Emissive Probe Diagnostics in Low Temperature Plasma – Effect of Space Charge and Variations of Electron Saturation Current

Abstract

We report a study of electron saturation current variations at varying probe heating that were found to be closely related to probe wire contamination. The study was performed in three types of low temperature argon plasma – the weakly magnetized plasma of a cylindrical magnetron, the non-magnetized plasma of a double plasma machine and a hollow cathode plasma jet, showing different trends. In the present work the effect of overestimation of the plasma potential by a strongly emitting probe is discussed and experimental data are compared with a theoretical model. (© 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim)

Full text

Page 1

28th ICPIG, July 15-20, 2007, Prague, Czech Republi?
Emissive probe diagnostic in low temperature plasma – effect of the space
charge and variations of the electron saturation current
A. Marek1, M. Jílek1, I. Picková1, P. Kudrna1, M. Tichý1, R. Schrittwieser2, C. Ionita2,
P.C. Balan2
1 Charles University in Prague, V Holešovičkách 2, 18000 Prague, Czech Republic
2Institute for Ion Physics and Applied Physics, Leopold–Franzens University of Innsbruck, Technikerstr. 25,
A-6020 Innsbruck, Austria
Emissive probe diagnostic is the focus of this contribution. Two subjects were investigated:
a) applicability of the strongly emitting probe technique in low temperature plasma; b) variations
of the electron saturation current collected by the probe for varying probe heating. It was found
that electron saturation current variations were probably induced by probe surface cleaning for in-
creasing probe heating and probe contamination at decreasing probe heating in our type of plasma.
At strong heating the electron emission from probe wire plays also a role in electron saturation
current variations. The applicability of the strongly emitting probe technique in low temperature
plasma depends on the ratio of the temperatures of the emitted electrons and plasma electrons. Ex-
perimental data were compared with theoretical model.
1. Introduction
Emissive probes are used as suitable experimen-
tal tools for determining the plasma potential in
many types of plasmas ranging from hot isothermal
plasma [1] through low-temperature and low-pres-
sure applications [2] to dense plasma [3]. They can
be used also in presence of electron drifts or high
energy tails and even in non-neutral plasmas of
cathode sheaths [4]. Several techniques of the
plasma potential determination were established.
Their basic overview is given e.g. in [5]. Two tech-
niques for plasma potential determination by emis-
sive probes are frequently used – the strongly emit-
ting probe technique and the inflection point in the
limit of zero emission technique.
In the emissive probe characteristic the ther-
mionic emission of electrons from the probe appears
as an increase of the ion saturation current in the ion
accelerating region, since the emission current su-
perimposes on the ion saturation current. For in-
creasing emission from the probe the floating poten-
tial of the probe moves towards the plasma potential.
This fact is the basis of the plasma potential deter-
mination via a strongly emitting probe.
The electron saturation current is expected to re-
main unaffected by the emission of electrons from
the probe since the emitted electrons are attracted
back to the probe when the probe voltage becomes
positive with respect to the plasma potential. How-
ever, electron saturation current variations were ob-
served in various experiments.
2. Experimental set-up
Experiments were performed in the low tem-
perature weakly magnetized argon plasma of the cy-
lindrical magnetron devices [6] and [7] and in
the un-magnetized low temperature plasma of a
double plasma (DP) machine [8]. The cylindrical
magnetrons [6] and [7] differ substantially only in
the length of the discharge region and in the elec-
trode diameters. Each consists of two coaxially situ-
ated cylindrical electrodes. The outer discharge
electrode is grounded and serves as anode. The inner
water cooled electrode serves as cathode. The dis-
charge region is bordered by a pair of limiters held
on cathode potential. The shorter magnetron [6] has
a cathode with a diameter of 10 mm, an anode with a
diameter of 60 mm and a length of the discharge re-
gion of 120 mm. The longer magnetron [7] has a
cathode with a diameter of 18 mm, an anode with a
diameter of 58 mm and a length of the discharge re-
gion of 300 mm. In both magnetrons the plasma is
confined by homogenous magnetic fields parallel to
the axis. These are created by magnetic coils and can
vary up to 40 mT. The systems are evacuated by tur-
bomolecular vacuum pumps backed by oil free pis-
ton vacuum pump. The ultimate pressure in both
systems is on the order of 10–3 Pa. The discharges
in the magnetrons are usually produced in noble
gases at pressures of 1-10 Pa at typical discharge
currents of 100-400 mA.
Typical electron densities achieved in the mag-
netrons are on the order of 1016 m–3. In the positive
columns the electron temperature ranges typically
from one to a few electron volts depending on the
discharge conditions.
The DP-machine [8] consists of a vacuum cylin-
der of 44 cm diameter and 90 cm length. Heated
double filaments of 0,2 mm diameter tungsten wire
serve as hot cathodes for a low-temperature dis-
charge in argon. The inner side of the entire chamber
is covered by rows of strong permanent magnets
Topic number: 06
1606

