plasma sources iv: high density sources: icp, ecr, mw · this is not a magic but a legal number by...
TRANSCRIPT
1 U. Czarnetzki, ISLTPP, Bad Honnef 2016 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Plasma Sources IV:
High Density Sources:
ICP, ECR, MW
U. Czarnetzki
International School on
Low Temperature Plasma Physics:
Basics and Applications
Physikzentrum Bad Honnef, October 2-8, 2016
2 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Outline
• Some Applications.
• Some general wave properties.
• ICP discharges.
• ECR discharges.
• Microwave discharges.
3 U. Czarnetzki, ISLTPP, Bad Honnef 2016 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Some Applications
4 U. Czarnetzki, ISLTPP, Bad Honnef 2016 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Application:
Etching of micro structures in silicon
Creation of ions and radicals and synergistic directional etching:
Reactive ion etching and plasma aided chemical vapor deposition.
5 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Application: Deposition of Thin-Films
6 U. Czarnetzki, ISLTPP, Bad Honnef 2016 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Some General Wave Properties
7 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Basic Wave Equation General wave equation (linear equation in E!):
2 2
0 0 0 0E E/c / E+ / .extj j
No space charge for transversal waves:
(only this case is considered here) 0/ E E 0.ik
• Only time varying currents contribute.
• External currents jext are source terms.
• They can be included in the boundary conditions.
• Without magnetic field, the conductivity is isotropic,
i.e. it is a scalar, with magnetic field it is a tensor.
.j E
The current in the plasma is proportional to the electric field:
The conductivity follows from the momentum equation.
8 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Dispersion Relation
The conductivity can be complex.
Then also k is complex (for real w).
The wave equation together with the conductivity leads to
the dispersion relation (for simplicity here scalar ):
2 2 2
0 0
0
( ) , using: 1/ck i c
w w
Example of a homogeneous
(unmagnetized) plasma:
22
0 0
pee
e
e n
m i i
w
w w
• The imaginary part of the conductivity contributes to the
real part of the wave vector.
• The real part of the conductivity contributes to the
imaginary part of the wave vector.
9 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Damping of the wave
Complex wave vector:
Electric field: expE i k r tw
Damped wave:
' i k''k k
exp '' exp ' r tE k r i k w
• The imaginary part of k (real part of ) leads to damping
of the wave (k’’ > 0).
• In general, waves can also be excited in a plasma (k’’ < 0).
• This happens in case of instabilities.
10 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Classical Skin Depth
Imaginary part:
Effective parameter:
Real part:
Skin depth (1/e amplitude depletion):
2
2
2
1''
2 1
''' ''
1
m
pe
w
w w wk
c w
kk k
w w
w
w
1/ '' exp /skin skink E z z
11 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Typical Parameters
RF discharge:
Pressure p = 10 Pa, hydrogen:
Effective parameter:
Skin depth:
Plasma frequency:
8 14 10m s
8 113.56 0.85 10f MHz sw
0.21 1w
(a)
MW discharge:
Effective parameter:
Skin depth:
10 12.45 1.54 10f GHz sw
385 1w
(b)
11 3 4 1 10 110 5.64 10 1.8 10pen cm s n sw
25.1skin
pe
ccm
w
w
1.6skin
pe
ccm
w
12 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Effect of Collisions
1 1( )
21
0 ( )
peskin
pe
cfor wfor w HF casec
for wfor w DC casew
ww
Special cases:
Collisions (dissipation)
enlarge the skin depth.
m
ww
13 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Electric Field Profile
Note: Also the real part of k depends on w.
• Without dissipation the wave is (classically)
simply evanescent.
• With dissipation, damped oscillations can occur.
(generally no practical consequences due to low amplitudes).
0 2 4 6 8 1010
-4
10-3
10-2
10-1
100
E /
E0
z / Skin
w / m = 0
w / m = 2
w / m = 10
14 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Phase Velocity
• With collisions, the evanescent wave has a
low phase velocity that can be much smaller than c.
• Without collisions, the phase velocity is infinite.
01
23
4
5
04
812
1620
0
>> w
za
mp
litu
de
(a
.u.)
w t
01
23
4
5
04
812
1620
0
<< w
z
am
plit
ud
e (
a.u
.)
w t
15 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Failure of the Local Picture
• The classical approach presented before breaks down in a
collisionless plasma (mean free path l >> skin depth skin).
• Then the conductivity is no longer a local quantity.
• For high collisionality, the local velocity depends only on
the local field at the local time.
• Without collisions, it depends on the entire history,
i.e. the trajectory in space and time.
• The correct description can no longer be performed
in a fluid dynamic picture but requires a kinetic treatment,
since the individual particle velocities matter.
• The entire interaction between wave and particles needs
to be treated in parallel.
• This affects the skin depth and the power deposition!
16 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Principle of the Non-Local Effect • Thermal flux of electrons towards the antenna window.
• Reflection of the electrons at the floating potential sheath
in front of the window.
• Gain and loss of energy occurs in different regions of
space.
• Due to the field inhomogeneity in the skin layer,
the net effect is non-zero.
01
23
4
5
04
812
1620
0
<< w
z
am
plit
ud
e (
a.u
.)
w t
antenna window
bulk plasma skin layer
Important:
The electric field
oscillates perpendicular
to the direction of the
(thermal) particle motion
and the field gradient!
17 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Anomalous Skin Effect
experiment and theory
• Alternation of positive and
negative power deposition.
• Net effect is positive: Heating.
18 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Consequences of the Non-Local Effect
• Net power deposition even without collisions:
Stochastic heating.
• The field decay is not exponential.
