maximizing the potential of rotatable magnetron …...maximizing the potential of rotatable...
TRANSCRIPT
Maximizing the Potential of Rotatable Magnetron Sputter Sources for Web Coating Applications V.Bellido-Gonzalez, Dermot Monaghan, Robert Brown, Alex Azzopardi, Gencoa, Liverpool UK
• Anode importance in planar and rotatable magnetrons and effect on substrate heating • Magnetic options for rotatable magnetrons for web coating and heat load effects • Case study: electrical and optical properties of reactive and non-reactive AZO layers formed with different rotatable magnetic geometries and varying substrate temperatures • Conclusions
NREL
Structure of presentation
A magnetron sputtering plasma
+ + -
Neg
ativ
ely
bias
ed ta
rget
-V
High density plasma by exb field
Resulting erosion of the sputter target
Confinement between a negatively biased target and ‘closed’ magnetic field produces a dense plasma.
High densities of ions are generated within the confined plasma, and these ions are subsequently attracted to negative target, producing sputtering at high rates.
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.
Electrons will spiral around field lines until enough energy is lost to escape
the magnetic trap.
If an anode intersects a magnetic field line it will collect the electrons, so they are
lost to the plasma and do not add to substrate heating
Comparison of the plasma expansion with an anode that intersects with the magnetic field and
one moved just 1mm to avoid a magnetic interaction
Whilst for a planar magnetron discharge and anode can be used to confine the plasma,
typically for rotatable magnetron no anode is close-by
Rotatables great for target life and target use but not good for substrate heat load
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
There are several factors that contribute to the overall heat load on a substrate: • Positive ions from the plasma. • Electrons (primary and secondary) from the plasma. • Thermal energy input due to the heat of condensation of the atoms. The thermal energy from the coating flux is comprised of the standard enthalphy (heat of condensation) of the given material plus the kinetic energy of the atoms. That leaves the heating effect from the plasma electrons and ions. For a DC based magnetron discharge this can be as high as 75% of the heat load and 95% for an RF magnetron based plasma [1]. Since the enthalphy is unavoidable during coating, the major means of reducing heat on the substrate is via the plasma control.
Heat load on substrates from magnetron sputtering plasmas
Heat load on substrates contributions dominated by primary and secondary electrons
and argon ions – plasma heating rather than atomic
M.Andritschky et al, Vacuum/volume 44, pages 809 to 813, 1993, 0042-207X
Heat load on substrates from different DC & AC plasma and
with different anode arrangements
Magnetic design and anode position will affect the substrate heating for rotatable
magnetrons in the same way as planar magnetrons
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 – definitely not perfect for temperature of web!
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-
Plasma to substrate interaction by assymetric magnetics and tilting
New 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
(ind
icat
or)
probe position
Temperature on probes across (every 25 mm)
T across DLIMT across BOC
Comparison of substrate temperature in-front of a double AC rotatable magnetron
DLIM has a 20̊C lower temperature for same conditions
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
DC discharges the angle of the magnetic pack relative to the magnetic anode can be
adjusted to drive the plasma away from the substrate
The anode has a combined magnetic trapping with electron acceleration due to either a positive bias or as the floating earth return for the power supply.
Supplementary magnetic anodes for rotatable cathodes with DC & DC pulsed power
More stable environment to avoid process drifts
The introduction of an optional magnetically guided hidden auxiliary anode and gas bar offers the following benefits: • lower plasma heating of the substrate – x3 power possible for web coating •reduced substrate movement influence on the plasma impedance • lower discharge voltages – lower impedance – lower TCO resistivity • less drift of plasma impedance and instability for non-conducting layers • more consistent uniformity •gas injected uniformly and protects hidden anode surface
Gencoa Rotatable Magnet Bar
Products Applications Component
parts supplied
LS Low strength
Low strength for higher voltage sputtering
RF Radio frequency
Strength optimized for RF power modes with active anode
SSF Standard Strength
Focused
Standard field strength of 550 Gauss over the target with balanced field design
PP-RT Unbalanced ion
assist processes
Single and multi-cathode unbalanced magnetic designs for high levels of ion assistance for deco and hard coatings
HSS700 HSS850
HSS1000 High strength options
High Strength Single for thicker targets or lower discharge voltages – range of 700, 850 & 1000 Gauss versions available
TCO Transparent
conduction oxide films