Page 2

28th ICPIG, July 15-20, 2007, Prague, Czech Republi?
with opposite polarity to minimise the plasma-wall
interaction. The vacuum cylinder is evacuated by a
diffusion pump backed by a rotary pump. The ulti-
mate pressure in the system is on the order of
10–3 Pa. For discharge currents between 50 and 300
mA and argon pressures between 10–2 and 10–1 Pa,
the achievable plasma density lies in the range be-
tween 1015 and 1016 m–3.
The construction of the emissive probe used in
the experiments is shown in Fig. 1. The probe wire
was of tungsten or thoriated tungsten with a typical
diameter d = 150 µm. The electrical contact between
the probe wire and the feed line was realised by fine
copper wires that were tightly spliced around the
probe wire. This construction ensured excellent
electrical contact and relatively easy probe prepara-
tion. The probe was inserted into a double bored
degusit tube with 2,4 mm of major diameter.
3. Results
3.1. Strongly emitting probe technique in low
temperature plasma
The strongly emitting technique is used in many
types of plasma for measurements of the plasma po-
tential, since this technique is relatively simple and
provides a direct measurement of this parameter.
The applicability of this technique for low tem-
peratures depends on the ratio Te/TeW and on n,
where Te is temperature of electrons in the plasma,
TeW is the temperature of the emitted electrons and n
is the plasma density. This is illustrated in Fig. 2 for
the case of a tungsten probe and different values of
Te for a fixed plasma density. The data in Fig. 2 were
computed by applying floating probe conditions in a
1D model of a strongly emitting plane developed by
Takamura et al. [9]. It is visible that if Te and TeW are
comparable the plasma potential determined by the
strongly emitting probe can be markedly overesti-
mated. If TeW can be neglected with respect to Te (hot
plasma), the plasma potential determined by the
strongly emitting probe is underestimated by ap-
proximately one kTe/e as was shown e.g. in [10].
Model [9] was developed for the space charge
limited regime. The model potential used in [9] is
depicted in Fig. 3 together with the difference be-
tween the normalized potential of the virtual cathode
formed in front of the strongly emitting wall and the
plasma potential in dependence on Te/TeW.
Measurements with a strongly emitting probe
were performed in the low temperature plasma of the
shorter cylindrical magnetron at several radial posi-
tions and at two different pressures ensuring differ-
ent Te. The plasma potential determined by the
strongly emitting probe technique UplEM was com-
pared with the plasma potential determined by a
Langmuir probe Upl. The results of the experiment
are summarised in tables 1 and 2.
According to model [9] and Fig. 2, the plasma
potential determined by the strongly emitting probe
for p = 1,5 Pa should be more markedly underesti-
mated than in case of p = 6 Pa since the electron
temperature was higher in the latter case while the
plasma densities are comparable. However, this
trend is not visible in our data. It is probably because
probe wire
double-bore
degusit tube
bounded
copper wires
Fig. 1: Scheme of the emissive probe construction.
0,200,220,240,260,28 0,300,32
-1
0
1
2
3
4
5
6
Te=0.5eV
Te=3eV
Te=10eV
Ar, n = 4x10
16 m
-3, Ti=300 K
3695 K - melting point of tungsten
φW
TeW[eV]
Fig. 2: Results of 1D model [9] – difference between the
normalized potential (φ = eϕ/kTe) determined at the
strongly emitting tungsten wall and the plasma potential
for increasing TeW for three values of Te.
Fig. 3: Difference of normalized potential at the virtual
cathode and the plasma potential in dependence on the
temperature of emitted electrons TeW and schematics of
the 1D model of the sheath in front of a sufficiently emit-
ting probe [9].
1 101001000
-1.0
-0.9
-0.8
-0.7
-0.6
-0.5
-0.4
-0.3
-0.2
0
x
plasmasheath
wall
virtual
cathode
plasma ions
plasma electronsemitted electrons
α
φ=φV C
φ=φW
φ
β
φVC
Te/TeW
1607