• Different scaling of the skin depth (larger than classical): 1/31/3
2
2
6
8 8
9 2 9 2
8.10 /
pe th thskin
pe pe
eth
vc c v
c
kTv typ m s
m
w
w w ww
• Remarks:
- The factor 8/9 is often skipped.
- Requirement for stochastic heating:
- Difficult to realize at low pressures.
- Most effective at lower RF frequencies, e.g. 1 MHz.
10 1
11 3
1, 13.56 :
2 10
1 10
pe th
pe
e
cAt f MHz
v
s
n cm
w
w
w
19 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Frequency and Pressure Dependence
• Experimental result from an
ICP discharge.
• The ratio between the total
(collisional) and the Ohmic
overall power deposition
(volume integrated) is
determined.
• The scaling is similar for
all frequencies.
• The non-local effect
becomes effective only for
pressures of less than 1 Pa.
20 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Some General Technical Consequences • The electron density in metals is much higher.
• Therefore the skin depth is much smaller.
• In copper typical values for RF frequencies are 10 m.
• For microwaves the skin depth is only about 1 m.
• The current is flowing only at the surface.
• Large surfaces are requires.
• The cross section does not matter.
• The surface quality matters.
• Contacts can cause strong losses.
• Polishing and coating with silver are common measures.
• Further: Impedances scale with frequency.
• Dielectric isolators can easily transmit displacement current,
tiny bends in wirings can cause noticeable inductances.
• Without impedance matching strong reflections occur.
21 U. Czarnetzki, ISLTPP, Bad Honnef 2016 U. Czarnetzki, ISLTPP, Bad Honnef 2016
ICP discharges
22 U. Czarnetzki, ISLTPP, Bad Honnef 2016 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Basic Principles of CCP and ICP Discharges
current in
the plasma
heating
time
sheath
voltage
time
heating
time
+
+
+
+
++
-
--
-VRF
electrode
plasma
sheath
I
I
A
Pl
antenna
plasma
E-field
B-field
CCP-mode:
„piston“-
principle
ICP-mode:
transformer-
principle
23 U. Czarnetzki, ISLTPP, Bad Honnef 2016 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Radio Frequency (RF)
• processing of non-conducting dielectrics (CCP)
• extra heating mechanism for electrons (CCP)
(not existing in DC discharges)
• operation at low pressures (< 1 Pa)
• high plasma densities (up to 1012 cm-3)
• large scale homogeneous discharge possible
• high efficiency for the transformer (ICP)
CCP and ICP discharges operate in the radio frequency regime:
The most common frequency is 13.56 MHz (and harmonics).
This is not a magic but a legal number by international
telecommunication regulations (also harmonics).
24 U. Czarnetzki, ISLTPP, Bad Honnef 2016 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Some General Characteristics I
• ICP discharges operate usually at an order
of magnitude higher powers than CCPs.
• They produce an order of magnitude more
dense plasmas.
• The plasma potential is low and quiet for
pure ICP operation.
• They can sometimes operate in a hybrid CCP/ICP mode,
since the antenna can also act as an electrode
(especially at low densities).
25 U. Czarnetzki, ISLTPP, Bad Honnef 2016 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Some General Characteristics II
• Often they are operated together with a separate CCP:
ICP for plasma generation, CCP for RF-bias.
• This allows independent control of the plasma density
(ion flux) and the bias (ion energy).
• The antenna coil is usually either a cylindrical or planar coil.
• A dielectric window (quartz or aluminium oxide) is required.
26 U. Czarnetzki, ISLTPP, Bad Honnef 2016 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Typical Antenna Coils
Faraday shield (spoke like structure):
• only poloidal fields can penetrate
• capacitive coupling is suppressed.
• disadvantage: No self-ignition,
high voltage between coil and the shield
Cylindrical Antenna Planar Antenna
27 U. Czarnetzki, ISLTPP, Bad Honnef 2016 U. Czarnetzki, ISLTPP, Bad Honnef 2016
0 2 4 6 8 100,0
0,2
0,4
0,6
0,8
1,0
1,2
1,4
1,6
1,8
2,0
XL ()000225a.o
rg
50
40
30
20
C2C1
resistance ()
capacitance (nF)
ICP Matching Unit
C
CL
R
R 2
1
in out
VRF
0
28 U. Czarnetzki, ISLTPP, Bad Honnef 2016 U. Czarnetzki, ISLTPP, Bad Honnef 2016
-120 -100 -80 -60 -40 -20 0 20 40 60 80 100 120
50
40
30
20
10
0B
ver [-0,009 ~ 0,061 mT/A]
R (mm)h
(m
m)
-120 -100 -80 -60 -40 -20 0 20 40 60 80 100 120
50
40
30
20
10
0E
pol [0 ~ 55 (V/m)/A]
R (mm)
h (
mm
)
Calculated antenna fields in vacuum:
Bz and Ej
The relevant induced electric field is in azimuthal direction.
29 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Induced Azimuthal Electric Field
(calculated for vacuum)
The radial current is causing distortion of the symmetry.
30 U. Czarnetzki, ISLTPP, Bad Honnef 2016
-150 -100 -50 0 50 100 150-150
-100
-50
0
50
100
150
radial positon (mm)
rad
ial p
osito
n (
mm
)
70.00 -- 75.00
65.00 -- 70.00
60.00 -- 65.00
55.00 -- 60.00
50.00 -- 55.00
45.00 -- 50.00
40.00 -- 45.00
35.00 -- 40.00
30.00 -- 35.00
25.00 -- 30.00
20.00 -- 25.00
15.00 -- 20.00
10.00 -- 15.00
5.000 -- 10.00
0 -- 5.000
Electric Field Induced by a Flat Coil Antenna
Flat coil with four parallel spirals, each 2.5 turns.
