

100

Pow

er D

issi

pate

d(W

)

0

2

4

6

8

10

TransmissionPower Dissipated

31

Transmission through Diode vs Horizontal Distance(2.5 ms Tn and Tp)

Horizontal Distance (um)

0 500 1000 1500 2000 2500

Tra

nsm

issi

on (

%)

0

5

10

15

20

25

30

35

40

0.4 V0.5 V0.6 V

32

BUT…

• There are issues…as always– Surface Recombination Velocity– Auger Recombination– Shockley-Read-Hall (SRH) Recombination– Photon transitions (indirect band gap)– Impact ionisation (high E.Field only)– Tunneling (not possible)– Many-body effects on diffusivity

33

What is Surface Recombination?

For GeS=1-100cm/s

Conduction band

Valence band

Carriers diffuse to the surfaceAnd ‘kill’ the effective lifetime.

34

What is SRH Recombination

• Trap could be due to presence of foreign atoms such as Cu, Ni, Au or a structural defect in crystal

• The ultra-pure material is essential free of this

• But the heavily doped regions are damaged….

35

Concentration Dependent SRH

• No literature available on CONSRH in Ge, pure guess values!!

• Quite Pessimistic values assumed, a factor of 1/10,000 i.e. from 10mS to 1 uS at highest concentration!

• The effect only operates over a very small volume of the device

36

What is Auger Recombination?• Three particle effect

• γ3=1.1e-31 cm6/s

• Some theoretical estimates lower (better)

• But not that well established

at densities relevant to us.

37

Conclusions

• Some device modeling still needed for optimization

• Process simulation now more important• Alloyed contacts may be used for fabrication?• Lifetime preservation in processing

absolutely crucial• The Auger parameter is a concern• But it does look possible