Pulsed Plasma Thrusters for Space Propulsion

and Industrial Processing

Henri P. Wagner, Monika Auweter-Kurtz

Institut fur Raumfahrtsysteme

Universitat Stuttgart

Abstract

Pulsed plasma thrusters (PPT) are of particular interest for space exploration and exploita-tion since they essentially combine the advantage of continuously operated high-power magneto-plasmadynamic (MPD) thrusters with low average electric power consumption and manageableheat generation. This available-power capacity together with their ability of delivering small im-pulse bits for orbit and attitude maintenance makes them especially well suited for small satellitesand/or constellations of satellites where precise orientation-keeping is mandatory for mission suc-cess. The principle of these short-pulsed plasma accelerators can also be used to produce fast heatand particle bursts for industrial materials processing and surface modication. Many requirementsare the same for both space propulsion and industrial processing, even if the accelerators may dierin size and power. In a new line of research and development at the Institut fur Raumfahrtsysteme(IRS) of the University of Stuttgart, a large pulsed 20-GW coaxial plasma accelerator for industrialprocessing has been designed and is in the process of being manufactured and assembled. In paral-lel, a smaller PPT for orbital manoeuvres has also been designed and will be installed on a thruststand. The pulsed plasma activities at the IRS will be described and the rst results obtained willbe presented and discussed.

1 Introduction

Pulsed plasma thrusters (PPT) are of particular interest for space exploration and exploitation sincethey essentially combine the advantage of continuously operated high-power magneto-plasmadynamic(MPD) thrusters, i.e. reduced propellant mass through high specic impulse, with variably low averageelectric power consumption and manageable heat generation. This available-power capacity togetherwith their ability of delivering small impulse bits for orbit and attitude maintenance makes themespecially well suited for small satellites and / or constellations of satellites where precise orientation-keeping is mandatory for mission success.The rst theoretical and experimental investigations of short-pulsed magneto-plasmadynamic (MPD)accelerators were done in 1960 [1] and a rst optimization in view of possible space missions in1962 [2]. The rst recorded operation of PPTs (and the rst-ever operation of MPD thrusters) inspace took place in 1964 when 6 units were successfully red on the ZOND-2 probe. In the years upto 1988 PPTs demonstrated their operational reliability and compatibility with satellite electronicsand communications [3][7]. The most recent missions and projects with pulsed plasma thrusters areEARTH-OBSERVING-1 (launched in 2001) as well as MIGTHYSAT-II [8] and DAWGSTAR [9]. Withstored energies of less than 100 J per pulse and thruster, these engines were essentially designed forsecondary propulsion, i.e. station-keeping, attitude control and drag compensation for small to mediumsatellites. Nonetheless there are also eorts made to investigate and optimize high-energy / high-powerdevices for primary propulsion [10] and even interplanetary (manned Mars) missions [11].The principle of these pulsed accelerators can also be used to build plasma generators for eÆcient andenvironmentally conscious industrial processing. Many surface treatment applications require a shortburst of incident energy. Then in its simplest approximation, an accelerated plasma acts as a largearea, intense, pulsed heat or particle source which can be used to modify the chemistry or morphology

Associate Fellow AIAA

1

of a thin surface layer of the treated material. Typical power uxes around 100 MW/cm2 and pulsewidths of less than 100 s, resulting in moderate energy loads around 10 J/cm2 can easily be achieved.The working propellant can be an inert gas such as Helium or Argon, or it can be chemically reactiveif specic reactions are desired on the target surface, in addition to the pure thermal / kinetic eects.The high particle uxes on the target make such a pulsed accelerator then well suited for thin layerdeposition [12][18], biological or radioactive decontamination, waste reduction or plasma etching.Given their capabilities of delivering intense heat pulses as well as bursts of highly energetic ionsover large target areas, pulsed plasma accelerators are expected to play a major role in the eld ofsurface engineering. It has been shown that thermal and / or shock hardening in association with ionimplantation leads to increased surface hardness and improved tribological and erosion characteristics.Furthermore, these accelerators can also be used to obtain high non-equilibrium atomic concentrationsduring surface alloying processes [19][23].For both space propulsion and materials processing the advantages of the pulsed plasma technologyare the considerable knowledge on magnetodynamic plasma acceleration which was accumulated sinceseveral decades (among others in fusion experiments and weapons research), the simplicity and robust-ness of their electrical and mechanical design as well as the use of inexpensive and reliable technology.Moreover, the most important requirements for successful operation and commercialization, i.e. life-time, eÆciency, reliability and reproducibility of the results, are the same for both applications, evenif the accelerators may dier in size and power. In 2001, a new line of research and development wasinitiated at the Institut fur Raumfahrtsysteme (IRS) of the University of Stuttgart in order to addressthese issues while making use of the synergies.