(HV in the centre, ground at the outer connectors)
D L Crintea, D Luggenhölscher, V A Kadetov, Ch Isenberg and
U Czarnetzki, J. Phys. D: Appl. Phys. 41 (2008) 082003
31 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Wave Equation for the Induced Field
• Two dimensional problem.
• Displacement current can be neglected since the plasma
frequency is much higher than the RF frequency.
• Analytical solution only for a homogeneous plasma.
• The flat antenna is approximated as a conducting disc
of radius R.
• Plasma is assumed as homogeneous!
• Inhomogeneous plasma: Numerical solution!
32 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Plasma Current Density
Local isotropic conductivity:
Ej
• For mean free paths of the order of the skin depth
or larger the conductivity is no longer local.
• Then the product of conductivity and field becomes
a convolution.
33 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Geometry and Boundary Condition
Conditions are set at the walls
and the interfaces:
generally
34 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Complex Solution for the Electric Field
The coefficients are:
Radially Bessel functions, axially exponentials:
ln: n-th zero of the
Bessel function 2
0pk iw
35 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Relevance of the Individual Terms
Only very few terms contribute, actually mainly 2!
distance from
antenna
number of Bessel
zero
Dominik Winter, Master Thesis, RUB (2010)
36 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Electric Field Depletion
Typical values:
• In reality, mostly a mixture of geometry and skin depth!
• Approximately exponential decay of the field.
• The depletion length has a geometrical and a plasma
contribution: 22
11
geometry skin effect
: 1/ , 3.83pe
L c
wl w l
1 typ. 0.1pe
cO ml
w
largechambers : classical skin depth behaviour
smallchambers : geometrical depletion
L
L
37 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Radial Profile
• Fundamental radial mode profile is a Bessel function.
• Only the first two Bessel functions contribute
significantly.
• Higher modes shift the maximum to smaller radii.
• Rule of thumb: Maximum at r = 2/3 R.
1 1 1 2 1 2
r rE r A J A J
L Ll l
Physics:
• B-field is a dipole.
• Naturally, no electric field
on axis and no field at the wall.
• Zero at the conducting wall due to
eddy currents.
38 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Comparison with numerical integration
analytical solution
numerical solution
(Comsol)
Dominik Winter, Master Thesis, RUB (2010)
39 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Self consistent numerical solution I
Plasma density and ion flux
Dominik Winter, Master Thesis, RUB (2010)
40 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Self consistent numerical solution II
Induced power and electron temperature
Dominik Winter, Master Thesis, RUB (2010)
41 U. Czarnetzki, ISLTPP, Bad Honnef 2016 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Radial Profiles of Te
-40 -20 0 20 40 60 80 100 120 1400,0
0,5
1,0
1,5
2,0
2,5
3,0
3,5
p = 10 Pa H2
1200 W
800 W
500 W
300 W
Te (
eV
)
radial position (mm)
Collisional losses are strong in hydrogen and the energy is balanced
locally. The electron temperature reproduces schematically the torus
structure of the induced electric field (radial direction).
Victor Kadetov, PhD Thesis, RUB (2004)
42 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Phase Resolved Velocity Distribution
Function in an ICP (Ar, 0.5 Pa, 1 kW, Thomson Scattering)
-6 -4 -2 0 2 4 60
100
200
300
400
500
600
700
t = 60 ns
inte
nsity (
a.u
.)
wavelength shift (nm)
t = 20 ns
Argon ICP discharge, f = 13.56 MHz, P = 1 kW, p = 0.5 Pa.
Accumulation over 90,000 shots (30 minutes) at each phase.
Temporally independent EEPF, oscillating drift velocity.
0 5 10 151
10
100
1000
EE
PF
(a
.u.)
energy (eV)
kTe = 2.6 eV
ne = 1.0x10
11 cm
-3
t = 20 ns, 60 ns, 80 ns
D L Crintea, D Luggenhölscher, V A Kadetov, Ch Isenberg and
U Czarnetzki, J. Phys. D: Appl. Phys. 41 (2008) 082003
43 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Harmonically Oscillating Drift Velocity
V/cm 0.66Eˆ
v wm
Ee
0 20 40 60 80-10
-5
0
5
10
15
20
dri
ft v
elo
city (
10
4 m
/s)
Phase (ns)
0 v )tsin( v v jw
D L Crintea, D Luggenhölscher, V A Kadetov, Ch Isenberg and
U Czarnetzki, J. Phys. D: Appl. Phys. 41 (2008) 082003
44 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Self consistent numerical solution III
Electric field amplitude and phase
Dominik Winter, Master Thesis, RUB (2010)
45 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Measurement (B-dot probe)
Very similar to simulation.
Dominik Winter, Master Thesis, RUB (2010)
46 U. Czarnetzki, ISLTPP, Bad Honnef 2016 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Optically Measured Field Penetration
Intensity and field are radially integrated and the axial dependence
and time evolution are displayed as coloured contour plots.
0,0 0,2 0,4 0,6 0,8 1,050
40
30
20
10
0Pathe Shift
Time/Period
ve
rtic
al p
ositio
n (
mm
)
Electric field
Victor Kadetov, PhD Thesis, RUB (2004)
01
23
4
5
04
812
1620
0
>> w
z
am
plit
ud
e (
a.u
.)
w t
47 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Measurement (optical modulation)
Very good agreement throughout.
Dominik Winter, Master Thesis, RUB (2010)
48 U. Czarnetzki, ISLTPP, Bad Honnef 2016 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Transformer Model
2
2
2
2
2
122
LR
LRRS
w
w
2
2
2
2
2
1221
LR
LLLLS
w
wwww
The plasma inductance and
resistance are transformed to
an effective antenna
inductance and resistance.
