1Department of Solid State Sciences
Title
Discharge voltage variations when switching from pure metallic to poisoned sputtering
R. De Gryse, D. Depla, G. Buyle, J. Haemers
AIMCAL
2005
MYRTLE
BEACH
2
Chemisorption: experiment
AIMCAL
2005
MYRTLE
BEACH
360
340
320
300
280
260
targ
et v
olta
ge (
V)
5 6 7 8 9
12 3 4 5 6
oxygen flow (sccm)
Constant current: 200 mATarget: AluminiumConstant Ar pressure: 2x10-3 mbar
5 L/s28 L/s42 L/s228 L/s469 L/s
Situating the problemPoisoning : drop in target voltage
: exceptions!
3
Reactive sputtering: different oxides (I)
AIMCAL
2005
MYRTLE
BEACH
600
500
400
300
200
100
0
Dis
cha
rge
vo
ltage
(V
)
Ag Al Au Ce Cr Cu Mg Nb Pt Re Ta Ti Y
target material
time
sta
tus
sta
tus
voltage
sta
tus
argon
oxygen
magnetron
Dt
Experimental conditions : 0.3 A, 0.3 Pa, waiting time : 5s, target diameter : 50 mm
VAr
VO2
Vox,Ar
How behave metals ?
4
Reactive sputtering: different oxides (II)
AIMCAL
2005
MYRTLE
BEACH
The full line shows the discharge voltage behaviour for oxidized targets sputtered in pure Ar, i.e. each curve shows the transition from VoxAr
to VAr, or stated differently the sputter removal of the formed oxide layer on the target surface.
Experimental conditions: constant current 0.3 A, argon pressure 0.3 Pa. The dotted line shows the discharge voltage behaviour during the sputtering (in pure Ar) of a target exposed to oxygen (total exposure approximately 1.4x105 L)
410
400
390
0.001 0.1 10
350
300
0.001 0.1 10
490
480
470
460
0.001 0.1 10
250
200
150
0.001 0.1 10
380
370
360
350
0.001 0.1 10
420
400
0.001 0.1 10
300
200
100
0.001 0.1 10
450
400
350
0.001 0.1 10
500
480
460
440
0.001 0.1 10
440
420
400
380
0.001 0.1 10
400
350
0.001 0.1 10
400
380
360
340
0.001 0.1 10
250
200
150
0.001 0.1 10
time (s)
dis
ch
arg
e v
olta
ge
(V
)
Ag Al
Au Ce
Cr Cu
Mg Nb
Pt Re
Ta Ti
Y plasma exposure
oxygen exposure (1.4x105 L)
5
Target voltage simulation : pumping speed
AIMCAL
2005
MYRTLE
BEACH
360
340
320
300
280
260
targ
et v
olt
age
(V)
6543210flow O2 (sccm)
-1.0
-0.8
-0.6
-0.4
-0.2
0.0
0.2
qtc
-qti
6543210
5 L/s 28 L/s 42 L/s 224 L/s 469 L/s
360
340
320
300
280
260
targ
et v
olt
age
(V)
6543210flow O2 (sccm)
-1.0
-0.8
-0.6
-0.4
-0.2
0.0
0.2
qtc
-qti
6543210
5 L/s 28 L/s 42 L/s 224 L/s 469 L/s
simulation
experiment
Chemisorption
Shallow implantationOxide layer
6
Chemisorption vs. implantation
AIMCAL
2005
MYRTLE
BEACH
PAr= 2x10-3 mbar SO2 = 0.1
PO2= 5x10-4 mbar SO = 1
PO= 5x10-5 mbar
2 2A
Pz 2 . 6 3 x 1 0
M T=
A. Chemisorption
Impingement rate ZA (O2) = 1.34x1017 cm-2s-1
ZA (O) = 1.9x1016 cm-2s-1
Condensing rate CA (O2) = 1.34x1016 cm-2s-1
CA (O) = 1.9x1016 cm-2s-1
å CA (O) = 3.2x1016 cm-2s-1
B. Shallow implantation of ions (subplantation)
Discharge current density = 25 mA.cm-2 or 1.56x1017 cm-2s-1
mole fraction oxygen (O) : f = 0.2 3.12x1016 cm-2s-1
Ar+ O
O2+
O2O+
7
Overview : two effects
AIMCAL
2005
MYRTLE
BEACH
Neutral SpeciesChemisorption at
target surface
Ionic Species 2( R , R )+ +Shallow implantation
Compound growth
Dissolved reaction gas
POISONING: 2 EFFECTS
o o( R , R )2
8
AIMCAL
2005
MYRTLE
BEACH
336
334
332
330
abso
lute
tar
get
volt
age
(V)
0.60.50.40.30.20.10.0mole fraction nitrogen
on addition on removal
No hysteresis
Constant Ar flow
Constant power (50 W)
Reactive sputtering of Ag in Ar/N2
Why maximum ?
