OXIDE AND INTERFACE TRAPPED CHARGES, OXIDE THICKNESS

mewlingfawnΗμιαγωγοί

2 Νοε 2013 (πριν από 3 χρόνια και 11 μήνες)

716 εμφανίσεις

OXIDE AND INTERFACE TRAPPED

CHARGES, OXIDE THICKNESS

Yameng

Bao

Yameng.bao@aalto.fi


Electron Physics Group

1

Outline


1.Introduction


2. Fixed
Oxide Trapped, and Mobile Oxide
Charge


3. Interface
Trapped
Charge


4. Oxide Thickness


5. Strengths
and Weaknesses

2

1.Introduction


Capacitance
-
voltage
and oxide
thickness measurements
must be
more
carefully interpreted
for thin,
leaky
oxides


Charges and defects in the oxide


Variable Capacitance


Insulation and passivation


High dielectric constant


Low leakage current and low tunnel
current
-
lower power waste lower
temperature of device


Focus
on SiO
2
-
Si system


3


(
1). Interface Trapped Charge(
Q
it

,N
it

,
D
it
)


(
2). Fixed Oxide
Charge(
Q
f

,
N
f

)


(3) Oxide Trapped Charge (
Q
ot

, N
ot

)


(4) Mobile Oxide Charge (Q
m
,N
m
)

4

Oxide Charges


(1) Interface Trapped
Charge(
Q
it
)


Due
to structural defects,
oxidation
-
induced
defects, metal
impurities, or other
defects caused
by radiation or similar
bond
breaking
processes


Unlike fixed charge or trapped
charge interface
trapped
charge is in electrical communication with the underlying
silicon


Could be neutralized by low T H
2

or forming Gas(N
2
&H
2
)

5

Oxide Charges


(2) Fixed Oxide Charge(
Q
f
)(near the interface)


Coming from oxidation process


Usually measure after Annealing to eliminate
the effect of the interface trapped charge


It depends on final oxidation temperature


Always present in any cases

6

Oxide Charges


(3) Oxide Trapped Charge(
Q
ot
)


Due to the ionizing radiation, avalanche injection and
so on


Sometimes could be annealed by Low
-
T treatment

but the neutral traps still remain


(4) Mobile Oxide Charge(
Q
m
)


Caused by Na
+
, Li
+
, K
+

and so on


Chlorine atom may reduce this charge

7

2. Fixed Oxide Trapped, and Mobile Oxide Charge

(1) Capacitance
-
voltage Curve



8

Q
G
is gate charge density

V
G

is
gate voltage

Q
G
=
-
(Q
s +
Q
it
)

Q
s
is
semiconductor charge density

Q
it

is
interface charge density

V
G

= V
FB

+ V
ox

+
φ
s

V
FB
is

flatband

voltage

V
ox
is

oxide voltage


φ
s
is surface potential

Q

S
= Q

p
+Q

b
+ Q

n

Q
p

is
hole charge
density,

Q
b

is space
-
charge
region bulk charge
density


Q
n

is electron
charge
density


2. Fixed Oxide Trapped, and Mobile Oxide Charge

(1) Capacitance
-
voltage Curve



9

V < 0

V
>
0

V >> 0

Accumulation

Depletion

Strong inversion

For P type substrate

2. Fixed Oxide Trapped, and Mobile Oxide Charge

(1) Capacitance
-
voltage Curve



10

For negative gate
voltages

Accumulation
:

1) Big negative voltage

Q
p

dominates .
C
p

is short circuit

2. Fixed Oxide Trapped, and Mobile Oxide Charge

(1) Capacitance
-
voltage Curve



11

Depletion

Small negative voltage and small
positive voltage

Qb
=
-
qN
A
W


In week
inversion

C
n

begin to appear

Strong
inversion

C
n

domains


a) If the inversion charge could
follow the
HF
-
AC
,


C=C
ox


b) if the inversion could not
follow,


C=
Cox+Cb

2. Fixed Oxide Trapped, and Mobile Oxide Charge

(1) Capacitance
-
voltage Curve



12

When the dc bias voltage is changed rapidly with
insufficient time for inversion
charge generation
, the
deep
-
depletion curve results. Its high
-

or low
-
frequency
semiconductor capacitance is
C
dd

Effect of sweep direction and sweep rate on
the
hf

MOS
-
C capacitance on p
-
substrate
,

2. Fixed Oxide Trapped, and Mobile Oxide Charge

(2)
Flatband

Voltage



13

The
flatband

voltage is determined by the metal
-
semiconductor work function
difference
φ
MS

and the various oxide charges through the relation

Determine the V
FB

Metal
-
S work
function
different

Fixed
charge

Interface
trapped
charge

Charges in metal

Charges in oxide

2. Fixed Oxide Trapped, and Mobile Oxide Charge

(3) Capacitance Measurement



14

High
Frequency: High
-
frequency C


V curves are typically measured at 10 kHz


1
MHz.

Using
a phase
sensitive detector, one can determine the conductance G or
the capacitance
C, knowing
R and ω = 2πf

