Plasma chemistry fluctuations in a reactive arc plasma in the presence of magnetic fields

Article abstract of DOI:10.1063/1.1483128

The effect of a magnetic field on the plasma chemistry and pulse-to-pulse fluctuations of cathodic arc ion charge state distributions in a reactive environment were investigated. The plasma composition was measured by time-of-flight charge-to-mass spectrometry. The fluctuation of the concentrations of Al⁺, Al²⁺, and Al³⁺ was found to increase with an increasing magnetic field strength. We suggest that this is caused by magnetic field dependent fluctuations of the energy input into cathode spots as seen through fluctuations of the cathode potential. These results are qualitatively consistent with the model of partial local Saha equilibrium and are of fundamental importance for the evolution of the structure of films deposited by reactive cathodic arc deposition.

Full text of DOI:10.1063/1.1483128

Plasma chemistry fluctuations in a reactive arc plasma in the presence of magnetic
fields
Johanna Rosén, André Anders, and Jochen M. Schneider
Citation: Applied Physics Letters 80, 4109 (2002); doi: 10.1063/1.1483128
View online: http://dx.doi.org/10.1063/1.1483128
View Table of Contents: http://scitation.aip.org/content/aip/journal/apl/80/22?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
Characterization of plasma chemistry and ion energy in cathodic arc plasma from Ti-Si cathodes of different
compositions
J. Appl. Phys. 113, 163304 (2013); 10.1063/1.4802433
Departures from local thermodynamic equilibrium in cutting arc plasmas derived from electron and gas density
measurements using a two-wavelength quantitative Schlieren technique
J. Appl. Phys. 109, 063302 (2011); 10.1063/1.3552304
Measurement and modeling of a diamond deposition reactor: Hydrogen atom and electron number densities in
an Ar ∕ H 2 arc jet discharge
J. Appl. Phys. 97, 113306 (2005); 10.1063/1.1906288
Dense and strong plasma initiated by Ca 0.7 Sr 0.3 TiO 3 ceramic
Phys. Plasmas 11, 3715 (2004); 10.1063/1.1760773
Magnetic-field-dependent plasma composition of a pulsed aluminum arc in an oxygen ambient
Appl. Phys. Lett. 78, 150 (2001); 10.1063/1.1339847
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.252.67.66 On: Tue, 23 Dec 2014 06:49:59

APPLIED PHYSICS LETTERS
VOLUME 80, NUMBER 22
3 JUNE 2002
Plasma chemistry fluctuations in a reactive arc plasma in the presence
of magnetic fields
a)
´
Johanna Rosen
¨
¨
Department of Physics, Linkoping University, Linkoping SE-58183, Sweden
´
Andre Anders
Lawrence Berkeley National Laboratory, Berkeley, California 94720
Jochen M. Schneider
Materials Chemistry, RWTH-Aachen, D-52056 Aachen, Germany
͑Received 4 February 2002; accepted for publication 8 April 2002͒
The effect of a magnetic field on the plasma chemistry and pulse-to-pulse fluctuations of cathodic
arc ion charge state distributions in a reactive environment were investigated. The plasma
composition was measured by time-of-flight charge-to-mass spectrometry. The fluctuation of the
concentrations of Alϩ, Al2ϩ, and Al3ϩ was found to increase with an increasing magnetic field
strength. We suggest that this is caused by magnetic field dependent fluctuations of the energy input
into cathode spots as seen through fluctuations of the cathode potential. These results are
qualitatively consistent with the model of partial local Saha equilibrium and are of fundamental
importance for the evolution of the structure of films deposited by reactive cathodic arc deposition.
© 2002 American Institute of Physics. ͓DOI: 10.1063/1.1483128͔
Vacuum arc plasma is produced in micrometer-sized
cathode spots of very high current, power, and plasma den-
reactive arc plasma in the presence of magnetic fields,5 how-
ever, without investigating the statistics of the large fluctua-
tions inherent to cathodic arc plasmas. Ions are extracted
from the plasma source, forming a beam. A set of annular
electrostatic TOF gates is used to select a 200 ns sample of
the beam, which is detected by a magnetically suppressed
Faraday cup located 1.03 m from the TOF gate. The plasma
composition can be calculated from the measured ion
current–time dependence.
sity ͑ϳ1012 A/m2, ϳ1013 W/m2, and ϳ1026 mϪ3
,
respectively͒.1 The high power density is associated with ex-
plosive phase transitions, resulting in a plasma that is highly
ionized and which can be used for thin film synthesis, ion
implantation, and ion injection into accelerators. The indi-
vidual ion charge states (Qi) are of importance in plasma
processing because the kinetic energy gain (⌬Ei) across the
potential difference between plasma and substrate (⌬U) is
given by ⌬EiϭQi ⌬U. It is well known that the kinetic
energy, and therefore the charge states, affect the microstruc-
ture evolution and hence the film properties.2
The plasma was generated from an aluminum cathode.
The arcs were triggered by a high voltage flashover across a
ceramic tube between a trigger electrode and the cathode.
The arc current was 400 A and the arc pulse duration was
250 ␮s at a repetition rate of 1 Hz. All data were taken at 150
␮s after the arc was ignited at an oxygen partial pressure of
9ϫ10Ϫ2 Pa. Prior to introduction of the oxygen the system
base pressure was about 9ϫ10Ϫ5 Pa. A magnetic field could
be induced in the immediate vicinity of the cathode by dis-
charging a 250 ␮F capacitor bank via a high-current thyristor
switch and solenoid coil. The coil was 30 mm in length and
20 mm in diameter. The magnetic field at the cathode surface
was up to 0.28 T with coil or 0 T when the coil was not used.
From 100 individual time-of-flight spectra the ion concentra-
tions for Alϩ, Al2ϩ, and Al3ϩ ions were calculated. From
the statistical analysis of these data one can learn about
plasma chemistry fluctuations with a resolution of 200 ns.
The relative standard deviation (Srel) of the individual alu-
minum ions is shown versus the magnetic field strength in
Fig. 1. The magnetic field strength strongly affects the pulse-
to-pulse fluctuations (Srel): Without magnetic field, Srel for
Alϩ, Al2ϩ, and Al3ϩ equals 4%, 7%, and 26%, respectively.
As the magnetic field is increased up to a maximum of B
ϭ0.28 T, Srel increases drastically to values of up to 46%,
66%, and 91%, respectively.
Experimental as well as theoretical data describing the
plasma composition and the resulting average ion charge
states of vacuum arc plasmas are available for more than 50
cathode materials.3 The effect of both the reactive gas pres-
sure as well as the presence of a magnetic field on the pulse-
to-pulse fluctuations of the ion species distribution in the
plasma have not been investigated, despite the fact that thin
film compound synthesis by cathodic arc is usually achieved
in the presence of reactive gas as well as magnetic fields.
We used pulsed vacuum arcs to study the plasma chem-
istry and ion charge state fluctuations in an oxygen back-
ground with a varying magnetic field. We found that increas-
ing the magnetic field strength results in an extensive
increase in plasma chemistry fluctuations. This may have
important consequences for the microstructure evolution dur-
ing thin film growth.
The analysis of the plasma composition was carried out
using a time-of-flight ͑TOF͒ charge-to-mass spectrometer at
Berkeley Lab, described in detail in Ref. 4. This setup has
been used in the past to determine the plasma chemistry of
a͒
Electronic mail: johro@ifm.liu.se
We have also measured the arc burning voltage, i.e., the
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
0003-6951/2002/80(22)/4109/3/$19.00 4109 © 2002 American Institute of Physics
128.252.67.66 On: Tue, 23 Dec 2014 06:49:59

