Comparison of reactive and ceramic AZO and

ITO from dual rotatable magnetrons

V.Bellido-Gonzalez, Dermot Monaghan, Robert

Brown, Alex Azzopardi, Gencoa, Liverpool UK

• Overview of basic magnetic designs for rotatable

magnetrons for DC and AC sputtering

• Anode importance in rotatable magnetrons and effect

on substrate heating and plasma interaction

• Magnetic options for rotatable magnetrons with positive

guiding of plasma electrons

• Case study: electrical and optical properties of reactive

and non-reactive AZO layers formed with different

rotatable magnetic geometries

• Case study: electrical and optical properties of reactive

ITO layers formed with different rotatable magnetic

geometries

• Conclusions

NREL

Structure of presentation

Whilst for a planar magnetron discharge and anode can be used to confine the plasma,

typically for rotatable magnetron no anode is close-by

Accurate positioning of the magnetic field to ensure erosion to the end of the target

No reaction product on the surface – cleans itself

Absence of anode can be seen in a plasma spread away from the target area

DC AC

Anode’s in magnetron plasma’s

• A plasma is effectively an electric circuit with the

target a negatively biased cathode and the

chamber or separate mean providing the anode

for the circuit return.

• Anodes are commonly earthed, although a

positive charge is also possible.

• Whilst the plasma confinement in the near target

area is governed by the magnetic field, the

plasma spread away from the target is primarily

an anode interaction effect.

For single magnetrons or for DC discharges anodes needs to be different to the AC pair case,

hence a magnetically linked auxiliary anode is used

Effect of active magnetically guided anode on the sputter target voltage for a GRS75 – 75mm

OD dual DC powered arrangement & Al target material

Magnetic design for a double magnetron used in industry currently

The above is the conventional magnetic arrangement for rotatables

used by all manufacturers.

AC power mode and electron movement

e-

- +

• AC provides excellent arc suppression – perfect

for reactive oxides and TCO’s

• But increases the plasma at the substrate –

potentially damaging some layer structures and

substrates!

Industry standard magnetics with AC power mode and electron movement

70 mm

100 mm

120 mm

AC current “leaks”

Lower impedance ‘linked’ magnetics as a solution for better plasma control away

from the target area

e-

- +

e-

70 mm

100 mm 120 mm

AC current “leaks”

Plasma to substrate interaction by assymetric magnetics and tilting

Gencoa patent

NREL

Magnetic field – Gencoa DLIM bars – no AC leakage DLIM stands for Double Low

Impedance Magnetics

70 mm

100 mm

120 mm

AC current “channelled”

Plasma control by Double Low Impedance Magnetics - DLIM

Adjustment of angle relative to substrate position

DC

AC

100

110

120

130

140

150

160

0 2 4 6 8 10 12

Tem

pera

ture

(D

eg

C)

probe position

Temperature on probes across (every 25 mm)

T acrossDLIM

T acrossBOC

Comparison of substrate temperature in-front of a double AC rotatable magnetron

DLIM has a 20C̊ lower temperature for same conditions

CASE STUDY use of DLIM and standard magnetics to compare AZO layers from

ceramic targets with AZO layers deposited reactively

Ceramic AZO on rotatable – Good Concept, but!

Some areas to improve

• Moderately expensive ceramic targets and bonding

• Micro-arcing – leads to variable & non-optimum

product quality – adds power modes and material costs

• Long target burn in before stable film properties can be

> 24hrs

• Possible plasma damage of growing film - increasing

resistivity,

• Limitation of composition and crystal structure – good

and bad

* SCI – Sputtering Components Inc

Hard arc count during pulsed-DC sputtering of

ceramic AZO (ENI DCG + Sparc-le V)

0

100

200

300

400

500

600

3 4 5 6 7 8 9 10 11 12 13

Power (kW)

