IOP PUBLISHING PLASMA SOURCES SCIENCE AND TECHNOLOGY

Plasma Sources Sci. Technol. 21 (2012) 034002 (8pp) doi:10.1088/0963-0252/21/3/034002

Argon metastable dynamics in afilamentary jet micro-discharge atatmospheric pressureB Niermann, R Reuter, T Kuschel, J Benedikt, M Boke and J Winter

Ruhr-Universitat Bochum, Institute for Experimental Physics II, Universitatsstraße 150, 44780 Bochum,Germany

E-mail: benedikt.niermann@rub.de

Received 14 September 2011, in final form 9 December 2011Published 16 April 2012Online at stacks.iop.org/PSST/21/034002

AbstractSpace- and time-resolved measurements of Ar (3P2) metastable atoms at the exit of anatmospheric pressure radio-frequency micro-plasma jet are performed using tunable diodelaser absorption spectroscopy. The discharge features a coaxial geometry with a hollowcapillary as the inner electrode and a ceramic tube with a metal ring as the outer electrode.Absorption profiles of metastable atoms and optical emission measurements reveal thedynamics and the filamentary structure of the discharge. Single filaments with a diameter of125 µm and metastable densities of about 1013 cm−3 are observed. The average spatialdistribution of Ar metastables is characterized with and without a target in front of the jet,showing that the target potential and, therewith, the electric field distribution substantiallychange the filaments’ expansion. Together with a detailed analysis of the ignition phase andthe discharge’s behavior under pulsed operation, the results give an insight into the excitationand de-excitation mechanisms.

(Some figures may appear in colour only in the online journal)

1. Introduction

In recent years the research on atmospheric pressure micro-plasmas has become a strong focus in plasma science, mostlydue to their high potential for new plasma applications withoutthe need for expensive vacuum equipment. Among the largevariety of micro-plasma sources, which make use of dc, pulseddc and ac voltages ranging from mains frequency to RF, alarge number of configurations have been described for micro-discharges operating at atmospheric pressure. Examples aremicro-hollow cathode discharges, capillary plasma electrodedischarges, dielectric barrier discharges and micro-plasmajets [1, 2].

Understanding the energy transfer processes in micro-plasmas is one of the key requirements for developing newapplications and a reliable process control. In this contextmetastable species play a decisive role. The metastable densityin micro-discharges is several orders of magnitude lower thanthe density of the ground-state atoms. However, comparedwith most other species, the density is significant and the

electron collision excitation cross sections of some argon levelsout of the metastable states exhibit values which are severalorders of magnitude larger and have much lower thresholdsthan those for the ground state [3, 4]. Among the Ar metastablestates the Ar (3P2) level plays a decisive role, since argon isused as a feed gas in many micro-discharges.

Reliable techniques are needed for the systematicinvestigation of plasma characteristics and dynamics. Thisis essential for the optimization of plasma sources andprocess control. Application of conventional diagnostics tomicro-plasmas, especially invasive ones, is often impossibleregarding the small dimensions, high operating pressuresand high power densities. In this context absorptionspectroscopy is a widely used technique to measure, e.g.absolute concentrations of particles in plasma discharges, sinceit is non-invasive, highly sensitive and provides a sufficientspatial resolution [5, 6].

Among the variety of micro-plasma sources, radio-frequency (RF) driven atmospheric pressure plasma jetsprovide an effective platform for various plasma chemical and

0963-0252/12/034002+08$33.00 1 © 2012 IOP Publishing Ltd Printed in the UK & the USA

Plasma Sources Sci. Technol. 21 (2012) 034002 B Niermann et al

Figure 1. Scheme of the micro-plasma jet discharge with aphotograph of the plasma effluent in Ar/He atmosphere. Position Aindicates the position of the laser beam for measurements presentedin figures 3 and 4.

surface reactive processes [7–10]. Here we present the resultsfor a micro-plasma jet with a hollow capillary as the innerelectrode and a ceramic tube with a metal ring as the outerelectrode. This design has already been used for coating andsurface treatment applications [11, 12]. While this jet designshows large potential in these domains, there is still a lack ofunderstanding of the fundamental discharge dynamics. Weapplied tunable diode laser absorption spectroscopy (TDLAS)to record the spectral line profile of the lowest argon metastablestate (3P2) from the 1s5 →2p9 transition, deducing absolutedensities for various discharge conditions. In addition toTDLAS we used optical emission spectroscopy with hightemporal resolution to observe the discharge’s filamentarystructure.