2ˆ2
1RFSdiss IRP
M. A. Lieberman and A. J. Lichtenberg,
Principles of Plasma Discharges and Materials Processing
(Wiley, New York, 1994).
49 U. Czarnetzki, ISLTPP, Bad Honnef 2016 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Low and High Density Operation
Low density operation
(skin depth >> plasma size):
e
P
SPP nR
RLR 1
w
eP
PSPPn
RRLR11
wHigh density (normal) operation
(skin depth << plasma size):
Transformed resistance:
22
2
12
)( PP
PSLR
LRR
w
w
RS
electron density
50 U. Czarnetzki, ISLTPP, Bad Honnef 2016 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Dissipated Power and
Operation Threshold
threshold for
ICP operation
Ploss
I3 < I
2
I2 < I
1
I1
Pdiss
electron density
2ˆRFediss InP
2ˆ1RF
e
diss In
P
eloss nP
Low density:
High density:
Power lost by the discharge
(collisions, flux to the wall):
• There is a minimum power (current) for ICP operation.
• Below this value, the discharge operates capacitively.
51 U. Czarnetzki, ISLTPP, Bad Honnef 2016 U. Czarnetzki, ISLTPP, Bad Honnef 2016
10-1
100
101
0,001
0,01
0,1
1
S = K (U / d)3 n-3/2
US = U4 / U
p3
= C (U / d)4 / n2
0011
05A
.OR
G P
lot1
US /
U
U / Up
Capacitive Coupling by the Antenna
• The quartz window and the sheath form a non-linear
capacitive voltage divider between the antenna and
the plasma.
• At high densities and/or low antenna voltages,
the capacitive coupling is reduced strongly.
52 U. Czarnetzki, ISLTPP, Bad Honnef 2016 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Effect of CCP-ICP Mode Transition on
the Matching
0,1 1 10 1000
200
400
600
800
1000
1200
C1 divergence
guide to the eye
optical emission
H2 gas
po
we
r (W
)
pressure (Pa)
Rs 0 C1
• Divergence of the matching capacitance C1 at
the transition point to CCP operation.
ICP
CCP
Victor Kadetov, PhD Thesis, RUB (2004)
s
s
53 U. Czarnetzki, ISLTPP, Bad Honnef 2016 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Effect on the Modulation
of the Optical Emission
Discharge in hydrogen at
p = 10 Pa:
mode transition CCP to ICP
at P = 150 W
Observation of the Balmera
emission line (integrated
spatially):
%100I
)t(Imodulation
Victor Kadetov, PhD Thesis, RUB (2004)
54 U. Czarnetzki, ISLTPP, Bad Honnef 2016 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Observation of the Mode Transition in the
Modulation of the Optical Emission
ICP discharge in hydrogen at p = 10 Pa, Balmera emission
Pulsed mode: P = 300 W, f = 1 kHz, 1:1
0
100
200
300
4000.0
0.5
1.0
1.5
2.0
60
80
100
120
140
160
modula
tion (
%)
power
(W)
time/period
Victor Kadetov, PhD Thesis, RUB (2004)
• ICP discharges always ignite as CCPs.
• Then the density rises, the sheath shrinks,
and the mode transition happens.
%100I
)t(Imodulation
55 U. Czarnetzki, ISLTPP, Bad Honnef 2016 U. Czarnetzki, ISLTPP, Bad Honnef 2016
• The mode transition between ICP and CCP has a strong
effect on the ion energy distribution function.
• ICP: Quasi mono-energetic, low-energy ions from the
floating potential sheath.
• CCP: High energy wide distributed ions from the RF sheath.
Effect on the Ion Energy
Victor Kadetov, PhD Thesis, RUB (2004)
56 U. Czarnetzki, ISLTPP, Bad Honnef 2016
1) Pulsed discharges and afterglow.
2) Electronegative discharges and instabilities.
3) Gas heating and electron pressure effects.
4) Atmospheric pressures.
5) Hybrid discharges ICP – CCP.
6) Many applications…..
Further Topics in ICPs
57 U. Czarnetzki, ISLTPP, Bad Honnef 2016 U. Czarnetzki, ISLTPP, Bad Honnef 2016
ECR Discharges
(and a little bit of Whistlers, Helicons)
58 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Addition of a Weak Axial Magnetic Field
Argon p = 0.1 Pa, B = 10 mT, lme = 1 m = 2 x diameter.
ICP,
without
B-field
Helicon,
with
B-field
static magnetic
field coil
antenna
Y. Celik et al., POP 18, 022107 (2011)
59 U. Czarnetzki, ISLTPP, Bad Honnef 2016
RF Modulation Spectroscopy Upstream
oscillatory
velocity (2w)
azimuthal
drift
velocity (1w)
Y. Celik et al., POP 18, 022107 (2011)
60 U. Czarnetzki, ISLTPP, Bad Honnef 2016
RF Modulation Spectroscopy Downstream
phase (1w)
• Modulation can still be observed downstream.
• A helicon wave propagates along the field lines.
• The diamagnetic drift (grad(Te)) displays the RF phase front.
• The phase front is bended due to the inhomogeneous
dispersion by the density profile
drift velocity (1w)
Y. Celik et al., POP 18, 022107 (2011)
61 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Helicon Wave Magnetic Field (B-Dot Probe)
-10 -5 0 5 10 15 20-40
-20
0
20
40
60
80
100
120
Bz (
T
)
r (cm)
ICP (no static B-field) Helicon (with static B-field)
• The antenna field determines
the Bz(r) wave field.
