a helicon source requires a dc magnetic field.. u. wisconsin
DESCRIPTION
...and is based on launching a circularly polarized wave in the plasma UCLATRANSCRIPT
A helicon source requires a DC magnetic field..
U. Wisconsin
...and is based on launching a circularly polarized wave in the plasma
UCLA
+
_+
_ _
+
+ +
_ _
_
B
k
(a)
(b)
(c)
+
The antenna can be twisted to match the helicon's helical waveform
UCLA
B, k
Right Helical Antenna
B
Nagoya III Antenna
(a) (b)
The R-wave propagates to the right, andthe L-wave to the left (for this antenna helicity)
UCLA
--
+ --
+
--
+ --
+
(a)
(b)
(c)
B
But the L-wave is very weak, and this antenna is unidirectional
Why do helicons ionize so well?
UCLA
Trivelpiece-Gould mode
Helicon mode
Landau damping (1985)F.F. Chen, Plasma Phys. Control. Fusion 33, 339 (1991).
This was disproved (1999)F.F. Chen and D.D. Blackwell, Phys. Rev. Lett. 82, 2677 (1999).
Mode-coupling to TG modes(1996)K.P. Shamrai and V.B. Taranov, Plasma Sources Sci. Technol. 5, 474 (1996).
Parametric excitation of ion acoustic waves (2005).B. Lorenz, M. Krämer, V.L. Selenin, and Yu.M. Aliev, Plasma Sources Sci.Technol. 14, 623 (2005).
Two commercial helicon reactors
UCLAThe Boswell source
The PMT (Trikon) MØRI source
Distributed source: first attempt
UCLA
QUARTZ TUBE
PVC PIPE
ANTENNA
MAGNET WINDING
7 cm
5 cm
13 cm
BNC connector
5 mm
17 mm
1 cm
1 cm
10 cm
ELECTROSTATIC CHUCK
WAFER
MULTI-TUBE HELICON PLASMA SOURCE
PE
RM
AN
EN
T M
AG
NE
T A
RR
AY
Conceptual RF plasma source for etching and depostion of semiconductor wafers and flat-panel display substrates.
Each tube with a solenoidal coil and helical m = +1 antenna
A 7-tube circular array.This failed to produce high density.
Reason: Diverging field lines
UCLA
-10
0
10
20
30
z (cm)
Distributed source: Second attempt
UCLA
ELECTROSTATIC CHUCK
WAFER
PE
RM
AN
EN
T M
AG
NE
T A
RR
AY
SOURCE WITH LARGE MAGNETS
This was better
Distributed source: Third attempt
UCLA
ROTATING PROBE ARRAY
PERMANENT MAGNETS
3"
DC MAGNET COIL
18"
54 mm2.4 mm
6.4 mm10 cm
2.5
cm
The “stubby” tube
This worked beautifully! But…
Plasmas merged; density is uniform
UCLA
Power scan at z = 7 cm, 5 mT A, 20 G, 13.56 MHz,
0.0
0.5
1.0
1.5
2.0
0 5 10 15 20 25 30R (cm)
N (
1012
cm
-3) 3.0
2.52.01.51.0
P(kW)
7-tube m=0 array
ARGON
…but the size is limited by the single large electromagnet.
Characteristics of permanent magnet rings
UCLA
Internal field
External field
Experiments with 7-cm diam tube
UCLA
Gate Valve
To Turbo Pump
34 cm
36 cm
D
Z1
Z2
0
20
40
60
80
100
120
5 10 15 20 25 30 35 40z (cm)
Bz
(G)
B (0)2319283338
D (cm)
-300
-250
-200
-150
-100
-50
0
50
100
150
0 5 10 15 20 25 30z (cm)
Bz
(G)
CalculatedMeasured
External field
Internal field
Radial density profiles at Z1 and Z2
UCLA
0
2
4
6
8
10
-5 0 5 10 15 20r (cm)
n (1
010cm
-3)
Z1, 40Z1, 35Z1, 30Z1, 21Z1, 1
D (cm)500W, 1mTorr
0
1
2
3
4
5
6
7
-5 0 5 10 15 20r (cm)
n (1
010cm
-3)
Z2, 40Z2, 35Z2, 30Z2, 21Z2, 1
D (cm)
500W, 1 mTorr
Upper probe Lower probex 1010 cm-3
Proof of principle: discharge in the external field gives much more plasma downstream.