2. The atmospheric pressure micro-plasma jet

Figure 1 shows a scheme of the micro-plasma jet. A detaileddescription of the setup has been published previously [13–15].A stainless steel capillary tube is inserted into a ceramic tubeleaving an annular gap of 250 µm between the tubes. Thecapillary ends 2 mm away from the end of the ceramic tube.Around the ceramic tube and 1 mm apart from its end, analuminum tube serves as the counter electrode. A 13.56 MHzRF power supply can be attached to the capillary or to theouter electrode through a matching network. Plasma-forminggas (He or Ar in this case) is introduced into the annular spacebetween the ceramic tube and the capillary with a flow rateof 3000 sccm (main flow). Additionally a flow of typically160 sccm (capillary flow) is guided through the capillary inorder to maintain similar gas velocities in the annular spacebetween the electrodes and in the capillary. This is especiallyimportant when some reactive gas is added to the capillaryflow. The plasma can be ignited easily in He by applying aroot mean square (RMS) voltage of about 100 V. The ignitionin Ar is not possible without an external high-voltage pulse and

therefore an Ar-plasma operation is realized through ignitionin He and switching gas flows to Ar. Nevertheless, it is possibleto operate the plasma pulsed, as long as the power-off time isless than 8 µs.

The whole jet setup is located in an airtight stainless steelvessel; its atmosphere was filled with a helium and argonmixture at a ratio of 5 : 3 after pumping it to a base pressureof 1 Pa. This gas mixture was chosen to correspond to theexperimental conditions used in another experiment for surfacetreatment [12].

To observe the ignition phase of the jet, the plasma wasoperated in pulsed mode, with a frequency of 1 kHz and a dutycycle of 99.4%. The short power off time is sufficient due tothe short metastable lifetimes on the order of some 100 ns. Asteady-state condition of the discharge is reached about 150 µsafter the ignition. The plasma extends up to 2 mm from the exitnozzle and appears homogeneous to the naked eye, but it hasa strong filamentary structure, similar to the results describedby Schafer et al [19], as will be shown in this paper.

3. Spectroscopic setup

Figure 2 shows a sketch of the spectroscopic setup. The laserbeam from the diode laser is guided through an optical isolator,to protect the diode from back-reflections, and passes throughtwo beam splitters to create three equivalent branches.

The first branch leads to a low pressure reference cell; inour case a hollow cathode discharge filled with argon. Thereference cell provides a longer absorption length and since itis running at low pressures, around a few tens of millibars, thelines are narrow and dominated by the Doppler width of theline profile. The second branch is guided to a confocal Fabry–Perot interferometer with 1 GHz free spectral range, necessaryfor the frequency calibration of the spectrum.

The part of the beam transmitted through the first beamsplitter is attenuated by neutral density filters with an opticaldensity of the order of 3, and focused into the discharge with abeam power of less than 2 µW and 100 µm spot size, to avoidany saturation effects. After passing through the discharge,the beam was guided through a set of apertures and filtersto suppress the emission from the plasma by reducing thecollection angle and blocking wavelengths different from theobserved transitions. The transmitted beam intensity wasmeasured by very fast low-noise photodiodes for highly time-resolved measurements. The laser beam was focused intothe discharge with a focal length of 4 cm. Assuming ideallenses and taking into account the aperture size of 2 mm andused wavelength of 811 nm, the spot size of the beam inthe focal point can be calculated to be about 30 µm. In thereal system, measurements show that the spot size is around100 µm. The time resolution of the system is about 40 ns. Foran effective measurement of the absorption signal across thejet axes, the discharge casing was mounted on a small movablestage featuring three electronically controlled stepping motorsto adjust the plasma jet position in all spatial dimensions withhigh precision. This setup allows the positioning of the jet withan accuracy of about 30 µm. All measurements presented are

2

Plasma Sources Sci. Technol. 21 (2012) 034002 B Niermann et al

Figure 2. TDLAS setup for absolute argon metastable density measurements.

line integrated on they-axis. The time-averaged measurementspresented in this paper are accumulated over 500 pulse cycles.