• Maxwell’s equations then
determine the Helicon Bj(r):
)()( rBr
rB z
j
Y. Celik et al., POP 18, 022107 (2011)
62 U. Czarnetzki, ISLTPP, Bad Honnef 2016 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Electromagnetic Fields Propagating Along
Static Magnetic Fields
ie tzki
1EE 0
w
• The wave vector k points along the field lines.
• All electric fields are perpendicular to the wave vector k.
• Two fundamental polarization states:
- R-wave (rotating clockwise, with the electrons, “+”)
- L-wave (rotating counter clockwise, with the ions, “-”)
• x- and y-components are 90° out of phase.
• The Lorentz force is important in this case:
e
mv E v Bm
c
e
e B
mw
• Electrons rotate clockwise with the cyclotron frequency:
63 U. Czarnetzki, ISLTPP, Bad Honnef 2016 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Current density from the momentum equation:
Solution of the equation (determines also the conductivity):
zc
peej
cji
w
ww
0
2
2E
E
1
12
2
0
w
w
ww
c
pe
cji
Insertion into the wave equation gives the dispersion relation:
w
w
www
w
c
pe
cckji
ck
1
1EE
2
2
2
22
02
22
Determination of the Dispersion Relation
64 U. Czarnetzki, ISLTPP, Bad Honnef 2016 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Solving for k2 yields:
Normalization by the electron plasma frequency helps
identifying the characteristics:
2
2 2 " " : , " " :
1
pe
c
ck R wave L wavew
ww
w
cpecpepe
c
kkcmit
k
wwwwwww
w
ww
/,/,/
1
122
Structure of the Dispersion Relation
65 U. Czarnetzki, ISLTPP, Bad Honnef 2016 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Cut-off frequencies (k = 0) are easily identified:
• Three intersections with the w-axis.
• R-wave at w = 0 and w = wR.
• L-wave at wL, in between the two cut-offs of the R-wave.
• Only the R-wave has a resonance (k .
• This resonance is at the cyclotron frequency w = wc.
• Natural result with view on the electron gyration.
cLR
ccLRc
wwww
wwww
w
ww
0,42
1
1
10 2
/
2
Cut-Offs and Resonances
66 U. Czarnetzki, ISLTPP, Bad Honnef 2016 U. Czarnetzki, ISLTPP, Bad Honnef 2016
• The dispersion relation is shown for wc = 2 wpe.
• Then the L-wave cut-off is below the R-wave resonance
(wc).
• In case wc < wpe, the L-wave cut-off is between wR and wc.
• The R-wave has a band-
gap between wc and wR.
• The L-wave has band-gap
between 0 und wL.
• For large k both waves
behave like waves in
vacuum with a phase
velocity c.
R- and L-Wave Dispersion
0
1
2
3
4
5
0 1 2 3 4 5
R wave
c k / wpe
w /
wp
e
L wave
R wave
wR
wL
wc
resonance, ECR
whistlers (helicon)
Faraday
rotation
67 U. Czarnetzki, ISLTPP, Bad Honnef 2016 U. Czarnetzki, ISLTPP, Bad Honnef 2016
• At low frequencies only R-waves exist (cut-off for L-waves).
• The dispersion relation can then be expanded to first order
(w << wc) which provides a simple expression for the phase
velocity:
pe
c
ph ck w
wwwv
• The phase velocity increases monotonically with w.
• Higher frequencies propagate faster and arrive earlier.
• If all frequency start at the same time a x = 0 and detection
is after a distance s, each frequency arrives at its own time:
R-Wave Dispersion at Low w and k:
Whistler Waves
2
0
1 1s
ph
dxt t
v tw
w
68 U. Czarnetzki, ISLTPP, Bad Honnef 2016 U. Czarnetzki, ISLTPP, Bad Honnef 2016
• Lightning flashes at the southern hemisphere excite whistler
waves within a broad spectral range (kHz).
• Waves propagate along the earth magnetic field lines.
• Due to dispersion, higher frequencies arrive earlier.
F1
F2
• A falling glissando can be heard (coining the term “whistler”).
• The time dependent frequency spectra scale like 1/t2.
• A fit “by eye” shows very good agreement (red lines).
Natural Whistlers
http://www-pw.physics.uiowa.edu/mcgreevy/ Stanford vfl group
69 U. Czarnetzki, ISLTPP, Bad Honnef 2016 U. Czarnetzki, ISLTPP, Bad Honnef 2016
• Helicon waves are whistler waves which are radially
bounded.
• This requires a finite kr and the dispersion relation changes.
• The difference is particularly pronounced at R < lz .
• For large radii the Helicon wave eventually converges
to the normal whistler wave.
2
2
2
2 2 2
1 Helicon
Whistler
Note:
z z
r rpe
c rz
r
z r
k k
k k
c kk
k
k k k
ww
w
0.0 0.5 1.0 1.5 2.0
0
1
2
3
4
5
Helicon wave
Whistler wave
w
wc (
wp
e/
c k
r)2
w
kz / k
r R / l
z
Helicon and Whistler Dispersion
Comparison:
70 U. Czarnetzki, ISLTPP, Bad Honnef 2016 U. Czarnetzki, ISLTPP, Bad Honnef 2016
• At high frequencies R- and L-waves both travel almost with
the speed of light in vacuum.
• However, still the R-wave is a little faser.
• Therefore, with distance L an increase phase difference j
develops. In second order in wpe/w this difference is:
nBc
LkL
pecRL 0
2
w
wwjjj
Transformation of a linearly polarized wave into a left and right
polarized wave: xyxyx eeieeie
2
Faraday Rotation
• The phase difference between the R- and L-components
rotates the direction of polarization.