´
4110
Appl. Phys. Lett., Vol. 80, No. 22, 3 June 2002
Rosen, Anders, and Schneider
by a set of Saha equations ͑Saha equilibrium͒. As the plasma
expands, the rate of ionizing and recombining collisions de-
creases due to decreasing plasma density. The plasma
reaches a temperature and density dependent transition zone,
characterized by ionization equilibrium between ions of
neighboring charge states that are separated by relatively
small ionization energy, while the ionization and recombina-
tion rates of ions with high ionization energy are not equal
due to insufficient frequency of inelastic collisions. The ion
charge state ratios freeze successively in the expanding
plasma, where the ionization transition with the largest ion-
ization energy freezes first. After the plasma has gone
through the PLSE transition zone, the charge state distribu-
tion ͑CSD͒ remains practically constant when the plasma is
further expanding ͑‘‘frozen distribution’’͒. The PLSE model
is verified for vacuum arc discharges,9,10 and later we show
that our findings for the metal species of arcs in reactive
gases are qualitatively consistent with plasma expansion de-
scribed by the PLSE freezing model.
Fluctuations of the burning voltage represent fluctuations
of the power invested in the plasma, and therefore the plasma
parameters such as the electron temperature exhibit fluctua-
tions as well. The ratios of ion concentrations are coupled to
plasma density and electron temperature. Therefore, fluctua-
tions of electron temperature must be associated with fluc-
tuations of the freezing conditions in the transition zone and
the resulting charge state distribution. A magnetic field in-
creases the range of fluctuations ͑Fig. 1͒. The reason for the
increase of fluctuations is not precisely known but it is rea-
sonable to consider that the step distance between explosion
centers is increased,6,7 thereby ion assistance in spot forma-
tion is reduced. This would cause the impedance to go
through larger amplitude variations. Impedance fluctuations
are more likely to be amplified in magnetized plasmas than
in nonmagnetized plasmas.11 In addition to causing larger
fluctuations, the magnetic field also causes electron impact
ionization of the oxygen ambient5 and residual gas5,12 in the
vacuum chamber.
FIG. 1. Relative standard deviation of aluminum ion concentrations vs mag-
netic field, each data point is based on 100 plasma pulses.
cathode potential with respect to the grounded anode. It is
known6,7 that almost all of this voltage drops in the very thin
cathode fall, thus the burning voltage is directly proportional
to the power available for cathode spot operation. We again
performed a statistical analysis of 100 pulses per parameter
set. We found that an increase in magnetic field strength
causes an increase in ͑a͒ the average burning voltage as well
as ͑b͒ the fluctuations of the burning voltage, which are
shown in Fig. 2. Since the arc current is kept constant this
implies that ͑a͒ the power invested in the cathode surface is
increasing with increasing B-field and ͑b͒ the fluctuation of
the power is also increased with increasing B. These fluctua-
tions are due to the well-known nonstationary character of
cathodic arc spots,6,7 i.e., the rapid formation and extinction
of plasma production centers.
As the power invested in the cathode spot increases, the
electron temperature increases,8 which in turn affects the
plasma composition. The correlation of electron temperature
and plasma composition for an equilibrium plasma is given
by the Saha equations.9 The model of partial local Saha
equilibrium10 ͑PLSE͒ implies that the expanding plasma can
be divided in distinct zones. In the first zone very close to the
cathode spot, the plasma density is high enough to provide
for sufficiently frequent collisions between electrons, atoms,
and ions to result in ionization equilibrium, and thus the
ratios between differently charged particles can be described
The plasma composition is important since it is strongly
correlated to the ion energy of the impinging ions during film
growth. An expression for the kinetic energy of an ion is
given by EiϭE0ϩQe⌬U, where E0 is the initial kinetic
energy of the ion gained at the cathode spot, Q is the charge
state of the ion, and ⌬U is the potential difference in the
sheath in front of the substrate. Through ⌬U, the ion energy
can be controlled by selecting the substrate bias; however,
the fluctuations discussed here about the plasma chemistry
result in an extensive variation in ion energy. For example, if
⌬Uϭ500 V and the average charge state Q ϭ2 with a
͗ ͘
variation of Ϯ10%, the average kinetic ion energy will ac-
cordingly have a variation of Ϯ100 eV. Therefore, the fluc-
tuation of the CSD will affect the resulting microstructure of
the film.
In summary, an increasing magnetic field causes an in-
crease in the pulse-to-pulse fluctuations of aluminum ion
concentrations in a reactive environment. This is consistent
with the model of PLSE, suggesting that the fluctuations in
plasma chemistry are caused by fluctuations of the plasma
parameters in the cathode spot and transition zone to non-
FIG. 2. Average burning voltage vs magnetic field, the bars represent stan-
dard deviation based on 100 plasma pulses per measuring point.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
equilibrium. The plasma fluctuations in turn are the result of
128.252.67.66 On: Tue, 23 Dec 2014 06:49:59