2 Pulsed plasma activities at IRS

Based upon the experience gained by one of the authors during his stay at the Plasma Physics GroupP24 of Los Alamos National Laboratory (LANL), working on short-pulsed coaxial plasma accelerators[24][30], contact was made with industry in 2000 in order to discuss possible industrial applications forthis technology. Among others, one process came out to be of particular importance: The generation ofresidual compressive stresses in supercial layers of thin-walled engine parts subject to high dynamicloads and, hence, prone to fatigue-induced failure.

Fig. 1: Jet engine turbine blade

For dynamically loaded machine parts like jet engine com-pressor and turbine blades (Fig. 1), damage tolerance andfatigue strength are the determining factors for cost andweight reduction. This type of failure happens under a loadwell below the static yield limit and is due to the steadyprogression of cracks across the part's section. These cracksstart at the part's surface where the bending and twistingstresses are usually highest and where, in addition, surfaceroughness leads to a local stress amplication. The surfaceroughness is determined by the production process and,especially, the damages caused by corrosion or mechanicalabrasion (debris, grazing), even during normal operation.The sensitive areas of the blade are its foot, shoulder andleading edge as well as its upper edge.

Hence, the only practical way of extending a blade's operational life is to improve the fatigue resistanceof its material. This can be done with a high enough pressure burst acting on the endangered area(peening) and, thus, creating a supercial layer with residual compressive stresses which impede thegrowth of cracks. It was demonstrated that the operational life of damaged but peened blades couldbe even higher than that of undamaged and unpeened ones [31].In a rst method to achieve that surface state, conventional shot peening, a jet of fast and hardparticles creates the residual stresses directly upon impact through plastic deformation of the targetsurface. However, this tends to change the surface texture in a way which makes it diÆcult to detectthe appearance of new cracks during regular engine maintenance. There also exists a second method

2

using a pulsed high-power laser to explosively evaporate the material of a sacricial layer previouslydeposed onto the part's surface and, thus, sending a compressive shock wave into the substrate. This isan indirect process since it uses a fast thermodynamic phase change to create the plastic deformation.In order to generate the power densities required for a fast evaporation, the laser beam has to befocused down to a small area. This means that despite a potentially high pulse repetition capability ofthe laser, the surface throughput may be uneconomically low, in addition to the low energy conversioneÆciencies of pulsed lasers (usually less than 1%).A third method, the Plasma Impulse Peening (PIP) uses a coaxial plasma accelerator to create abeam of highly directed energy and is expected to help avoid these disadvantages. On one hand itresembles the indirect laser peening for it also creates a shock wave by evaporating a sacricial layerand, on the other hand, it resembles the direct shot peening because it uses a pulsed ow of highlyenergetic (atomic) particles which themselves create a pressure upon impact with the target material.Plasma impulse peening can be an economically viable process because of its potentially large beamproles resulting in large single-shot-treated areas and its high energetic eÆciency. Hence, one pulsedplasma activity at IRS is focused on the design and experimental investigation of a compact, eÆcientand self-eld coaxial plasma accelerator for plasma impulse peening demonstration and other possibleindustrial applications.With the PIP activity well in progress and its potential synergies with electric space propulsion, theresearch eld was then enlarged to include the design and development of pulsed plasma thrusters.Moreover, with the administrative reorganization of the Faculty of Aviation and Aerospace Engineeringand Geodesy concluded, it has become a declared goal of IRS to develop and launch its own series ofself-propelled micro-satellites.