Absolute metastable densities were derived from thetransmittance I/I0 and the Beer–Lambert law. Therefore, foursignals were acquired:

L(ν)—plasma and laser on,L0(ν)—plasma off, laser on,P(ν)—plasma on, laser off,B(ν)—plasma and laser off (background),

to calculate the transmittance spectra and correlate them withthe plasma properties by

I (ν)

I0(ν)= L(ν) − P(ν)

L0(ν) − B(ν)= e−k(ν)·l , (1)

where I (ν) and I0(ν) are the intensities of the transmittedradiation with and without absorbing species, k(ν) is theabsorption coefficient and l the path length through theabsorbing medium [16]. The absorption coefficient isconnected to the population density of metastable atoms by

k(ν) = 1

4πε0· π(e2)

cme· fik · Ni · F(ν), (2)

where fik is the oscillator strength of the line, Ni the densityof the lower level, and F(ν) a normalized function (

∫ ∞0 F(ν) ·

dν = 1) representing the absorption line shape. All other termshave their usual definitions. The absolute metastable densitycan then be given by

∫ ∞

0ln

(I0(ν)

I (ν)

)dν = S = e2fikl

4ε0mec· Ni, (3)

where S is the area under the absorption curve that providesthe line of sight-averaged density of the absorbing species. Toderive absolute values for the metastable density, the spectralprofile of the lines has to be analyzed, and the pressurebroadening and the peak intensities have to be measured. Thearea of the absorption line profile is given by

S =∫ ∞

0ln

(I0(ν)

I (ν)

)dν = −

∫ ∞

0ln(1 − A(ν))dν, (4)

where A(ν) = (I0(ν) − I (ν))/I0(ν) is the absorption rate atfrequency ν and νc is the central frequency of the absorptionline. The absorption line shape can be described by aVoigt function, i.e. a convolution of Lorentzian and Gaussianfunctions. The width of the Gaussian component �νD, causedby Doppler broadening is described by

�νD = 2

λ0·√

2 ln(2)kbT

M, (5)

where T is the temperature of the absorbing species (assumedequal to the gas temperature here) and M the mass of theabsorbing species [17]. A more detailed insight into the densitycalculations can be found in [18].

Additionally, the light emission profile of the dischargeeffluent is recorded with an ICCD camera (Andor, iStarDH734-18F-03) in the visible spectral range. The experimentand the camera are synchronized with the internal gate monitorof the ICCD camera.

4. Results and discussion

4.1. Density approximations and experimental uncertainties

Figure 3 shows the time-averaged spectral profile of theabsorbing transition, measured about 200 µm behind thenozzle (position A in figure 1) under steady-state conditionsabout 150 µs after the plasma ignition. The line showsdominating Lorentzian broadening with a half-width of about8.8 GHz. The line is significantly shifted compared withits position under low pressure and low electron densityconditions. The shift is about 3.3 GHz to a longer wavelength,which is almost 2 GHz more than in α-mode glow dischargesat atmospheric pressure [18]. Due to high electron densitiesin the filaments, the shift must probably be attributed toStark effects. Since the discharge is strongly fluctuating,both in time and space, the spectral profile is expected notto be constant. The time-averaged profile shape was used tocalculate the absolute metastable densities, because accountingfor the real absorption profile with a temporal and spatial

3

Plasma Sources Sci. Technol. 21 (2012) 034002 B Niermann et al

Figure 3. Spectral absorption profile of the argon metastable1s5 →2p9 transition, showing strong Lorentzian broadening andsignificant red shift of the line.

resolution of nanoseconds and micrometers was not in therealm of possibility. The wavelength where the absorptionprofile has its maximum is used in the following for the space-and time-resolved measurements.