Hard

arc

co

un

t

Ceramic AZO layer properties – variation of properties with process parameters

Problematic but presents an opportunity to improve

Variation of AZO properties for DLIM dual rotatable cathode with pulsed DC power

Variation of sheet resistance and resistivity with O2

Variation of AZO properties for DLIM dual rotatable cathode with pulsed DC power

Variation of sheet resistance and resistivity with T

Ts vs. Sheet resitance (ceramic AZO, 10 kW p-DC 100kHz,

2us, 500nm) DLIM

10

14

18

22

26

30

0 50 100 150 200 250

Ts (deg. C)

Sh

ee

t re

sist

an

ce

(Oh

m/

sq)

8.4e-4 9.2e-4

7e-4

* Szyszka et al

Controlled reactive sputtering is x 3 the rate in production than ceramic AZO

Price will be < 50% current ceramic based costs

Different sensor control modes possible for reactive AZO via feedback controller

Penning-PEM

Target V

Lambda

Process-

PEM

O2 gas

Basic process parameters for all depositions

target rotation speed: 5 rpm

Substrate static

T/S: 95 mm

Temp: Room Temp.

Dep. Time: 10 mins

ZnAl: 152 mm diam x 475 mm L

AC-MF: 5.3 kW (Huettinger)

Ar press.: 3E-03 mbar

0

50

100

150

200

250

300

350

400

450

500

Deposition conditions

deposited thic

kness

Thickness (nm) for 2.5 min deposition at 5.3 kW AC BOC reactive (RT)

DLIM reactive (RT)

DLIM ceramic (RT)

DLIM ceramic (150 deg C)

Comparison of deposition rates for reactive and ceramic and DLIM/BOC magnetics

Under conditions for optimum layer properties

Comparison of electrical properties for ceramic AZO, standard (BOC) & DLIM

without substrate heating and AC power

1.00E-04

1.00E-03

1.00E-02

1.00E-01

1.00E+00

0 2 4 6 8 10 12

resis

tivity,

Ohm

-cm

Sample position

Resisitivity DLIM ceramic AZO target at RT and 150 deg C (samples every 25 mm)

resistivity AZO DLIM (RT)

resistivity AZO DLIM (150deg C)

Comparison of ceramic AZO in-front of a double AC rotatable magnetron

Comparing 2 different substrate temperatures

1.00E-04

1.00E-03

1.00E-02

1.00E-01

1.00E+00

0 2 4 6 8 10 12

resis

tivity,

Ohm

-cm

Sample position

Resisitivity DLIM (reactive and ceramic AZO) at room temperature (static coating every 25 mm under double

magnetron cathodes)

resistivity AZO DLIM (RT)

resistivity reactive DLIM

Comparison of electrical properties for ceramic and DLIM for optimized layers

without substrate heating and with AC power

1.00E-04

1.00E-03

1.00E-02

1.00E-01

1.00E+00

0 2 4 6 8 10 12

resis

tivity,

Ohm

-cm

Sample position

Resisitivity BOC & DLIM at room temperature (every 25 mm)

resistivity BOC

resistivity DLIM

Comparison of reactive AZO in-front of a double AC rotatable magnetron

Comparing the 2 different magnetic designs

AZ+O2 film properties at Room Temperature and 150ºC with similar properties

R09 (at RT) and R17(at 150 deg C)

0

500

1000

1500

2000

2500

3000

0 2 4 6 8 10 12

sample (every 25 mm)

Thic

kness, nm

1.00E-04

1.00E-03

1.00E-02

1.00E-01

1.00E+00

Oh

m-c

m

t (at 150ºC)

t (at RT)

r (at 150ºC)

r (at RT)Log scale

Room temperature films have better optical density with DLIM magnetics

Optical Density at 550nm & Resistivity for

R09 (at RT) and R17(at 150 deg C)

0

0.02

0.04

0.06

0.08

0.1

0.12

0.14

0.16

0.18

0.2

0 2 4 6 8 10 12

sample (every 25 mm)