Fig. 2: DLR-Tubsat Micro-Satellite

Given the low levels of available electric power, one typeof thrusters likely to be installed on these small satellitesand able to fulll the mission requirements are PPTs.Thus, the other pulsed plasma activity at IRS will focus onthe design, development and experimental investigation ofPPTs for small satellite propulsion and, in medium-term,their qualifying operation in space. The IRS satellite isassumed to be similar to the DLR-Tubsat [32] (Fig. 2) ofthe University of Berlin, roughly a 30 30 30 cm cubewith a mass of 35 kg and an average of 14 W availablepower. In a rst step, a PPT for drag compensation of thissatellite and raising of its circular orbit from 350 km to1000 km altitude is to be be designed and investigated on athrust stand. This propulsion system base model will thenbe varied in order to improve its eÆciency.

3 Pulsed plasma accelerator models

The coaxial plasma accelerator is a relatively simple device capable of generating, accelerating andguiding energetic plasmas. In the simplest form, a high electric current is discharged across two coaxialelectrodes and through a working gas. The Lorentz force, caused by the interaction of the dischargecurrent with its own magnetic induction eld, drives the plasma down the barrel and out the muzzle ofthe accelerator. External magnetic eld coils can be used to create a guide eld that forms a magneticnozzle. A conceptualization of such an accelerator with its magnetic guidance is shown in Fig. 3. Atthe present time however, there appears to be no immediate need for an externally applied magneticnozzle eld.For short-pulsed plasma acceleration, the discharge time is always comparable to the traversal time ofthe plasma and a critical element of the operation. The discharge type is determined by the electrody-namic and magnetodynamic properties of the electrical driving circuit and the plasma accelerator, i.e.its geometry and propellant distribution. The latter also determines in which one of the two basicallydierent modes the plasma gun is operating.

3

arc discharge

directed plasma ow

arc discharge

magneticnozzle eld

Fig. 3: Conceptualization of a coaxial plasma accelerator system with magnetic guide eld coils

1. If an initial propellant ll with a given density 0 is established in the space between the electrodeswhen the voltage is applied, then a shock is formed in the accelerator on top of a plasma sheetwith a thin, highly conductive, current carrying layer at its base. On these time scales, theinduced magnetic eld B cannot diuse through that thin layer which then acts as a movingsolid piston. The piston runs parallel to the electrodes, simoultaneously generating, acceleratingand heating the plasma, while it is swept up by the shock as in a `snowplow' [33].

2. If an initial propellant quantity with a given mass m is introduced between the electrodes inthe vicinity of the current feeds when the voltage is applied, then a plasma sheet is formedin the accelerator with a thin, highly conductive current carrying layer at its base. Again, theinduced magnetic eld cannot diuse through that thin layer and the whole plasma sheet is thenaccelerated parallel to the electrodes like a `slug'.

Whether in the `snowplow' mode or `slug' mode, the operation of a short-pulsed plasma acceleratorcan be described with a simple set of equations. Because the discharge current is conned to a thinand highly conductive plasma layer, it can be assigned within a certain accuracy a well dened axialposition z(t) inside the accelerator (Fig. 4). As z(t) increases with time, the volume inside whichB(t) can spread increases too and, thus, the electro-dynamic behaviour of the plasma gun can bedescribed with a time-varying self-inductance L(t). The driving circuit usually consists of one or severalcapacitors which are charged-up to the voltage U0 prior to the discharge. The dynamic properties ofthe driving circuit are represented by the circuit inductance L0 and circuit resistance R0 [25, 34].

Fig. 4: Modelization of the pulsed plasma accelerator with the help of lumped circuit elements

With the help of Kirchho's law, a rst equation can be derived for the electrical discharge,

1

C0Q(t) +R0

dQ(t)

d t

2

+ [L0 + L(t)]d2Q(t)

d t2+

dL(t)

d t

dQ(t)

d t= 0 ; (1)

4

with Q(t) = C0U(t) being the momentary capacitor charge. With the help of Newton's 3rd law asecond equation can be derived for the acceleration of the propellant. Since the Lorentz force is theonly relevant force, this equation writes as

dm(t)v(t)

d t1

2L0

dQ(t)

d t

2= 0 : (2)