Although the absorption can be measured very accurately,the fluctuating and inhomogeneous character of the dischargeaffects the precise calculation of absolute densities fromthe time-averaged absorption data in this case, and makesevaluation of the errors difficult. Therefore, 2D maps of theline-integrated averaged absorption, which are the measurefor the probability of finding a plasma filament at a givenposition, will be presented. However, as we will describelater, the metastable density in one single filament and theaveraged metastable densities on the jet axis, calculated underthe assumption of a constant line width and line shift, will beestimated.

4.2. Non-averaged signals

Figure 4(a) shows a non-averaged time-resolved measurementof argon metastables in the first 160 µs after the ignition of thejet. It was taken on the jet axis, about 200 µm in front of thenozzle (position A in figure 1). The measurements reveal thefilamentary nature of the discharge by showing a strong peakin the spectrum whenever a filament crosses the laser beam.Outside the filaments the absorption is more than one order ofmagnitude lower. Taking this into account, we can concludethat only the high density in the filaments contributes to thespectral absorption profile shown before. Together with anestimation of the absorption length in one filament, which isgiven at the end of this section, the density inside a filamentcan be approximated to be at least on the order of 1013 cm−3.This is in agreement with the theoretical results obtained bySchafer et al [19].

Figure 4(b) shows a Fourier analysis of the temporaldensity evolution. In addition to the random noise, whichwould appear as a Brownian 1/f 2 spectrum, the frequencyspectrum shows two anomalies. One is the peak at 13.56 MHz

×

Figure 4. (a) Non-averaged (blue) and averaged (red) time-resolvedmeasurement of argon metastables in the first 160 µs after theignition of the jet, taken on the jet axis, about 200 µm in front of thenozzle. (b) Fourier analysis of the temporal density evolution,showing the metastable events in a strongly broadened peak around1 GHz.

caused by the RF frequency used to power the jet discharge.The other is a broad peak shaped structure, distorting thespectrum at frequencies around 1 MHz. This broad peakrepresents the characteristic frequency range of the filamentscrossing the laser beam. It is determined by a convolutionof several factors given by the spot size of the laser focus, riseand decay times of the metastable atoms and the lifetime of thefilament itself. The strong broadening of the frequency peak iscaused by a variety of factors: fluctuations in the gas mixture,the gas temperature, as well as quenching effects by highelectron and metastable densities inside the filaments stronglyinfluencing the species lifetime. The pile up of subsequentfilaments passing the point of measurement leads to a smearingof the frequencies to lower values.

Many factors influence the lifetime of the metastables.It is determined by the three-body collisions with impurities,superelastic collisions with electrons or pooling reactions.Based on our previous measurements of the lifetime of Armetastables in helium plasma and its dependence on impuritylevel, we expect that their lifetime is on the order of 100 ns [21].

To confirm the assumption of a filamentary discharge,fast optical emission measurements were used to record the

4

Plasma Sources Sci. Technol. 21 (2012) 034002 B Niermann et al

Figure 5. Emission profiles of the effluent region and the capillaryin the visible range, measured with an exposure time of 1 µs. Thered boxed area indicates the position of the ceramic capillary. Thelight emission from the surface of the ceramic tube is plasmaemission originating from the filaments inside the tube. Radiationgets scattered in all directions while passing through the ceramic.

emission profiles in the visible range with a high temporalresolution on the order of 1 ns. As indicated by the metastablemeasurements the movement of the filaments is slow, on theorder of 100 m s −1. Figure 5 shows two typical emission mapsof the effluent and the capillary in the visual range. The mapsare time integrated over 1 µs, still providing sufficient spatialconfinement due to the slow movement of the filaments. Theupper diagram shows the typical structure of a filament. Itoriginates from the inner powered metal electrode and extendsto a distance of about 2 mm from the nozzle. The course of thefilament is typically slightly curved, but it constantly tends tohead back to the central axis. This may be due to the electricfield distribution in the effluent or the fact that in the centralchannel the working gas is mostly unaffected by the ambientgas environment and the propagation of filament is easiest here.The diameter of the filaments is about 125 µm, which is in goodagreement with the values modeled by Schafer et al [19].