2. Fixed Oxide Trapped, and Mobile Oxide Charge

(3) Capacitance Measurement



15

Low
Frequency:
Current
-
Voltage

Low Frequency: Current
-
Voltage: The low
-
frequency capacitance of an MOS
-
C
is usually
not
obtained by measuring the capacitance, but rather by measuring a
current or
a charge,
because capacitance measurements at low frequencies are very noisy
.

Low F

High F

2. Fixed Oxide Trapped, and Mobile Oxide Charge

(3) Capacitance Measurement



16

Low
Frequency:
Current
-
Voltage and Charge
-
Voltage

Q
-
V is more suitable for
MOS measurement

2. Fixed Oxide Trapped, and Mobile Oxide Charge

(4) Fixed Charge



17

a)
The
fixed charge is determined by comparing the
flatband

voltage shift of an
experimental C


V curve with a theoretical curve and measure the voltage
shift

To determine
Q
f

,one
should eliminate or at least reduce the effects of all other oxide
charges and
reduce the
interface trapped charge to as low a value as possible.
Q
it

is
reduced by
annealing in
a
forming gas.

b)
Second method using differing
t
ox

Plot V
FB

versus
t
ox

with slope
Q
f

/
K
ox

ε
0

and intercept
φ
MS

. This
method, requires
MOS capacitors with differing
t
ox

.However
, it is more accurate because it is
independent of
φ
MS

.
K
ox

is semiconductor dielectric constant

2. Fixed Oxide Trapped, and Mobile Oxide Charge

(5) Work function difference




18

φ
MS

depends on oxidation temperature, wafer orientation, interface trap density,
and
on the
low temperature
D
it

anneal

(6) Oxide Trapped Charge(
Q
ot
)





The distribution of
Q
ot

must be known
for proper
interpretation of C

V curves. Trapped
charge distributions are measured
most commonly
by the
etch
-
off
and the
photo I

V
methods


A determination of the charge distribution in the oxide is tedious and therefore not

routinely done. In the absence of such information, the
Vfb

shift
due to
charge
injection
is generally interpreted by
assuming the charge is at the
oxide
-
semiconductor
interface

using the expression

2. Fixed Oxide Trapped, and Mobile Oxide Charge

(5) Mobile Charge




19

Mobile charge in
SiO

2
is due primarily to the ionic impurities
Na
+
, Li
+
, K
+
,
and
perhaps H
+
.
Sodium is the dominant contaminant.


Bias
-
Temperature
Stress( BTS): Measured at 250C, under gate bias, measure CV
then cool down to 25C, then measure CV, the
Q
m

is
determined by
V
fb

shift
.


Triangular voltage sweep
(TVS)
method:

C
lf

and
C
hf

measured at T=250C, The
Q
m

is determined
from the area between the two curves

3. Interface trapped charge

(1) Low frequency(Quasi
-
static) methods




20

Effect of D

it
on MOS
-
C capacitance
-
voltage curves. (a) Theoretical high
-
frequency
,(
b)
theoretical low
-
frequency and (c) experimental low
-
frequency curves. Gate voltage stress
generated
interface traps

This stretch
-
out is not the
result of
interface traps contributing excess
capacitance, but rather it is the result of
the C

V curve
stretch
-
out along the
gate voltage axis

Interface traps do respond to the
probe
frequency at
LF
,
and the curve distorts
because the
interface traps
contribute
interface trap capacitance
C
it

and
the
curve stretches out along the
voltage axis

HF

LF

Experimental
-
LF

3. Interface trapped charge

(1) Low frequency(Quasi
-
static) methods




21

Δ
C/C
ox

=
C
lf

/
C
ox


C
hf

/
C
ox

3. Interface trapped charge

(2) Conductance Method



22

The conductance is measured as
a function
of frequency and plotted as G

P
/ω versus ω
.