• This effect is called Faraday rotation.
71 U. Czarnetzki, ISLTPP, Bad Honnef 2016 U. Czarnetzki, ISLTPP, Bad Honnef 2016
The R- and L- parts are recombined after the wave has
experienced a certain phase shift:
2/sin
2/cosE
11
2
EEE
2/
00
j
jjjjj RR iii
LR ei
ei
e
The initially only x-direction polarized wave has been rotated
by an angle -j/2 in direction of the y-axis.
Linear Polarization Rotation
j
z direction (wave propagation)
x,y direction (wave polarization)
72 U. Czarnetzki, ISLTPP, Bad Honnef 2016 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Applications:
• Measuring magnetic fields in astro physics.
• Determination of magnetic fields in fusion experiments.
• Building optical diodes for laser applications.
Application of Faraday Rotation
H.R. Koslowski, H. Soltwisch, W. Stodiek,
Plasma Physics and Controlled Fusion 38, 271 (1996)
B.M. Gaensler et al., Science 307m 1610 (2005)
Large Magellian Cloud
73 U. Czarnetzki, ISLTPP, Bad Honnef 2016 U. Czarnetzki, ISLTPP, Bad Honnef 2016
• ECR discharges take advantage of the resonance of the
R-wave at the electron cyclotron frequency.
• For commercial applications they are generally operated at
a microwave frequency of f = 2.45 GHz.
• This requires a magnetic field strength of B = 87.5 mT.
• The magnetic field is generally non-uniform.
• Waves are launched from the high field side (wc > w).
Electron Cyclotron Resonance (ECR)
Kouichi Ono, Mutumi Tuda, Hiroki Ootera, and Tatsuo Oomori,
Pure & Appl. Chem., Vol. 66, No. 6, pp. 1327-1334, 1994.
74 U. Czarnetzki, ISLTPP, Bad Honnef 2016 U. Czarnetzki, ISLTPP, Bad Honnef 2016
• Ion source for highly charged ions, e.g. at accelerators.
• High density plasma source in the semiconductor industry.
• Satellite propulsion.
THE SACLAY (FRANCE) HIGH-CURRENT
PROTON AND DEUTERON ECR SOURCE
Roth & Rau: commercial ECR source
z.B. for semiconductor treatment,
At P = 1 kW, n = a few 1017 m-3 .
Pressure: p = 0,1 – 10 Pa.
Examples of ECR Sources
75 U. Czarnetzki, ISLTPP, Bad Honnef 2016 U. Czarnetzki, ISLTPP, Bad Honnef 2016
POSTECH (Korea)
W. C. Lee, S. I. Moon, H. J. Jang, Y. S. Bae, M. H. Cho, W. Namkung (POSTECH)
* Work supported by KAERI’s KOMAC Project
Control of the Resonance Location
by the Magnetic Field
76 U. Czarnetzki, ISLTPP, Bad Honnef 2016
1 MW at f = 140 GHz
Figure shows a so called
“gyratron”, a high power
microwave source.
So far 10 gyratrons installed
at Wendelstein with a total
power of 10 MW.
ECR Heating (ECRH) at the New
STELLARATOR Wendelstein 7-X (W7-X)
77 U. Czarnetzki, ISLTPP, Bad Honnef 2016
ICRH (Ion Cyclotron Resonant
Heating) at the TOKAMAK
Joint European Torus (JET):
23 – 57 MHz, up to 32 MW
in 16 channels of 2 MW.
Ion Cyclotron Resonance Heating (ICRH)
78 U. Czarnetzki, ISLTPP, Bad Honnef 2016
The equation of motion for a single particle is solved:
Collisionless Heating in ECR Discharges
0 0 0, , 0mv e E v B v z z t v
• The time varying electric field E is assumed as
homogeneous, i.e.it is only a function of time
(corrections -> later).
• The static magnetic field B points only in the z-direction
and is only a function of z (magnetic mirror effects are
ignored).
• The aim is the calculation of the change in energy
between the initial moment and a later moment.
79 U. Czarnetzki, ISLTPP, Bad Honnef 2016
• The initial condition can always be met by adding the
solution for a free gyrating electron.
Free Gyrating Particle
0 0, , 0f h fmv ev B v z z t v
• This solution cancels out later when calculating the energy
change and averaging over all possible initial phases.
• Solution for a homogeneous B field:
0 0
0 0 0
0
sin cos
sin sin ,
cos
c
f c c
te B
v v tm
w j
w j w
• If B varies with z only the time
dependence changes.
• Note: Ions gyrate counter
clockwise and have a much
higher mass (lower frequency).
80 U. Czarnetzki, ISLTPP, Bad Honnef 2016
• The fields and initial conditions are now:
Homogeneous B-Field
0, , 0 0i t
z x y
e Be E E e i e e v t
m
ww
• The solution is:
0
cos
ˆ ˆsin , ,
0
e Ev v t v
m
j
j j j ww
• The velocity amplitude increases linearly with time!
• The energy increase quadratically:
22ˆ
2
mv j
• Therefore, very high energies can be reached.
81 U. Czarnetzki, ISLTPP, Bad Honnef 2016
• An inhomogeneous B field is now approximated by a
linear variation (first order expansion around the
resonance point):
Inhomogeneous B Field
1 z
e B ze
m Lw
• The solution is slightly more tricky but in the end effectively
the linear term j is replaced by an integral and some phase
shift appears:
0
cos / 4
ˆ ˆ/ 4 , ,
0
e Ev v cos t v
m
j j
j j j j ww
• For typical parameters, the integral converges to a constant:
• The integral reads:
' 20 0
00
' 2 ,2
i z
z
L v ze d
v L L
jj w
j j j j jw
2 10 1 6
0 10 10 10 /zL v m s m sw
82 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Since the velocity along z is constant, phase j
and z-position can be directly converted:
Energy Gain Along the z-Axis
0 0zz v z
L L Lj
w
Naturally, the width of the resonance zone is about L.