Page 3

28th ICPIG, July 15-20, 2007, Prague, Czech Republi?
TeW and the difference in Te were not sufficient. In
both cases (tables 1 and 2), the plasma potential de-
termined by strongly emitting probe technique was
underestimated by approximately 0,3 kTe/e.
3.2 Variations of the electron saturation current
at varying probe heating
Thorough studies of the electron saturation cur-
rent variations at varying probe heating were per-
formed in cylindrical magnetron and DP-machine
plasma [11,12]. The magnitudes of the electron satu-
ration current variations were characterized, and it
turned out that the variations were more pronounced
in case of shorter probes. In [12] various possible
processes leading to variations were discussed. New
experimental data were recently obtained that yield a
better understanding of the phenomena. The profiles
of variations given in [11,12] were obtained from the
values of the emissive probe current-voltage char-
acteristics at fixed voltages. In this study not the
whole current-voltage characteristics was recorded
but a fixed probe bias was chosen and the probe
heating was varied. This approach enabled us to as-
sess better temperature induced processes.
An example of the electron saturation current
variation with varying probe heating is shown in
Fig. 4. The corresponding dependence of the float-
ing potential of the emissive probe on probe heating
is depicted in Fig. 5. The electron saturation current
(Iesat) in Fig. 4 first slightly increases with increasing
probe heating and then decreases by 40% for re-
peated increases of the probe heating. According to
Fig. 5 electron emission from the probe starts ap-
proximately for a heating current of IHEAT = 2,4 A.
The electron saturation current in Fig. 4 increases
with increasing electron emission for strong heating.
The initial increase of Iesat followed by a fast de-
crease is probably connected with probe surface
cleaning. The increase of Iesat in the region above
IHEAT = 2.4 A can be ascribed to the effect of ther-
mionically emitted electrons. Note that Iesat can
markedly change due to probe surface contamination
or cleaning, respectively, giving rise to errors in
Langmuir probe diagnostics [13]. Time dependence
of the emissive probe contamination after switching
off the heating in the cylindrical magnetron plasma
is depicted in Fig. 6.
Corresponding data measured by the same tech-
nique in the DP-machine plasma are shown in Fig. 7.
In this case the onset of emission occurred at IHEAT =
2 A. The temporal evolution of the probe contami-
nation is depicted in Fig. 8.
The difference between Figs. 4 and 7 is probably
caused by different processes of contamination in
the cylindrical magnetron and the DP-machine. The
Ar, 1.5 Pa, 20 m T, 50 m A
r [mm] Te [eV ] ne [m-3]
21 4.7
18 4.0
154.3
12 4.5
94.6
(UplE M-Upl)/kTe
7.00E+15
1.20E+16
1.80E+16
2.30E+16
2.30E+16
-0.26
-0.23
-0.33
-0.29
-0.30
Tab 1: Normalized difference between plasma potential
determined by strongly emitting probe and Langmuir
probe in cylindrical magnetron.
Ar, 6 Pa, 20 m T, 50 m A
r [mm] Te [eV ] ne [m-3]
212.1 8.00E+15
182.2 1.20E+16
15 2.3 1.60E+16
12 2.2 2.40E+16
9 2.6 2.80E+16
(UplEM-Upl)/kTe
-0.19
-0.32
-0.43
-0.50
-0.23
Tab 2: Normalized difference between plasma potential
determined by strongly emitting probe and Langmuir
probe in cylindrical magnetron.
-0.50.00.51.01.52.02.53.03.5
3.0
3.5
4.0
4.5
5.0
5.5
realization
%1
%2
W, d = 0.14 mm, l = 1 cm
Ar, 4 Pa, 20 mT, 200 mA
Electron saturation current [mA]
Heating current [A]
Fig. 4: Variation of the electron saturation current at
varying probe heating measured in the longer cylindrical
magnetron.
-12
W, d = 0.14 mm, l = 1 cm
Ar, 4 Pa, 20 mT, 200 mA
-0.50.00.51.01.52.02.53.03.5
-24
-22
-20
-18
-16
-14
Floating potential [V]
Heating current [A]
Fig. 5: Dependence of the floating potential on the emis-
sive probe heating in longer cylindrical magnetron.
1608