4.3. Spatial profiles in the continuous plasma

Figure 6 (top) shows the time-averaged 2D map of the line-integrated Ar metastable absorption behind the nozzle, whichreached the steady-state condition about 150 µs after theignition of the jet. The map shows significant absorptionfar outside the capillary, although the gas velocity on theexit of the nozzle is slightly below 100 m s −1, and lifetimesare on the order of a few hundred nanoseconds. This limitsthe decay length of this species to some tens of micrometerswithout further excitation. Taking this into account, thedensity distribution suggests significant electron densities andtemperatures at a distance of more than 1.5 mm from thenozzle, leading to high production rates of metastable atoms.This observation supports our previous suggestion, basedon the phase-resolved emission spectroscopy data, that thedischarge has a streamer-like character and is filamentary [9].

The absorption has, on average, a symmetrical shape andextends up to 2 mm from the jet nozzle. Comparing theseresults with the bended structure of the filaments shown in the

Figure 6. Top: averaged steady-state absorption map of the argonmetastable distribution, measured 150 µs after the ignition of the jetbehind the nozzle. Bottom: steady-state density distribution on thecentral axis.

last section, the course of the filaments explains the star-shapedexcitation profile in the effluent, which is shown in figure 6(top).

Figure 6 (bottom) gives absolute metastable densities onthe central axis from the nozzle of the jet. The densitiesare calculated by taking the full-width at half-maximum ofthe measured absorption profile as the absorption length fromthe spatial profile. The density values are not connected to thedensity inside the filaments but give a time-averaged densityapproximation for certain point in the effluent. The averageddensity on the axis decreases first up to a distance of about200 µm and then increases again. This is due to the curvedtrajectory of the filaments resulting in a hollow density profilewith lower density on the axis. This is in agreement withthe observed deposition profiles of hydrogenated carbon filmswith small admixtures of C2H2, which are hollow with a lowerdeposition rate on the jet axis [14].

4.4. Spatial profiles during the ignition phase

Since the absorption was measured with a time resolution ofabout 40 ns the ignition phase of the jet could be observed indetail. Figure 7 shows an area of 2.2 × 2.2 mm2 behind thenozzle at certain times after the ignition. The graphs showmajor steps in the temporal evolution of the 2D absorptionmaps. It should be stressed here that these maps represent justthe probability of finding a single filament at a given positionin front of the jet nozzle. They are averaged over 500 ignitioncycles. In the first 2 µs the filaments form in a channel ofroughly 1 mm diameter width at the tip of the nozzle. After3.5 µs a maximum in the metastable absorption is reached,

5

Plasma Sources Sci. Technol. 21 (2012) 034002 B Niermann et al

Figure 7. Absorption maps of argon metastable distributions for selected points in time, representing the ignition phase of the jet in the first20 µs.

Figure 8. Averaged metastable absorption in the first 160 µs afterthe ignition, measured on the central axis of the jet for variousdistances from the nozzle.

located about 200 µm in front of the nozzle. In the following7 µs the filaments expand over about 2 mm in the verticaldirection and produces significant metastable densities in up to2 mm horizontal distance. The discharge reaches a steady-statedistribution about 150 µs after the breakdown.

These changes in the temporal evolution may be aninteresting observation for applications such as surfacetreatment processes, where the materials are at some distancefrom the nozzle. If in the first breakdown phase of thedischarge a significantly higher averaged metastable densitycan be achieved far away from the nozzle, a high-frequencypulsing of the discharge may lead to improved results.