83 U. Czarnetzki, ISLTPP, Bad Honnef 2016
• The result for the energy gain is then:
Energy Gain in an Inhomogeneous B-Field
• The gain scales inversely with the initial velocity in
z-direction, i.e. it is proportional to the transit time
through the resonance zone.
• Remarkably, this result is independent of the initial velocity.
• The power per unit area is obtained by multiplying with
the flux density n v0z:
2
0
0 0
ˆˆ2 with 1,
2 z z
mv L L eEv
v v m
w w
w
2
0e E LnS
m
w
• Since the frequency is very high collisions do not play a
significant role up to a quite high pressure of about 100 Pa.
84 U. Czarnetzki, ISLTPP, Bad Honnef 2016
• A correct description of the heating process must take
the propagation and depletion of the wave into account.
Wave Equation Model
• The collisonless wave equation is (R - dispersion relation):
• A solution to this equation was found by K.G. Budden
in1966 but it is rather complex.
• Using again the linear B-field and constant plasma density
approximation and some normalization, the equation
reads:
222
0 021 0,
perr
c
Ek E k
z cz
w w
w w w
22
021 0, ,
perr
LEE zk
c
w
w
K.G. Budden, Radio Waves in the Ionosphere,
Cambrige University Press, UK 1966
85 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Reflection and Transmission
• One of the most interesting aspect about the solution is
that it allows also calculation of the reflection and
transmission coefficients. These results are quite simple:
• Quite unexpected, there is no reflection.
• Typical values are k0 = 50 m-1, L = 0.1 m which leads to
a moderate density condition:
2 2
021 , , 0,
ref pe peabs trans
inc inc inc
S LS Se e k L
S S S c
w w
w w
2
10 3
20.2 1.5 10
pen cm
w
w
86 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Tunneling Through the Band Gap
• The result of Budden implies that the wave is tunneling
through the band gap which exists between wc and wR.
• Since wc and z are interchangeable variables in the model,
it is illustrative to re-plot the dispersion relation in an
alternative representation.
• For low plasma densities
the band gap is small and
the wave can tunnel.
• Within the band gap, the
wave is evanescent.
band gap
87 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Further Aspects
• Doppler shift of the resonance:
z ck z v zw w
• This shifts the resonant magnetic field by typically a few
percent.
• Heating at higher harmonics of the cyclotron frequency.
• Parametric instabilities and non-linear power absorption.
• Confinement of electrons in a mirror field.
• Effect of the chamber geometry on resonances
(cavity effect at longer wavelengths or smaller chambers).
• Effects related to the transversal magnetic field.
88 U. Czarnetzki, ISLTPP, Bad Honnef 2016 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Microwave Discharges
89 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Some Advantages of Microwave Sources
• Simplicity of plasma generation at high (>100 W/cm3) and
low (< 1 W/cm3 ) powers.
• Wide range of operating pressures (from 10 Pa up to
atmospheric pressure).
• Control of the density profile by the antenna structure.
• Size of the discharge chamber can be variable.
• No electrodes.
• Large areas / volumes can be treated.
• High power sources at relatively low price available.
Y.A. Lebedev, Journal of Physics:
Conference Series 257, 012016 (2010)
90 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Operation Schemes
For most microwave sources the operation principle is either:
• Standing waves in a resonator.
• Propagation of surface waves (with or without resonator).
A. Schulz et al., Contrib. Plasma Phys. 52, 607 2012) Yi-Hsin Chou, et al. Dalton Trans., 2010,39, 6062-6066
Plasma torch:
91 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Surface Waves
• Dielectric (x > 0) - plasma (x < 0) interface.
• Wave is evanescent on both sides (exponential decay).
• Propagation along the surface (z – direction).
• Wave guide effect.
• Simplification: Homogeneous plasma.
x
y
z plasma
dielectric
Ex
Ez
By
2
0
, , exp exp exp
,
0 : 1 1
0 :
x z y m m
pe
p
D
E E B x i k z i t
m P D
P x ii
D x
w
w
w w w
• Transversal electric and magnetic fields (y – direction).
• Longitudinal (z) electric field due to waveguide effect!
k
92 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Surface Wave Parameters
• The boundary conditions couple the wave properties in the
two regions.
( ) ( )
( ) ( )
( ) ( )
1 1( )
( )
P D
y yP D
z z
p D
P D
y y
B Bi E E
x x
ii B B
2
P Dp D
P D
mm
P D
k k kc
c
w
w
Boundary condition (x = 0):
Solution (Note: The quantities are complex due to collisions!):
93 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Dispersion Relation (collisionless)
Simple case (no collisions):
2
2
2
2
2
2
1
1
1
1:
1
: 1
Resonance : :
1
1
pe
p
pe
peD peD
pe
D
pe
pe
r
r
pe D
c k
k
w
w
w
w
w w
w
w ww
ww w
w w
w
w
surface wave: w < wr
bulk wave
resonance bandgap
94 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Damping Evanescent Wave (collisonless)
• The decay of the wave in the plasma is always faster
than in the dielectric. (This can change in cylindrical
geometry.)
pe
pc
w
Dielectric Plasma
imaginary
(oscillating)
imaginary
(oscillating)
95 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Sources Types I
• Cavity and waveguide sources: Typically for flow-tube
reactors with a quartz or ceramic tube.