Optimization of magnet geometry
UCLA
Result: Field strength magnet volumeSpacing improves uniformity slightly
actual
actual
Optimization of discharge tube: HELIC code
UCLA
a
b
c
Distant conducting shell
antenna
plasma
LLc
dEndplates have arbitrary reflectivity
D. Arnush, Phys. Plasmas 7, 3042 (2000).
Radial profiles are arbitrary, but B and n must be uniform axially.
HELIC gives not only the wave fields but also R, the loading resistance.
The low-field peak
UCLA
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1E+11 1E+12 1E+13n (cm-3)
R (
ohm
s)
100 G
63 G
40 G
25 G
16 G10 G Typical HELIC result
Relation of R to plasma density
UCLA
pin rf
p c
RP P
R R
7 1/ ½ 1.0 10 secsdN dt Snc n
c i eW E W W 101.1 10 WattsoutP n
Rp << Rc
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1E+11 1E+12 1E+13n (cm-3)
R (
ohm
s)
100 G
63 G
40 G
25 G
16 G10 G
Relation of R to plasma density
UCLA
pin rf
p c
RP P
R R
7 1/ ½ 1.0 10 secsdN dt Snc n
c i eW E W W 101.1 10 WattsoutP n
10
100
1000
1E+11 1E+12 1E+13n (cm-3)
Pin (
W)
1000500200100Loss
P rf (W)
Rp > Rc
HELIC and expt. matrices varied a, L, f, p, endplate
UCLA
Matrix A: tube dimensions Matrix B: frequency-diameter
2" long 2.0 MHz
4" long 13.56 MHz
6" long 27.12 MHz
Matrix C: pressure-frequency Matrix D: Boundary condition
2.0 MHz ei = -1
13.56 MHz ei = +1
Use MatrixD.csv input file Goes to higher density.27.12 MHz Diam = 4", freq. = 27.12
DL2ins DL4ins DL6ins
ZA = 0.975 ZA = 0.949 ZA = 0.924
ZA = 0.975 ZA = 0.949 ZA = 0.924
DL2cond DL4cond DL6cond
2" long 4" long 6" long
ZA = 0.949
Cp1f27_
ZA = 0.949
Cp3f27_
Cp1f13_
a = .038
Cp3f13_
a = .038
ZA = 0.949
ZA = 0.949 ZA = 0.949Standard
ZA = 0.949
a = .038ZA = 0.949
AD2L6_ AD3L6_ AD4L6_ BD3f27_ BD4f27_
2" diam 3" diam 4" diam
ZA = 0.949 ZA = 0.949 ZA = 0.949a = .025 a = .038 a = .051
ZA = 0.949 ZA = 0.949
ZA = 0.949 ZA = 0.949Standard
a = .038 a = .051
BD2f13_ BD3f13_ BD4f13_
BD3f2_ BD4f2_
a = .025 a = .038 a = .051
BD2f2_
ZA = 0.949
a = .025
BD2f27_
ZA = 0.924
ZA = 0.949 ZA = 0.949
ZA = 0.924 ZA = 0.924
StandardAD2L4_ AD3L4_ AD4L4_
a = .051
a = .051
AD3L2_ AD4L2_
4" diam
ZA = 0.975
a = .025
a = .025
ZA = 0.975
AD2L2_
ZA = 0.975
ZA = 0.949a = .038
a = .038
1 mTorr 3 mTorr 10 mTorr
a = .038a = .025 a = .0512" diam 3" diam
a = .038ZA = 0.949
Cp10f2_
a = .038 a = .038 a = .038
Cp1f2_ Cp3f2_
a = .038ZA = 0.949
Cp10f13_
a = .038ZA = 0.949
Cp10f27_
Examples: Tube diameter, frequency
UCLA
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
1E+11 1E+12 1E+13n (cm-3)
R (
ohm
s)
d = 10.2 cmd = 7.6 cmd = 5.1 cm
Tube diameter
0.0
0.5
1.0
1.5
2.0
1E+11 1E+12 1E+13n (cm-3)
R (o
hms)
100795020805020
13 MHz: solid points2 MHz: open points
B(G)
Larger diameter gives higher plasma resistance, but this is
not practical.