In (2) the driving force has been expressed with the help of the inductance per unit length L0 of theaccelerator. For coaxial geometries with constant cathode radius rc and constant anode radius ra,

L0 =0

2ln

rc

ra

and L(t) = L0 z(t) ; _L(t) = L0 _z(t) :

These equations can be solved using the following set of initial conditions

Q(0) = 0 ;dQ(0)

d t= 0 ; z(0) = 0 ;

d z(0)

d t= 0 : (3)

The system of equations (1), (2) and the initial conditions (3) represent a general description of thedischarge circuit behaviour and are valid whether the accelerator is operated in the `snoplow' or the`slug' mode. In the latter case however, a further simplication can be made by assuming that in arst approximation the accelerated mass m remains constant. This constant mass is often referred toin literature as the `mass-bit' mbit.In order to evaluate the investigated congurations, further criterions have to be introduced. Becauseof the impedance mismatch between the driving circuit and the accelerator, not all of the energyinitially stored in the capacitor will be transferred to the moving plasma. The dierence is either lostin R0 or is re ected back towards the capacitor and will be, in the worst case, lost in subsequentsecondary discharges. To describe the energy transfer at the end of the discharge pulse (t = p), anelectrical eÆciency el is introduced, with

el = 1L0

C0U2

0

dQ(p)

d t

2L(p)

C0U2

0

dQ(p)

d t

2

2R0

C0U2

0

Z p

0

dQ(t)

d t

2dt : (4)

For the purpose of electric space propulsion with kinetic energy being the only energy of interest, thethrust eÆciency t is used,

t =mbit v(p)

2

C0U2

0

: (5)

Although not explicitly appearing in this mathematical model, energy sinks for phase changes andionization, heat and momentum losses of the accelerated plasma towards the electrodes are neverthelessincluded in the experimentally determined values for el and t. Within the theoretical model, thedierence between the two eÆciencies is that el also includes the eventual compression work of theaccelerated plasma. In the `slug' mode t el whereas in the `snowplow' mode t < el.

4 Discussion of the results

For industrial processing it was decided to keep the freedom of choice of the propellant used (inertor reactive) and, hence, to design a plasma accelerator which would operate in the snowplow' mode[34, 35]. In order to create a plastic deformation of an Inconel 718 part's surface, a Hugoniot-pressureof H = 2:3 GPa needed to be achieved for a duration of H 100 ns. With Nickel as sacricialmaterial, this in turn required a power deposition of P 1:5 GW/cm2 and an exhaust velocityv(p) > 150 km/s. The latter ensures a penetration of the plasma into the sacricial layer and, thus,an eÆcient energy deposition and the use of el as a design criterion. With a systematic variation ofthe parameters C0; R0; L0; U0; 0; ra; rc, where the equations were numerically solved, a geometricaland electrical conguration as well as a nominal operation mode could be determined (Fig. 5).

5

32.0 kV

368.0 kA

441.0 kA

11.0 kV

227.9 km/s

242.0 km/s 41.2 %

µ

∆

ρ Ω

= 25.0 mmr = 35.0 mmr = 10.0 mm = 100.0 mg/m^3R = 3.00 m L = 12.0 nHU = 32.0 kVC = 22.0 F

|1076

|955

|835

|715

|595

|475

|355

|235

|115 [ns]

[mm]

c

a

0

0

0

0

0

1.59 GW/cm^2

1.74 GW/cm^2

0 50 100 150 200 250

axial snowplow position z, discharge time t

Fig. 5: Optimization results: Gun geometryand discharge parameter evolution

Ar

H

0

1

2

3

4

0 0.2 0.4 0.6 0.8 1

P

GW

cm2

i [s ]

Fig. 6: Operational window: Power densitydeposition vs. target interaction time i

The optimized nominal conguration is given by C0 = 22 F, L0 = 12 nH, R0 = 3 m, U0;nom = 32kV, 0;nom = 100 mg/m3, rc = 35 mm, ra = 10 mm and the electrode length lb = 250 mm. With agiven geometrical and electrical conguration, there exists the possibility to vary the charge voltageU0 and the ll density 0. The results of this variation are given in the operational window (Fig. 6)versus the plasma-target interaction time i for two propellants, H2 and Ar.

