Open Archive Toulouse Archive Ouverte (OATAO)

(1)

To cite this document: Goiffon, Vincent and Virmontois, Cédric and Magnan, Pierre and Girard, Sylvain and Paillet, Philippe Analysis of Total Dose Induced Dark Current in

CMOS Image Sensors from Interface State and Trapped Charge Density Measurements.

(2010) In: IEEE Nuclear and Space Radiation Effects Conference (NSREC 2010), 19 July 2010 - 23 July 2010 (Denver, United States). (Unpublished)

OATAO is an open access repository that collects the work of Toulouse researchers and makes it freely available over the web where possible.

This is an author-deposited version published in: http://oatao.univ-toulouse.fr/

Eprints ID: 11394

Any correspondence concerning this service should be sent to the repository administrator: [email protected]

(2)

Vincent Goiffon

1

_{, Cédric Virmontois}

1

_{, Pierre Magnan}

1

_,

Sylvain Girard

2

_{and Philippe Paillet}

2

2010 IEEE Nuclear and Space Radiation Effects Conference 20 July 2010, Denver, Colorado

1

_{Institut Supérieur de l’Aéronautique et de l’Espace (ISAE)}

Image Sensor Research Team, Toulouse University, France

(3)

20 July 2010 NSREC2010 - V. Goiffon 2

Context / Motivations

Dark current increases in CIS are

often reported

but

Its variation with TID has not been explained in detail so far The annealing behavior is not fully understood

Radiation hard photodiode structures compatible with the majority of commercial deep sub-micrometer CMOS process is still missing

Purpose of this work:

Precise the role of TID induced trapped charge and interface states

in the generation of dark current in CIS

Dark current

increase is the main issue in

CMOS Image

Sensors

(

CIS

) exposed to ionizing radiations in:

Space / Military / Scientific / Medical applications

Dark current (

I_DARK)

Parasitic current when the sensor is not illuminated

Reduces the useful voltage swing and brings shot noise

(4)

Talk Outline

Experimental details

Results & Discussion

Nature of TID induced dark current

Role of trapped charges

(5)

20 July 2010 NSREC2010 - V. Goiffon

Photodiode environment

in pixel arrays

As can be seen on the 3x3 pixel array illustration

photodiodes (PN junctions) are

surrounded by Shallow

Trench Isolations

(

STI

) in CMOS image sensors

The

depleted region ends

at the bottom

STI interface

(6)

Photodiode environment

in pixel arrays

As can be seen on the 3x3 pixel array illustration

photodiodes (PN junctions) are

surrounded by Shallow

Trench Isolations

(

STI

) in CMOS image sensors

The

depleted region ends

at the bottom

STI interface

3x3 pixel array

Photodiode STI Photodiode

(7)

Photodiode environment

in pixel arrays

As can be seen on the 3x3 pixel array illustration

photodiodes (PN junctions) are

surrounded by Shallow

Trench Isolations

(

STI

) in CMOS image sensors

The

depleted region ends

at the bottom

STI interface

3x3 pixel array

Photodiode STI Photodiode

2 adjacent pixel cross section

Space charge region

(8)

Test structures

The following

test structures

have been used in this

work to

study the dark current

:

(9)

Test structures

CIS photodiode

The following

test structures

have been used in this

work to

study the dark current

:

CIS photodiode:

(10)

Test structures

FOXFET

CIS photodiode

The following

test structures

have been used in this

work to

study the dark current

:

CIS photodiode:

To measure the dark current

FOXFET:

To estimate the defect densities around the photodiode

Drain/Source = CIS photodiode

(11)

Test chip & Irradiation

Test structures : Photodiodes

Arranged in arrays of 300 diodes in parallel Junction size: 2 µm x 5 µm

Test structures : FOXFETs

To extract the defect densities at the photodiode vicinity

Several W/L ratios : from 300/0.7 to 300/100

Manufactured

On the same die

By using a commercial 0.18µm CMOS process dedicated to imagers

Irradiated at CEA DIF by

10 keV X-rays

Up to 1 Mrad(SiO₂) At ~0.5 krad/s

(12)

Talk Outline

Experimental details

Results & Discussion

Nature of TID induced dark current

Role of trapped charges

(13)

Irradiated FOXFET I-V characteristics

I

_DS

corresponds to the dark current for zero FOXFET gate

voltage:

I

_DS

≈

≈ I

≈

_DARK

for

V

_GS

= 0V

-20 0 20 40 60 10-14 10-12 10-10 10-8 10-6 10-4 10-2

Gate to source voltage (V)