13.56 MHz is much better than 2 MHz.
Final design
UCLA
5.1 cm
10 cm
5 cmANTENNA
GAS INLET
-16
-14
-12
-10
-8
-6
-4
-2
0
2
4
6
8
-10 -8 -6 -4 -2 0 2 4 6 8 10Very similar to “stubby” tube, designed by intuition! Only improvement is the metal top. A single NdFeB magnet
The magnets are dangerous!
UCLA
12.7 cm
7.6 cm
PLASMA
Material: NdFeB
Bmax = 12 kG
Attractive force between two magnets 2 cm apart:
516 Newtons = 53 kG
Wooden frame for safe storage
UCLA
Single tube, final configuration
UCLA0
50
100
150
200
250
300
350
-4 -2 0 2 4r (cm)
Bz
(G)
56789101214
z (cm from midplane of magnet)
0
2
4
6
8
10
-5 0 5 10 15 20r (cm)
n (1
011
cm-3
)
Z1, single magnetZ1, with neighborsZ2, single magnetZ2, with neighbors
Radial density profilesat Z1 = 7.4 cm and Z2 = 17.6 cm below discharge.
Radial Bz profiles at various distances below the magnet.
Discharge tube
Design of array
UCLA
0
2
4
6
8
10
12
15 20 25 30L (cm)
Rip
ple
(+/-
%)
L = 17.5 cm gives < 2% ripple at 17 cm below source
The density at Z2 is summed over nearest tubes.
For a single row, a distance L = 17.5 cm between two tubes gives less than 2% ripple in density.
Computed uniformity n(x) for various y
0
1
2
3
4
5
6
7
8
-30 -20 -10 0 10 20 30x (cm)
n (1
011
cm
-3)
SumRow1Row2Avg: all y
y = 0
(a)
0
1
2
3
4
5
6
7
8
-30 -20 -10 0 10 20 30x (cm)
n (1
011
cm
-3)
SumRow1Row2Avg: all y
y = 7.5
(b)
0
1
2
3
4
5
6
7
8
-30 -20 -10 0 10 20 30x (cm)
n (1
011
cm
-3)
SumRow1Row2Avg: all y
y = 15
(c)
0
1
2
3
4
5
6
-30 -20 -10 0 10 20 30x (cm)
n (1
011
cm
-3)
SumRow1Row2Avg: all y
y = 22.5
(d)
Half-way between rows 1/4-way between rows
Directly under a row Beyond both rows
An 8-tube linear test array
UCLA
165 cm
53.3 cm
17.8
17.8
17.8
17.8 cm73.7 cm
8.9 cm
x
y
The array source is vertically compact
UCLA
165 cm
30.5
17.8 cm
The magnets are to be stuck onto an iron plate, which holds them and also concentrates the flux.
Once placed, the magnets cannot easily be moved, so for testing we use a wooden support.
The wooden magnet frame is used in testing
UCLA
64"
21"
3.5"
12"
Wooden magnet support
2.00
5.00
3.75
2.5
24 ea. 2.5 x 5.5 x 1/2"4 ea. 2.5 x 64 x ¾”
1 ea. 21 x 64 x 1/2”16 ea. 1/4” diam dowels
An 8-tube staggered array in operation
UCLA
Possible applications
UCLA
• Web coaters• Flat panel displays• Solar cells• Optical coatings
A web coater