G
P

/ω has a maximum at ω
=1/τ
it and at that maximum
D
it
=2G
P
/

.
we
find ω ≈ 2/τ
it

and D
it

=2.5G
P
/


at the maximum. Hence we determine D
it

from the maximum G
P


and determine
τ
it

from ω at the peak conductance location on the ω
-
axis.

One
of
the most
sensitive methods to determine D

it
Interface
trap densities of

10
9

cm
− 2
eV
− 1
and lower
can be measured.

3. Interface trapped charge

(3)High Frequency Method



23

Terman

Method
: In HF CV, interface
traps do not respond to the ac probe frequency,
they do respond to
the slowly
varying dc gate voltage and cause the
hf

C

V curve to
stretch out along the
gate voltage
axis as interface trap occupancy changes with gate
bias

Δ
V
G

= V
G


V
G
(ideal) is the voltage shift of the
experimental from the ideal
curve, and
V
G

the
experimental gate voltage

The method is generally considered to be
useful for measuring interface trap densities

of 10
10

cm

2

eV
− 1
and
above

3. Interface trapped charge

(3)High Frequency Method



24

Gray
-
Brown and
Jenq

Method
:, the
C
HF

measured as a function of
T. Reducing
the
T
causes
the Fermi
level to shift towards the majority carrier band edge
and the interface trap
time constant
τ
it

increases at lower
T.
Hence interface
traps near the band
edges should
not respond to typical ac probe frequencies
at low
T
whereas at
room temperature
they do respond. This method should
extend the range of interface
traps measurements
to D

it
near the majority
carrier band
edge

Compared with DLTS?

3. Interface trapped charge

(4)Other Methods



25

1.Charge
Pumping

2. MOSFET
Sub
-
threshold
Current method


3. DC
-
IV
method

4.
deep
-
level transient
spectroscopy(DLTS)


3. DC
-
IV
method


5. charge
-
coupled
devices (CCD)


6. electron spin
resonance (
ESR)

3. Oxide thickness

(1)Capacitance
-
Voltage(equivalent electrical thickness)



26


C

V , I

V ,
ellipsometry
, transmission electron
microscopy(TEM
), X
-
ray photoelectron
spectroscopy (XPS), medium energy ion scattering
spectrometry
(MEIS), nuclear reaction
analysis (NRA), Rutherford backscattering (RBS),
elastic backscattering
spectrometry
(EBS), secondary ion mass spectrometry (SIMS),
grazing incidence
X
-
ray
reflectometry

(GIXRR), and neutron
reflectometry

3. Oxide thickness

(2)Current
-
Voltage



27

The current flowing
through an
insulator is either Fowler
-
Nordheim

(FN)
or direct tunnel current

(a) V
ox

<

B

(direct tunneling)

(b) V
ox

>

B

Fowler
-
Nordheim

tunneling

3. Oxide thickness

(3)Other methods



28

Ellipsometry
:
Suitable
for oxides into the 1

2 nm
regime. Variable
angle
,


spectroscopic
ellipsometry

is especially suited for oxide thickness

measurements

Transmission Electron Microscopy is very precise and usable

to very thin oxides, but sample preparation is tedious

X
-
ray Photoelectron Spectroscopy

4. Strength and Weakness

(1)Mobile Oxide Charge


29


Bias
temperature stress method



Requiring the
measurement of a C

V at different
Ts


Total mobile charge density will be measured, No separation


Triangular
voltage sweep method


Could differentiate different mobile charges, high sensitivity, fast


Increasing oxide leakage current for thin film


(2) Interface tapped charge(conductance and low
frequency method



Conductance method



high sensitivity, majority carrier capture cross sections


Limited surface potential range


Quasi
-
static method(I
-
V/Q
-
V)


Easy to measure, large surface potential range


I
-
V the requirement for I
-
V, current is low


For I
-
V and Q
-
V leakage current could be a big problem

4. Strength and Weakness

(3)Oxide Thickness


30



MOS C

V measurements
are most
common
.

Leakage current make the result much difficult



I
-
V used for thickness extraction



Ellipsometry is mostly used for thickness, very sensitive to thin
oxides



XPS suitable for very thin oxide

5. Questions?

31


All the charges seems affect each other during
the measurement.


For thin oxide, the tunnel current or leakage
current will effect the result.


Real measurement is always not as simple as
description in the book


O(∩_∩)O