D ra in c u rr e n t (A ) 0 krad 3 krad 10 krad 30 krad 100 krad 300 krad 1 Mrad V_DS = 3.3V W/L = 300/0.7

(14)

Irradiated FOXFET I-V characteristics

I

_DS

corresponds to the dark current for zero FOXFET gate

voltage:

I

_DS

≈

≈ I

≈

_DARK

for

V

_GS

= 0V

-20 0 20 40 60 10-14 10-12 10-10 10-8 10-6 10-4 10-2

No sign of gate induced - tunneling effect

- electric field enhancement

No subthreshold conduction up to 100 krad

(15)

Irradiated FOXFET I-V characteristics

I

_DS

corresponds to the dark current for zero FOXFET gate

voltage:

I

_DS

≈

≈ I

≈

_DARK

for

V

_GS

= 0V

-20 0 20 40 60 10-14 10-12 10-10 10-8 10-6 10-4 10-2

Above 100 krad,

subthreshold current dominates

(16)

Irradiated FOXFET I-V characteristics

I

_DS

corresponds to the dark current for zero FOXFET gate

voltage:

I

_DS

≈

≈ I

≈

_DARK

for

V

_GS

= 0V

-20 0 20 40 60 10-14 10-12 10-10 10-8 10-6 10-4 10-2

Above 100 krad,

subthreshold current dominates

(inter device leakage)

100 101 102 103 1010 1011 1012 TID (krad) ∆N o t (c m -2 ) a n d ∆D it ( c m -2 ⋅ e V -1 ) ∆D it ∆N_ot Extracted defect densities

(17)

Irradiated photodiode dark current

0 0.5 1 1.5 2 2.5 3 10-15 10-14 10-13 10-12 10-11 10-10 Reverse voltage (V) R e v e rs e c u rr e n t (A ) fresh 3 krad 10 krad 30 krad 100 krad 300 krad 1 Mrad

Slow dark current (I

_DARK

) increase with reverse voltage V

_R

Negligible diffusion contribution

No sign of tunneling effect / electric field enhancement

Negligible subthreshold conduction (because here W/L is more than 10 time smaller than the FOXFET one)

(18)

Dark current source location

Previous work showed that radiation induced dark current

was directly

proportional

to the photodiode junction

perimeter P

_J

(also confirmed in this work)

The dark current SRH source is located at the STI/depletion

region interface

(19)

Dark current Arrhenius plot

The dark current activation energy is close to Eg/2

Due to Shockley-Read-Hall generation (even @ 1 Mrad) Dark current simplified expression:

-50 -48 -46 -44 -42 -40 -38 10-15 10-14 10-13 10-12 10-11 10-10 -q/kT (eV-1) I da rk ( A ) 30 krad 100 krad 1 Mrad

it

dep

dark

K

W

D

I

=

×

Slope

≈

≈ Eg/2

Depletion width at the STI interface

(20)

Talk Outline

Experimental details

Results & Discussion

Nature of TID induced dark current

Role of trapped charges

(21)

Comparison dark current /

interface state density

First, if only the effect of

interface states is considered

:

The dark current increase

∆

∆I

_DARK

should be proportional

to the

interface density increase

∆

∆D

∆

_it

)

(

)

(

TID

K

P

W

D

TID

I

_dark

=

×

_J

×

_dep

×

∆

_it

(22)

Comparison dark current /

interface state density

If only the effect of

interface states is considered

:

The dark current increase

∆

∆I

_DARK

should be proportional

to the

interface density increase

∆

∆D

∆

_it

1011 1012 10-13

10-12 10-11

Interface state density increase (cm-2/eV)

D a rk c u rr e n t in c re a s e ( A ) Delta I dark y=a⋅x

No

proportionality

Interface states can not explain by themselves the

observed dark current increase

)

(

)

(

TID

K

W

D

TID

I

_dark

=

×

_dep

×

∆

_it

(23)

Influence of trapped charges

on dark current (1)

The trapped charge can

change the electrostatic equilibrium

at

the STI interface

Depletion width (W_DEP)variation with TID

If the dependence of

W

_DEP

on TID

is taken into account:

)

(

)

(

)

(

TID

K

W

TID

D

TID

I

_dark

=

×

_dep

×

_it

[

_it _it

]

ot dep it ot dark

N

D

K

W

N

D

I

(

∆

,

∆

)

=

×

(

∆

)

×

₀

+

∆

(24)

Influence of trapped charges

on dark current (1)