4.5. Excitation waves

While the temporal evolution of each cycle is highly chaotic,averaging the signal over about 500 cycles results in a stablesignal. Figure 8 shows the averaged metastable absorptionin the first 160 µs after the ignition, measured on the central

axis of the jet at various distances from the nozzle. Althoughevery new cycle shows an unpredictable sequence of filaments,there is a single commonality between each of them. In thefirst microseconds of the cycle a breakdown occurs close tothe nozzle and travels outward. As shown in figure 4(a) thefirst filament in the filament train is not stronger than thesubsequent ones, but since it has an almost fixed position in thetime domain, the signal sums up in the averaging process andgives the impression of an enhanced metastable production justafter the breakdown (see the averaged red absorption signalin figure 4(a)). Furthermore, the averaged signal shows astrong decrease after the first maximum, suggesting that theappearance of a filament in a certain time seems to lower theprobability of the appearance of an additional one shortly after.Measurements of the current and voltage waveforms takenon the jet electrodes during the ignition phase indicate thatthe observations described above cannot be explained by anovershoot of the power supply. Thus, the averaged temporalmetastable distribution gives valuable information about thestatistical distribution of the filaments in space and time.

Measuring the temporal profile in dependence of thedistance to the nozzle (figure 8) opens the possibility ofstudying the expansion of the filaments in the first fewmicroseconds after ignition. In this case the velocity wasmeasured on the central axis of the jet, resulting in velocitiesof more than 1500 m s −1 just behind the nozzle, and less than400 m s −1 at a distance of 1.6 mm. The speed of sound in argonand helium is about 300 m s −1 and 1000 m s −1, respectively,and can therefore not be responsible for the observed velocityprofiles. The expansion seems to follow an excitation waveas was described by Kong et al for a similar plasma jetdischarge [20].

4.6. Metastable distribution in front of targets

One way to significantly change the excitation profile ofmetastables in the effluent is to introduce a target at a

6

Plasma Sources Sci. Technol. 21 (2012) 034002 B Niermann et al

Figure 9. Top: steady-state absorption map of argon metastablesbetween the nozzle and a grounded target in 1.3 mm distance fromthe jet. Bottom: steady-state density distribution on the central axisbetween the nozzle and a grounded target at a distance of 1.3 mmfrom the jet.

short distance from the nozzle. These measurements are ofparticular interest as they highlight a connection to surfacetreatment experiments that have been performed with thisjet [11]. The target is made of a silicon wafer mounted ona metal plate. A critical element in this context is the electricalpotential of the metal plate directly behind the wafer. Whilean electrically floating plate leads to a density distributionthat is almost identical to the situation without a target, agrounded target significantly affects the characteristics of thedischarge. Figure 9 (top) shows the averaged steady-statespatial distribution between the nozzle and a target at a distanceof 1.3 mm. The spatial profiles differ substantially from themeasurements without a target and a floating target. Thedischarge shows a significantly more stable behavior and isnot fully governed by the filamentary character observed inthe free effluent. High metastable densities can be maintainedover the whole plasma column between the nozzle and thetarget. The plasma column is more confined with a morenarrow vertical profile, probably caused by the electric fielddistribution that pulls the filaments on the target electrode. Thestar-shaped profile directly behind that nozzle is lost. Thisactually corroborates that the star-shaped profile is causedby the presence of the outer electrode. The presence of thetarget has a much stronger effect and the star shape thereforedisappears. Close to the target’s surface the discharge spreadsover a diameter of more than 3 mm. Figure 9 (bottom) givesabsolute metastable densities on the central axis between thenozzle of the jet and the target. Densities are calculatedby taking the full-width at half-maximum of the measured

absorption profile as the absorption length from the spatialprofile.

Under grounded target conditions the plasma columncan be extended over a distance of about 3 mm, still havingaveraged metastable densities on the order of 5 × 1011 cm−3

on the target’s surface (not shown).