Optical D

ensity a

t 550nm

1.00E-05

1.00E-04

1.00E-03

1.00E-02

Oh

m-c

m

od (at 150ºC)

od (at RT)

r (at 150ºC)

r (at RT)

Log scale

With reactive processes transmission can tuned over a wide range and tuned

with electrical properties for different applications

Coating thickness for both is 1.8µm

3Ω/sq

0

20

40

60

80

100

120

325 525 725 925

Tra

nsm

issio

n

wavelength, nm

T(%) R09 (at RT) and R17 (at 150ºC)

T(%) R09 T(%) R17

AZ+O2 transmittance in the visible spectrum good low temp transparency

Coating thickness ~ 2.4 µm

Resistivity change with target voltage and

substrate temp. (see reference 2)

0123456789

1011

0 100 200 300 400 500

Substrate Temperature C

Re

sis

tivity (

x 1

0-4

Oh

m.c

m)

-400 V

-250 V

-110 V

For ITO & other sputtered TCO’s low damage on hot surfaces provide best quality

Crystal structure and doping is critical for all TCO’s

Jumbo Glass, TCO film property tuning using ‘Speedflo’ reactive sputtering

controller with a dual rotatable magnetron

InSn+O2 using Speedflo control for reactive

production of ITO

0

20

40

60

80

100

120

140

160

38 40 42 44 46 48 50 52

O2 Set-point (%)

Sh

ee

t re

sis

tan

ce

(o

hm

s)

0

10

20

30

40

50

60

70

80

90

Tra

nsm

issio

n (

%)

Sheet resistanceTransmission

Development

Optimised

0.00E+00

2.00E-04

4.00E-04

6.00E-04

8.00E-04

1.00E-03

1.20E-03

0 1 2 3 4 5 6 7 8 9 10 11

Resis

tivity

Ohm

/cm

Sample position every 25mm

DLIM RESISTIVITY

BOC RESITIVITY

Reactive ITO comparison of conventional magnetic design and DLIM with

AC power mode

1.00E-04

1.00E-03

1.00E-02

0 1 2 3 4 5 6 7 8 9 10 11

Res

isti

vit

y O

hm

/cm

DLIM RESISTIVITY

BOC RESISTIVITY

Reactive ITO comparison of conventional magnetic design and DLIM with

AC power mode at 150̊ C

3.00E-04

4.00E-04

5.00E-04

6.00E-04

7.00E-04

8.00E-04

9.00E-04

0 1 2 3 4 5 6 7 8 9 10 11

Resis

tivit

y,

Oh

m.c

m

DLIM resistivity 80 degrees C DLIM resistivity 120 degrees C

DLIM resistivity 150 degrees C DLIM resistivity 180 degrees C

Reactive ITO with DLIM, AC power mode and varying substrate

temperature

Parameters for ITO Ceramic Tests

• For ITO from ceramic targets several process parameters affect the electrical conductivity of the ITO film:

• Standard strength magnetics – 520 Gauss over target surface, average target voltage 370 Volts

• Deposition power 2.5kW per target – 2 targets – total 5 kW • Target to substrate separation 10cm & 15cm • Deposition time 30 sec – static substrates • Average ITO film thickness 130-140nm • Ar & O2 gas flow – introduced at central and / or outer gas bars • Central magnetically guided anode varied from earthed, floating

and +15 V. • Angle of the magnetic to the anode varied from 0, 30, 60 & 90 ̊ • Substrate temperature – RT 20̊C, 180̊ C.