The trapped charge can

change the electrostatic equilibrium

at

the STI interface

If the dependence of

W

_DEP

on TID

is taken into account:

Function to plot to observe the depletion width evolution:

)

(

)

(

)

(

TID

K

W

TID

D

TID

I

_dark

=

×

_dep

×

_it it it it ot dark ot dep

D

N

I

K

N

W

∆

+

∆

×

=

∆

0

)

,

(

1 )

(

i th eff J

v

k

Tn

qP

K

=

σ

π

K

=

K

'

×

σ

_eff

Known constant

[

_it _it

]

N

D

K

W

N

D

I

(

∆

,

∆

)

=

×

(

∆

)

×

₀

+

∆

(25)

Influence of trapped charges

on dark current (1)

The trapped charge can

change the electrostatic equilibrium

at

the STI interface

If the dependence of

W

_DEP

on TID

is taken into account:

Function to plot to observe the depletion width evolution:

)

(

)

(

)

(

TID

K

W

TID

D

TID

I

_dark

=

×

_dep

×

_it

Unknown parameters

i th eff J

v

k

Tn

qP

K

=

σ

π

K

=

K

'

×

σ

_eff

Known constant

it it it ot dark ot dep

D

N

I

K

N

W

∆

+

∆

×

=

∆

0

)

,

(

1 )

(

[

_it _it

]

N

D

K

W

N

D

I

(

∆

,

∆

)

=

×

(

∆

)

×

₀

+

∆

(26)

Influence of trapped charges

on dark current (2)

1011 1012 0 0.5 1 1.5 2 2.5

Trapped charge density increase ∆N_ot (cm-2)

In te rf a c e d e p le ti o n w id th ( µ m ) it it dark ot dep

D

I

K

N

W

∆

+

×

=

∆

0

1 )

(

2 16

cm

10

2 ×

− −

≈

eff

σ

The unknown parameters are chosen to yield realistic depletion

width values (they have a very small influence on the curve shape)

The

depletion width extension

with TID

explains

the observed dark

current evolution

eV

/

cm

10

10 2 0 −

≈

it

D

W

_DEP

increases

much with N

_ot

(and so with TID)

(27)

TCAD simulations

Qualitative Sentaurus TCAD simulations based on SIMS

measurements

2 photodiodes separated by 5 µm of STI (as in the studied photodiodes)

Simulation purpose:

Confirm the trapped charge influence on the depletion width at the STI interface

0 krad

N

P

Space charge region Space charge region Dark current sources

(28)

TCAD simulation: N

_ot

= 10

11

_cm

-2

0 krad

N

_ot

= 10

11 _cm

-2

~ 7 krad

W

₁

W

₂

Depletion region

extension at the STI

interface

(W2 > W1)

(29)

TCAD simulation: Not = 3x10

11

_cm

-2

N

_ot

= 3x10

11 _cm

-2

~ 25 krad

0 krad

W

₁

W

₂

Depletion region

extension at the STI

interface

(W2 ≈

≈

≈ 3xW1)

≈

(30)

TCAD Simulation: Not = 10

12

_cm

-2

N

_ot

= 10

12 cm

-2

~ 160 krad

0 krad

The STI interface is

completely depleted

:

Enhances strongly the generation rate

Can lead to inter device leakage if N_ot is large enough Adjacent photodiode are biased at different potentials

(31)

20 July 2010 30

Comparison TCAD/measurements

1011 1012 0 0.5 1 1.5 2 2.5 ∆N_ot (cm-2) W d e p ( µ m ) Experimental data TCAD

The same effect can be

observed on both curves

Discrepancies are acceptable considering the simplifying assumption used

(32)

Talk Outline

Experimental details

Results & Discussion

Nature of TID induced dark current

Role of trapped charges

(33)

Summary

After X-ray exposure (up to 1 Mrad(SiO

₂

)),

the CIS photodiode

dark current

Increased

quickly with TID

Due to

Shockley-Read-Hall

generation

at the

STI

/depleted region

interface

The

role

of TID induced

defects

was

clarified

:

Interface states

contribute directly to the dark current by

increasing the number of

generation centers

in the depleted

region

Trapped charges

enhance the dark current by

extending the

depleted region

at the STI interface (confirmed by TCAD

simulations)

At

high TID

(above ~300krad), the

STI interface is completely

depleted, and

inter-device leakage

can become an issue

100 101 102 103 10-14 10-13 10-12 10-11 10-10 10-9 TID (krad) D a rk c u rr e n t in c re a s e ( A / µ m )