Page 4

28th ICPIG, July 15-20, 2007, Prague, Czech Republi?
cylindrical magnetron is more complicated because
stainless steel from cathode is sputtered during the
experiment and covers all surfaces. In comparison to
that the situation in DP-machine is simpler – the
probe is contaminated only by residual gases and/or
by polymerised film created in discharge from oil
evaporating from diffusion pump. Relevant discus-
sion on the probe contamination was done e.g. in
[14].
4. Conclusion
Strongly emitting probe technique is applicable
in low temperature plasma with Te ≅ 2-3 eV and
plasma density n ≅ 1016 m–3. Variations of the elec-
tron saturation current at varying probe heating can
mainly be ascribed to probe surface cleaning and
contamination in laboratory argon discharge plasma.
5. Acknowledgments
This work is part of the project MSM
0021620834 that is financed by the Ministry of Edu-
cation, Youth and Sports of the Czech Republic.
Thanks are also due to Czech Science Foundation,
grants No. 202/03/H162,
202/07/0044. Additional support by the CEEPUS
Network AT-0063 is gratefully acknowledged.
202/05/2242 and
6. References
[1] P. Balan et. al, Rev. Sci. Inst. 74 (2003) 1583.
[2] J. W. Bradley et al., Plasma Sources Sci. Technol
13 (2004) 189.
[3] S. Yan et al., Rev. Sci. Inst. 67 (1996) 4130.
[4] Y. Okuno et al., J. Appl. Phys. 70 (1991) 642.
[5] O. Auciello and D. L. Flamm, Plasma Diagnos-
tics, Academic Press, London (1989).
[6] M. Holík et al., Czech. Jour. Phys, Suppl D, 52
(2002) D673.
[7] E. Passoth et al., J. Phys. D 30 (1997) 1763.
[8] D. G. Dimitriu et. al, Acta Phys. Slovaca 54
(2004) 89.
[9] S. Takamura et. al, Contrib. Plasma Phys. 44
(2004) 126.
[10] G. D. Hobbs and J. A. Wesson, Plasma Phys. 9
(1967) 85.
[11] A. Marek et al., Czech. Jour. Phys., Suppl B 56
(2006) 932.
[12] A. Marek, Doctoral thesis, Charles University
in Prague (2007).
[13] C. Winkler et al. Plasma Phys. Control. Fusion
42 (2000) 217.
[14] J. R. Smith et al., Rev. Sci. Inst. 50 (1979) 210.
050100150 200250300
3.0
3.2
3.4
3.6
3.8
4.0
4.2
4.4
4.6
4.8
5.0
Ar, 4 Pa, 20 mT, 200 mA, 380 V
probe: W, d = 0.14 mm, l = 1 cm
2 min at 2.5 A
1 min at 2.8 A
HEATING OFF
%1
%2
%3
Electron saturation current [mA]
Time [s]
Fig. 6: Time dependence of the probe surface contamina-
tion in the longer cylindrical magnetron.
050100150
Time [s]
200 250300 350
0.4
0.6
0.8
1.0
HEATING OFF
1 min at 2.4 A
DP-Machine
Ar, 0.1 Pa, 200 mA, UK= -70 V
probe:
W, d = 0.125 mm, l = 10 mm
Contamination of probe - Ar, 0.1 Pa
Electron saturation current [mA]
Fig. 8: Time dependence of the probe surface contamina-
tion in the DP-machine plasma.
0.00.51.01.5 2.0 2.5
0.35
0.40
0.45
0.50
0.55
0.60
0.65
0.70
0.75
0.80
0.85
0.90
0.95
1.00
DP-Machine
Ar, 0.1Pa, 200mA, UK=-70V
Electron saturation current [mA]
Heating current [A]
W, d = 0.125 mm, l = 10 mm
%1
%2
%3
Fig. 7: Variation of the electron saturation current at
varying probe heating in DP-machine.
1609