Single cathode magnetics with active anode for reduced resistivity TCO layers for DC and DC pulsed operation (patented)
LH/Web Single cathode with
lower heating of substrate
Single cathode magnetics with active anode for reduced heat loads during vacuum web coating for DC and DC pulsed operation – allows up to 3 x the power level compared to conventional magnetics (patented)
DLIM For better dual
cathode AC discharges
Double cathode Low Impedance Magnetics for high rate reactive deposition of oxides with lower substrate heating and plasma interference (patented)
DLIM-DC-TCO Single anode shared between 2 cathodes
for TCO
Double cathode low TCO resistivity magnetics for DC powered double magnetrons with an additional active anode (patented)
Gencoa have developed a wide range of magnet bar options for rotatable
magnetrons in order to control the plasma better
Magnet pack
Active magnetically guided anode
CASE STUDY use of DLIM 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 13Power (kW)
Har
d ar
c co
unt
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)
Shee
t res
ista
nce
(Ohm
/sq)
8.4e-4 9.2e-4
7e-4
* Szyszka et al
Controlled reactive sputtering will yield lower costs in production than ceramic AZO
Price will be 70-50% current ceramic based costs
Reactive gas controllers ‘common’ in other optical coating sectors
Reactive Sputtering Hysteresis
ZnAl dual rotatable + O2 - TRIPLE RAMP CONDITIONING
0
10
20
30
40
50
60
70
80
90
100
0 50 100 150 200 250Time, s
Sign
als,
% SetPoint (%)Sensor (%)Actuator (%)
Reactive Sputtering Control Plasma Emission
Plasma Emission for ZnAl + O 2 during control
-5000
0
5000
10000
15000
20000
25000
30000
35000
40000
250 350 450 550 650 750 850
Wavelength, nm
Pla
sma
Em
issi
on, c
ount
s
ZnAl + O2 in control
Different sensor control modes possible for reactive AZO
Penning-PEM
Target V
Lambda
Process-PEM
O2 gas
Reactive dual rotatable AZO deposition process window
ZnAl + O2 reactive control dual rotatable
0
10
20
30
40
50
60
70
80
90
100
0 100 200 300 400 500Time, s
MFC
feed
back
, scc
m
0
20
40
60
80
100
O2 P
EM a
nd ta
rget
V
Sens
ors
(%)
Gas Feedback (SCCM)O2 PEM value %O2 PEM setpoint %Target V %
Process control using plasma monitoring
DLIM magnetics dual magnetron AC power current / voltage characteristics
475 mm long Zn:Al target no O2, 0-8kW AC power
0
2
4
6
8
10
450 500 550 600 650 700 750 800 850
U, V
I, A
2mTorr5mTorr10mTorrExpon. (10mTorr)Expon. (5mTorr)Poly. (2mTorr)
0
50
100
150
200
250
300
350
400
450
500
Deposition conditions
depo
site
d th
ickn
ess
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 Mean thickness across the deposition zone
1.00E-04
1.00E-03
1.00E-02
1.00E-01
1.00E+000 2 4 6 8 10 12
resi
stiv
ity, O
hm-c
m
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
1.00E-04
1.00E-03
1.00E-02
1.00E-01
1.00E+000 2 4 6 8 10 12
resi
stiv
ity, O
hm-c
m
Sample position
Resisitivity BOC & DLIM at room temperature (every 25 mm)
resistivity BOCresistivity DLIM
Comparison of reactive AZO in-front of a double AC rotatable magnetron Comparing the 2 different magnetic designs
1.00E-04
1.00E-03
1.00E-02
1.00E-01
1.00E+000 2 4 6 8 10 12
resi
stiv
ity, O
hm-c
m
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 f ceramic AZO in-front of a double AC rotatable magnetron Comparing 2 different substrate temperatures
TCO film properties and rates also depend on target rotation speed
R07# NO TARGET ROTATION
0
500
1000
1500
2000
2500
3000
0 2 4 6 8 10 12sample (every 25 mm)
Thic
knes
s, n
m
1.00E-04
1.00E-03
1.00E-02
1.00E-01
1.00E+00
Ohm
-cm
thicknessresistivity
target rotation speed: 0 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
Higher resistivity on areas of high negative Oxygen ion bombardment
Log scale
Resistivity map for static deposition across 2 targets
Reactive AZO properties under same conditions but with target rotation
R08# as R07 with Target Rotation
0
500
1000
1500
2000
2500
3000
0 2 4 6 8 10 12sample (every 25 mm)
Thic
knes
s, n
m
1.00E-04
1.00E-03
1.00E-02
1.00E-01
1.00E+00
Ohm
-cm
thicknessresistivity
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
Log scale
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 12sample (every 25 mm)
Thic
knes
s, n
m
1.00E-04
1.00E-03
1.00E-02
1.00E-01
1.00E+00
Ohm
-cm t (at 150ºC)
t (at RT)r (at 150ºC)r (at RT)
Log scale
Room temperature films have better optical density
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 12sample (every 25 mm)
Opt
ical
Den
sity
at 5
50nm
1.00E-05
1.00E-04
1.00E-03
1.00E-02
Ohm
-cm 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
Tran
smis
sion
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
Room temp transparency for ceramic and reactive films – both 20 Ωsq films
Conclusions
Acknowledgements
• Alternative magnetic designs have been developed to control the plasma electrons in rotatable based sputtering arrangements in order to limit the heating of web based substrates. • The use of magnetic guiding of the electrons away from the substrate will limit heating. • For AC rotatable pairs the DLIM design is optimum and for DC power the use of an auxiliary magnetic anodes with or without a positive biasing. • 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.
• Special thanks to Heraeus for providing AZO and Zn:Al targets.