Fig. 7: 3-D view of the short-pulsed plasmaaccelerator for plasma impulse peening

A 3-D (exploded) front view of the technical real-ization is given in Fig. 7. The energy bank com-prises 22 capacitors of 1 F each at 40 kV ratedvoltage connected to circular plates of 1400 mmdiameter acting as the power transmission line.The overall length of the system is 1030 mm witha total mass of 420 kg, excluding controls, charg-ing supply and accelerator support structure.The simulation in Fig. 5 shows that a powerdensity of 1.6 GW/cm2 can be achieved undernominal operating conditions. The calculatedmaximum is around 3.6 GW/cm2 with a bankvoltage of 40 kV and H2 as working gas. Thedeposited energy density ranges from 20 to 650J/cm2, depending on the operating conditionsU0 and 0. The maximum instantaneous electricpower of the discharge circuit equals roughly 20GW. The high exhaust speeds of the plasma,230 km/s nominal and up to 600 km/s, makesure that the accelerated plasma penetratesthe sacricial layer and that all its kinetic andthermal energy is eÆciently delivered into thetarget material. The calculated energy transfereÆciency el is around 40% and the energy isdeposited with the calculated power density onan area of 12 cm2. As of today, the parts of theaccelerator are being manufactured.

In a parallel work [36], the possibility of using PPTs for small satellite primary propulsion has beeninvestigated. As mentioned above, the propulsion system is required to compensate the drag of theresidual atmosphere and, in addition, raise the altitude of the circular orbit of a 30 30 30 cmcubic micro-satellite of 35 kg mass from 350 to 1000 km. The average electric power continuouslyavailable for propulsion was estimated to be 14 W. With the assumption of a maximum solar activity,

6

the maximum drag was calculated to be 187 N at 350 km. Hence, the PPT was to be designed in afashion to produce more than this amount of thrust. With PTFE as the standard solid propellant forpulsed thrusters, the optimization was done using the `slug' mode. Further selection criteria were thethrust eÆciency t and a transport rate TR dened as

TR =msat mprop

mprot;

where msat = 35 kg is the total initial satellite mass, mprop the mass of the complete propulsionsystem including propellant and t the total trip time needed to spiral-up to 1000 km altitude. Withthe help of TR, propulsion system designs achieving both high payload fractions and short trip timescan be selected. The `slug' model has been tested on the high-energy PPT from Kurchatov [10] andthe predicted results were found to be in a 5% agreement with the experimentally determined ones.Furthermore, simple models for the estimations of the capacitor mass, mcap C0U

2

0, the structural

mass, mstr mcap as well as the ablated propellant, mbit C0U2

0were used for the optimization. The

propulsion system mass was then calculated according to

mprop mcap +mstr +mel +Nmbit ;

with N being the total number of pulses needed to reach 1000 km altitude and mel 0:75 kg themass of the control electronics.The subsequent numerical variation of the parameters C0; L0; R0; U0; ra; rc yielded the nominal con-guration C0 = 100 F, L0 = 40 nH, R0 = 2 m, U0;nom = 1000 V, rc = 40 mm, ra = 10 mmas well as lb=45 mm (barrel length). At nominal voltage U0;nom an impulse bit of Ibit = 867 Ns, aspecic impulse of Isp = 1312 s and t = 11:2 % can be achieved. With 14 W available power, a pulserepetition rate of p = 0:28 Hz, a thrust of F = 242 N and a mission v = 408 m/s are obtained. Thepropulsion system mass was calculated to be mprop = 5:63 kg with mPTFE = 1:09 kg of propellant.The orbit raising manoeuvre lasts 669 days with a total of 16.2 million pulses.With the ablated mass mbit varying with the stored energy, U0 is the only remaining independentoperational parameter. The evolution of the PPT's performance data as the capacitor charging voltageis varied is shown in Fig. 8. It follows that simply by adjusting U0, Ibit can be increased by a factor of4.4, Isp by 1.4 and t by 1.3, while at the same time the average thrust diminishes by 10%, along withthe pulse frequency. This high variability of pulsed plasma thrusters could play a major role when asingle thruster on the same satellite will be used for both primary and secondary propulsion purposes.The PPT described here will be built and investigated on a thrust stand in order to compare theorywith experiment.