Hysteresis Ag/N2
9
AIMCAL
2005
MYRTLE
BEACH
Target Poisoning
VD
VD
VD ?
Chemisorption of metal species
Dissolved gas
Compound growth
Shallow implantationof charged species
Overview : compound growth ?
10
AIMCAL
2005
MYRTLE
BEACHi
Thornton relation
0D
0 i
WV =
e e g
VD : Discharge or target voltage
W0 : effective ionisation energy
ei : ion collection efficiency (for magnetron : almost 1)
e0 : fraction of maximum possible number of ions VD /W0 produced by an electron
(for magnetron : almost 1)
g : the effective secondary electron emission coefficient
I S E Em Eg = g × ×
gISEE: Ion induced secondary electron emission
m : multiplication factor
E : Effective gas ionisation probability
Thornton equation
11
AIMCAL
2005
MYRTLE
BEACH
oD
i e
WV
E ( p ) I S E E m=
g × × e × e
1 i eD I S E E
oV E ( p ) m
w- e × e
= × × × g
1D I S E EV A B- = g +
Relationship discharge voltage ISEE
12
AIMCAL
2005
MYRTLE
BEACH
Experimental result
4.0
3.5
3.0
2.5
2.0inve
rse o
f th
e d
isch
arg
e v
olta
ge (
10
31
/V)
0.200.150.100.050.00
ISEE coefficient
AlAgAuCeCrCuMgNbPtReTaTiY
13
AIMCAL
2005
MYRTLE
BEACH
288.80.383162.20.178274.8Y
428.60.08391.90.114335.6Ti
431.90.058435.30.115321.1Ta
616.80.0394820.084379.7Re
588.30.022531.40.048461.0Pt
480.90.045468.50.130308.1Nb
270.10.414156.30.136330.0Mg
402.90.072411.10.082388.3Cu
428.30.094379.20.091359.6Cr
3660.285207.70.184277.9Ce
653.90.037487.70.057471.2Au
305.80.2259.10.091376.4Al
549.20.069410.30.110394.9Ag
VO2 (V)ISEE oxidized. target
VOxAr (V)ISEE metalVAr (V)Target material
Some numbers
14
AIMCAL
2005
MYRTLE
BEACH
30 msec before steady state is reached
Total ion flux: 1016 ions/cm2
à Sputtering of the oxide layer on the target
à Preferential sputtering of oxygen ?
à Formation of suboxides ?
Wittmaack(*): suboxide formation results in a
negligible yield enhancement
s
b
( M / O )R
( M / O )=
Reduction factor
M,O are the mole fractions of metal and oxygen at surface (s) and in the bulk (b)
(*)K. Wittmaack , Surf. Sci. 419 (1999) 249-264
VAr
VO2
Vox,Ar
Reduction factor : definition
time
sta
tus
sta
tus
voltage
sta
tus
argon
oxygen
magnetron
Dt30 msec
15
AIMCAL
2005
MYRTLE
BEACH
The influence of the reduction on the inverse of the ISEE coefficient. The reduction was calculated using SRIM. The noble elements (Ag, Au, Cu and Pt) have not been included as the SRIM calculation of their sputter yield was not in agreement with the experimental results
Influence of reduction
25
20
15
10
5
0
1/IS
EE
1.21.00.80.60.4
reduction R
Al
Ce
CrMg
Nb
Re
Ta
TiY
s
b
( M / O )R
( M / O )=
1 1.41.2 1.81.6
16
Preferential sputtering oxygen(*)Surface reductionSuboxide formation
AIMCAL
2005
MYRTLE
BEACH
Target Poisoning
Chemisorption of neutral species
Shallow implantationof charged species
Compound growth(oxides, nitrides…
Dissolved gas
Congruent sputtering(Al, Mg, Ce, Y)
(*)exact mechanism not known yet
VD
VD
VD
VD
Conclusions
17
Acknowledgements
AIMCAL
2005
MYRTLE
BEACH
Acknowledgements
The authors are indebted to the
for financial support
company
Department of Solid State Sciences
BACK TO LIST