´
Appl. Phys. Lett., Vol. 80, No. 22, 3 June 2002
Rosen, Anders, and Schneider
4111
a fluctuating power density due to a fluctuating burning volt-
age. The resulting CSD may have a profound influence on
the properties of the film since it is strongly connected to the
ion energy, which is well-known to affect the microstructure
evolution.
2 J. E. Greene, S. A. Barnett, J.-E. Sundgren, and A. Rocket, Ion Beam
Assisted Thin Film Growth, edited by Itoh ͑Elsevier, Amsterdam, 1989͒,
Chap. 5.
3 I. G. Brown, Rev. Sci. Instrum. 65, 3061 ͑1994͒.
4 I. G. Brown, J. E. Galvin, R. A. MacGill, and R. T. Wright, Rev. Sci.
Instrum. 58, 1589 ͑1987͒.
5 J. M. Schneider, A. Anders, and G. Y. Yushkov, Appl. Phys. Lett. 78, 150
͑2001͒.
Support from the Foundation for Strategic Research
͑SSF͒ Materials Research Consortium on Thin Film Growth
and from the U.S. Department of Energy ͑Contract No. DE-
AC03-76SF00098͒ is acknowledged. One of the authors
͑J.M.S.͒ acknowledges sponsorship of the Alexander von
Humboldt Foundation, the Federal Ministry of Education
and Research, and the Program for Investment in the Future.
6 G. A. Mesyats, Cathode Phenomena in a Vacuum Discharge: The Break-
down, the Spark and the Arc ͑Nauka, Moscow, 2000͒.
7 R. L. Boxman, D. M. Sanders, and P. J. Martin, Handbook of Vacuum Arc
Science and Technology ͑Noyes, Park Ridge, 1995͒.
8 A. Anders, B. Yotsombat, and R. Binder, J. Appl. Phys. 89, 7764 ͑2001͒.
9 A. Anders, Phys. Rev. E 55, 969 ͑1997͒.
10 A. Anders, IEEE Trans. Plasma Sci. 27, 1060 ͑1999͒.
11 R. J. Goldston and P. H. Rutherford, Introduction to Plasma Physics ͑IoP,
Bristol, 1997͒.
1
12
¨
¨
¨
¨
¨
A. Anders, S. Anders, B. Juttner, W. Botticher, H. Luck, and G. Schroder,
J. M. Schneider, A. Anders, B. Hjorvarsson, and L. Hultman, Appl. Phys.
IEEE Trans. Plasma Sci. 20, 466 ͑1992͒.
Lett. 76, 1531 ͑2000͒.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.252.67.66 On: Tue, 23 Dec 2014 06:49:59