Y.A. Lebedev, Journal of Physics:
Conference Series 257, 012016 (2010)
96 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Sources Types II
• Surface wave discharges: Flow tube as well as large area
application.
Y.A. Lebedev, Journal of Physics:
Conference Series 257, 012016 (2010)
97 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Resonator and Surface Wave: Surfatron
http://sincaf.icmse.csic.es/showhighlight.php?posic=3
https://plasimo.phys.tue.nl/applications/surfatron/index.html
• Often used as atomic
radical source, e.g. H,
N, O etc. (and excited
molecules).
98 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Inverted Geometry: Duo Plasma Line
• Inner conductor, outer plasma.
• MW feeding from both ends to compensate for
damping of the wave along the line.
• Large homogeneous (along the line) plasmas.
• Usually operated with a number of lines in parallel.
W. Petasch, H. Räuchle, H. Mügge, K. Mügge,
Surface and Coatings Technology 93, 112 (1997)
A. Schulz et al., Contr. Plasma Phys. 52, 607 (2012)
99 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Sources Types III
• Large area and large volume source:
Many slots with individual surface waves at each slot.
• Planar or ring shaped geometry.
Y.A. Lebedev, Journal of Physics:
Conference Series 257, 012016 (2010)
100 U. Czarnetzki, ISLTPP, Bad Honnef 2016 U. Czarnetzki, ISLTPP, Bad Honnef 2016
antenna slot
coupling antenna
field of viewCCD-camera
monitor antenna
shorting plunger
quartz tube
magnetron
ring resonator
laser
circulator
The SLAN Source
Dirk Luggenhölscher, Dissertation RUB, 2004
101 U. Czarnetzki, ISLTPP, Bad Honnef 2016 U. Czarnetzki, ISLTPP, Bad Honnef 2016
-8 -6 -4 -2 0 2 4 6 80
1
2
3
4
2500 s Bessel Function
4000 s Bessel Function
030226b.o
pj
ne(1
01
0cm
-3)
distance to the center (cm)
Argon dominated (90 % Ar, 10 % H2): p = 30 Pa , f = 200 Hz, 20 % on
R
r41.2Jnrn 00
-8 -6 -4 -2 0 2 4 6 80
1
2
3
4
5 1 s 500 s
10 s 1000 s
50 s 2500 s
100 s 4000 s
030226b.o
pj
ne(1
01
1cm
-3)
distance to the center (cm)
Density Decay in the Afterglow
• Source terms only at the wall due to small skin depth (cm).
• Collisional electron cooling due to high pressure (cm).
• Diffusion of recombining plasma leads to hollow profiles.
• In the afterglow, relaxation to the basic diffusion mode.
Dirk Luggenhölscher, Dissertation RUB, 2004
102 U. Czarnetzki, ISLTPP, Bad Honnef 2016 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Some Conclusions
• At high frequencies the penetration of the electromagnetic
wave into the plasma is a major issue.
• ECR discharges allow the wave to travel long distances
before the energy is deposited within a relatively short range.
• In ICP and microwave discharges the skin depth is short
and power is deposited in the vicinity of the antenna.
• Particularly microwave discharge have a skin depth of
typically only one cm. However, power can be distributed
by operation of a large number of antenna slots. These are
often connected to one common resonator.
• Microwaves can propagate in the sheath between surfaces
and the plasma via surface waves.
• All discharges can produce dense plasmas in the range
of typically n = 1010 cm-3 to 1013 cm-3 and electron
temperatures of a couple of eV and much higher in ECRs.
103 U. Czarnetzki, ISLTPP, Bad Honnef 2016 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Literature M. A. Lieberman and A. J. Lichtenberg,
Principles of Plasma Discharges and Materials Processing
(Wiley, New York, 1994).
P. Chabert and N. Braithwaite,
Physics of Radio-Frequency Plasmas
(Cambridge University press, Cambridge, 2011).
S. Ichimaru
Basic principles of plasma physics: a statistical approach
(Benjamin, Reading, 1973).
Y. Raizer,
Gas Discharge Physics
(Springer, Berlin, 1991)
Rolf Wilhem,
ECR Plasma Sources
(Springer, Heidelberg 1993)
K.G. Budden,
Radio Waves in the Ionosphere,
(Cambrige University Press, Cambridge 1966)
104 U. Czarnetzki, ISLTPP, Bad Honnef 2016 U. Czarnetzki, ISLTPP, Bad Honnef 2016
Literature J. Reece Roth,
Industrial Plasma Ingeneering,
(Institute of Physics Publishing, Bristol, 1995)
Yu. Aliev, H. Schlüter and A. Shivarova,
Guided Wave Produced Plasmas,
(Springer, Heidelberg, 2000)
Yuri P. Raizer, Mikhail N. Shneider, and Nikolai A. Yatsenko,
Radio-frequency capacitive discharges,
(CRC Press, 1995)
Thomas Howard Stix
Waves in Plasmas
(American Institute of Physics, New York 1992)
R. Geller
Electron Cyclotron Resonance Ion Sources and ECR Plasmas
(Institute of Physics Publishing, Bristol 1996)
Francis F. Chen
Introduction to Plasma Physics
(Plenum Press, New York, 1984)
Akademische Jahresfeier der Fakultät für Physik & Astronomie 2013 12.07.2013| Seite 105
RUHR-UNIVERSITÄT BOCHUM
RUHR-UNIVERSITÄT BOCHUM
GEC 2016 69th Annual Gaseous Electronics Conference Ruhr-Universität Bochum October 10th-14th, 2016 www.gec2016.de