5. Summary

Space- and time-resolved absorption of Ar (3P2) metastableatoms in an atmospheric pressure radio-frequency micro-plasma jet were measured using tunable diode laser absorptionspectroscopy. Metastable profiles and time-resolved opticalemission measurements revealed the Ar metastable dynamicsand the discharge’s filamentary structure. Spectral lineprofiles were recorded, allowing absolute metastable densitycalculations and revealed high electron densities inside thefilaments. Additionally, Ar metastable densities on the orderof 1 × 1013 cm−3 were estimated. The lifetime and theshape of the filaments were characterized by fast ICCDcamera measurements. Highly time-resolved TDLAS recordsshowed the expansion of the filaments after the ignition witha precision of 40 ns. The spatial distribution of the filamentswas characterized with and without a target in front of thejet, showing that the target potential and, therewith, theelectric field distribution substantially changed the filaments’expansion. Together with a detailed analysis of the ignitionphase and the discharge’s behavior under pulsed operation,the results gave an insight into the excitation mechanisms.The results showed ways to enhance or reduce surface andvolume reactive processes, which is interesting for a varietyof applications such as the deposition of coatings or medicalpurposes.

Acknowledgments

This project is supported by the DFG (German ScienceFoundation) within the framework of the Research UnitFOR1123 (Projects A2, A4 and C1) and the ResearchDepartment ’Plasmas with Complex Interactions’ at Ruhr-University Bochum.

References

[1] Becker K H, Schoenbach K H and Eden J G 2006 J. Phys. D:Appl. Phys. 39 R55–70

[2] Iza F, Kim G J, Lee S M, Lee J M, Walsh J L, Zhang Y T andKong M G 2008 Plasma Process. Polym. 5 322

[3] Katsch H M, Quand E and Schneider T 1996 Plasma Phys.Control. Fusion 38 183

[4] Flohr R, Melzer A and Piel A 1993 Plasma Sources Sci.Technol. 3 206

[5] Miclea M, Kunze K, Heitmann U, Florek S, Franzke J andNiemax K 2005 J. Phys. D: Appl. Phys. 38 1709

[6] Tachibana K, Kishimoto Y and Sakai O 2005 J. Appl. Phys.97 123301

[7] Schutze A, Jeong J Y, Babayan S E, Park J, Selwyn G S andHicks R F 1998 IEEE Trans. Plasma Sci. 26 1685

[8] Schulz-von der Gathen V, Buck V, Gans T, Knake N, Niemi K,Reuter S, Schaper L and Winter J 2007 Contrib. PlasmaPhys. 47 510–19

7

Plasma Sources Sci. Technol. 21 (2012) 034002 B Niermann et al

[9] Benedikt J, Hofmann S, Knake N, Bottner H, Reuter R,von Keudell A and Schulz-von der Gathen V 2010 Eur.Phys. J. D 60 539

[10] Schafer J, Foest R, Quade A, Ohl A and Weltman K D 2008J. Phys. D: Appl. Phys. 41 194010

[11] Raballand V, Benedikt J, Hofmann S, Zimmermann M and vonKeudell A 2009 J. Appl. Phys. 105 083304

[12] Reuter R, Ellerweg D, von Keudell A and Benedikt J 2011Appl. Phys. Lett. 98 111502

[13] Yanguas-Gil A, Focke K, Benedikt J and von Keudell A 2007J. Appl. Phys. 101 103307

[14] Benedikt J, Focke K, Yanguas-Gil A and von Keudell A 2006Appl. Phys. Lett. 89 2006

[15] Benedikt J, Raballand V, Yanguas-Gil A andvon Keudell A 2007 Plasma Phys. Control. Fusion49 B4197

[16] Sadeghi N 2004 J. Plasma Fusion Res. 80 767[17] Demtroder W 1996 Laser Spectroscopy (Berlin: Springer)[18] Niermann B, Boke M, Sadeghi N and Winter J 2010 Eur. Phys.

J. D 60 489–95[19] Schafer J, Sigeneger F, Foest R, Loffhagen D and

Weltmann K D 2010 Eur. Phys. J. D 60 531[20] Walsh J L, Iza F, Janson N B, Law V J and Kong M G 2010

J. Phys. D: Appl. Phys. 43 075201[21] Niermann B, Kanitz A, Boke M and Winter J 2011 J. Phys. D:

Appl. Phys. 44 325201

8