Ceramic ITO with DLIM – TCO (active anode, DC power mode and room

temperature substrate (no heating)

1.0E-04

1.0E-03

1.0E-02

Res

isti

vit

y Ω

.cm

Ceramic ITO rotatable, Active Anode +15V, Room Temp. Substrate, no O2 gas

DC

DC Pulsed

1.0E-04

1.0E-03

1.0E-02

Res

isti

vit

y Ω

.cm

Ceramic ITO rotatable, Active Anode +15V & 0V, Room Temp. Substrate, no O2 gas, pure DC target power

Anode Floating

Anode Grounded

Anode +15V

Ceramic ITO with DLIM – TCO (active anode +15v, 0, floating, DC power and

room temperature substrate (no heating)

1.0E-04

1.0E-03

1.0E-02

Res

isti

vit

y Ω

.cm

Ceramic ITO rotatable, Active Anode +15V, Room Temp. Substrate, no O2 gas, Pulsed DC Power Variation

50kHz pulsed DC

100 kHz pulsed DC

Pure DC Power

Ceramic ITO with DLIM – TCO (active anode +15v, varying DC modes and

room temperature substrate (no heating)

1.0E-04

1.0E-03

1.0E-02

Res

isti

vit

y Ω

.cm

Ceramic ITO rotatable, Active Anode +15V, Room Temp. Substrate, no O2 gas, Pulsed DC Power Variation

zero O2 added, 50 kHz pulse

2.5% O2 added

2% O2 added Pure DC Power

2% O2 added, 100 kHz pulsing

Ceramic ITO with DLIM – TCO (active anode +15v, DC power modes and

varying gas mixtures (no heating)

1.0E-04

1.0E-03

1.0E-02

Res

isti

vit

y Ω

.cm

Ceramic ITO rotatable, Active Anode, Room Temp. Substrate, Variable Gas Pulsing, Pulsed DC 50kHz

10 sccm constant gas, anode+15V

2.5 - 10 sccm pulsed gas profile, anode +15V

1.25 - 5 sccm gas pulsing, anode +15V

1.25 - 5 sccm gas pulsing, anode floating

Ceramic ITO with DLIM – TCO (active anode +15v & FL, DC 50kHz pulse

and varying gas modes (no heating)

1.0E-04

1.0E-03

1.0E-02

Res

isti

vit

y Ω

.cm

Ceramic ITO rotatable, Active Anode Variation, Room Temp. Substrate, Variable Tilt Angle, Pulsed DC 50kHz, 2-3% O2 constant flow

Zero mag bar tilt, anode earthed

zero mag bar tilt, anode +15V

30 deg mag bar tilt, anode earthed

Ceramic ITO with DLIM – TCO (active anode +15v & 0v, DC 50kHz pulse

and varying magnetics tilt angle (no heating)

Conclusions ITO Ceramic so far

• For ITO from ceramic targets several process parameters affect the

electrical conductivity of the ITO film:

• Average resistivities of 5 x 10-3 to 4.6 x 10-4 Ω.cm can be

achieved on room temperature substrates depending upon

process parameters

• Power mode – 50kHz DC optimum compared to pure DC or

100kHz

• Gas Injection position – more tests needed for conclusions

• Anode Bias - +15V best but earthed also good

• O2 gas flow – optimum needed for transparency and electrical

properties – gas pulsing can reduce resistivity peaks

• DLIM produces lower substrate heating

• DLIM TCO magnetics (with anode) lowers resistivity

• Under optimum conditions high resistivity peaks can be

eliminated

• More tests are needed to achieve close to the best parameters

and explore all possibilities

Conclusions AZO

Acknowledgements

• For AC rotatable pairs the DLIM linked magnetic design improves the

electrical properties of an AZO based TCO for both ceramic and

reactive processing routes.

• Reactive AZO deposited from dual rotatable magnetrons can be

readily tuned over a wide range and all have much lower internal stress

than the ceramic approach.

• Reactive AZO deposited with DLIM and MF power show equally good

or better properties at without substrate heating when compared to

elevated temperatures allowing high quality deposition onto temperature

sensitive substrates and energy savings.

• Reactive ITO is optimised with DLIM magnetics and elevated substrate

temperatures with a plasma interaction effect varying with temperature.

• Reactive ITO displays low resistivity with AC power

• Special thanks to Heraeus for providing AZO and Zn:Al targets and to

the Indium Corporation of America for the In:Sn target.