Fig. 8: Pulsed plasma thruster performance data versus charging voltage

7

5 Conclusions

In a new line of research at IRS, a systematic investigation of short-pulsed coaxial plasma acceleratorsfor electric space propulsion and for advanced industrial processing was started in 2001. In cooperationwith industry, a 20 GW plasma gun was designed for plasma impulse peening of dynamically loadedengine parts in order to improve their fatigue resistance. In a parallel work, a small PPT was designedand optimized in view of an orbit raising mission for a micro-satellite with limited electric power. Thesevere design requirements for both applications appear to have been met and the accelerators willnow be built and experimentally investigated at IRS in order to be able to compare the measuredperformance data with the the ones predicted by the common theory. The technology of the plasmaguns can then be improved while making a maximum use of the synergies.

6 Acknowledgements

This work is supported by grants from the companies EADS-Dornier in Friedrichshafen and MTUMotoren- und Turbinen-Union GmbH in Munich, Germany.

References

[1] J. Marshall, \Performance of a Hydromagnetic Plasma Gun", Phys. Fluids 3, p. 134, 1960.

[2] R.H. Lovberg, B. Hayworth, T. Gooding, \The Use of a Coaxial Gun for Plasma Propulsion",nal report, contract NAS-5-1139, NASA Lewis Research Center, Cleveland, OH, 1962.

[3] W.J. Gunman, D.M.Nathanson, \Pulsed Plasma Microthruster Propulsion System for Syn-chronous Orbit Satellite", Journal of Spacecraft, 7/4, pp. 409415, 1970.

[4] R.J. Vondra, K. Thomassen, A. Solbes, \Analysis of Solid Te on Pulsed Plasma Thruster", Jour-nal of Spacecraft, 7/12, pp. 14021406, 1970.

[5] S.-M. An, H.-J. Wu, X.-Z. Feng, W.-X. Liu, \Space Flight Test of Electric Thruster SystemMDT-2A", Journal of Spacecraft, 21/6, pp. 593594, 1984.

[6] M. Hirata, H. Murakami, \Impulse Measurement of a Pulsed-Plasma Engine on Engineering TestSatellite-IV", Journal of Spacecraft, 21/6, pp. 553557, 1984.

[7] R.L. Burton, P.J. Turchi, \Pulsed Plasma Thruster", Journal of Propulsion and Power, 14/5, pp.716735, 1998.

[8] J.R. LeDuc, D.R. Bromaghim, T. Peterson, E. Pencil, L. Arrington, W.A. Hoskins, N.J. Meckel,R.J. Cassady, \Mission Planning, Hardware Development and Ground Testing for the PulsedPlasma Thruster (PPT) Space Demonstration on MightySat-II.1", AIAA 97-2779, 1997.

[9] C. Rayburn, M. Campbell, W.A. Hoskins, R.J. Cassady, \Development of a Micro Pulsed PlasmaThruster for the Dawgstar Nanosatellite", AIAA-2000-3256.

[10] Y.A. Alexeev, M.N. Kazeev, \Performance Study of High Power Ablative Pulsed PlasmaThruster", IEPC-99-207, 1999.

[11] C.L. Dailey, R.H. Lovberg, \Pulsed Inductive Thruster Performance Data Base for Megawatt-Class Engine Applications", Proceedings 10th Symposium on Space Nuclear Power and Propul-sion, CONF 930103 No. 271, 2, pp. 14251434, 1993.

[12] K. Zdunek, \Mechanism of Crystallisation of Multi-Component Metallic Coatings Using the Im-pulse Plasma Method", J. Materials Sci. Eng., 26/16, pp. 44334438, 1991.

[13] K. Zdunek, \Formation of Metallic Coatings on Non-Heated Substrates by the Impulse PlasmaMethod", J. Materials Sci. Eng., 140/16, pp. 709714, 1991.

8

[14] M. Sokolowski, A. Sokolowska, B. Gokiela, A. Michalski, A. Rusek, Z. Romanowski, \ReactivePulse Plasma Crystallization of Diamond and Diamond-Like Carbon", J. Crystal Growth, 47,pp. 421426, 1979.

[15] M. Sokolowski, \In uence of the Pulse Plasma Chemical Content on trhe Crystallization of Dia-mond Under Conditions of its Thermodynamic Instability", J. Crystal Growth, 54, pp. 519522,1981.

[16] M. Solo;owski, \Deposition of Wurtzite Type Boron Nitride Layers by Reactive Pulse PlasmaCrystallization", J. Crystal Growth, 46, pp. 136138, 1979.

[17] A. Sokolowska, A. Olszyna, A. Michalski, K. Zdunek, \Diamond Layers Deposited from ImpulsePlasma", Surface and Coatings Technology, 47, pp. 144155, 1991.

[18] A. Michalski, Z. Romanowski, \Reactive Pulse Plasma Cristallization of TiN Layers on Substratesat 500 K", J. Crystal Growth, 61, pp. 675680, 1983.

[19] J.N. Feugeas, E.C. Llonch, C.O. de Gonzalez, G. Galmbos, \Nitrogen Implantation of AISI 304Stainless Stell with a Coaxial Plasma Gun", J. Appl. Phys., 64, pp. 26482651, 1988.

[20] B.A. Kalin, V.I. Polsky, V.L. Yakushin, Yu.S. Virgilev, V.I. Vasiliev, S.S. Tserevitinov, \Erosionand Microstructure Change of Materials During Disruption-Simulation Experiments in PlasmaAccelerator", J. Nucl. Mater. 220222, pp. 934938, 1995.

[21] V.L. Yakushin, B.I. Vasiliev, S.S. Tserevitinov, \Use of High Temperature Pulsed Plasma Fluxesin Modication of Metal Materials", to be published in Surface & Coatings Technology.

[22] V.L. Yakushin, B.A. Kalin, S.N.Votinov, V.I. Vasiliev, \Investigation of Vanadium Alloys in CTRPlasma Disruption-Simulation Experiments", to be published in J. Nucl. Mater..

[23] Yu.V. Skvortsov, \Research on Pulsed and Steady-State Plasma Guns and Their Applications inthe Troitsk Branch of Kurchatov Institute of Atomic Energy", Phys. Fluids B, 4 (3), pp. 750756,1992.

[24] K.F. Schoenberg, R.A. Gerwin, R.W. Moses,Jr., J.T. Scheuer, H.P. Wagner, \Magnetohydrody-namic ow physics of magnetically nozzled plasma accelerators with applications to advancedmanufacturing", Phys. Plasmas, 5/5, pp. 20902104, 1998.

[25] H.P. Wagner, R.W. Moses, R. Gerwin, K.F. Schoenberg, \A short-pulse coaxial plasma acceleratorfor industrial processing", in preparation, to be submitted to IEEE Transactions on PlasmaScience.

[26] H.P. Wagner, R.A. Gerwin, R.W. Moses, K.F. Schoenberg, \Short-Pulse Plasma Thrustersfor Space Propulsion", 6th NASA/JPL Advanced Space Propulsion Workshop, March 2123,Pasadena, California, 1995.

[27] H.P. Wagner, R.W. Moses, K.F. Schoenberg, \Design of a Short-Pulse Plasma Accelerator forIndustrial Processes", 37th Annual APS-DPP Meeting, November 610, Louisville, Kentucky,1995.

[28] H.P. Wagner, R.W. Moses, K.F. Schoenberg, \The Magnetoplasmadynamic Processing Tool",23rd IEEE International Conference on Plasma Science, June 35, Boston, Massachusetts, 1996.

[29] H.P. Wagner, R.W. Moses, K.F. Schoenberg, \Realization of an Industrial Magnetoplasmady-namic Processing Tool", 38th Annual APS-DPP Meeting, November 1115, Denver, Colorado,1996.

[30] H.P.Wagner, R.W. Moses, K.F. Schoenberg, \The Industrial Magneto-Plasmadynamic ProcessingTool", International Conference on Metallurgical Coatings and Thin Films, April 2125, SanDiego, California, 1997.

9

[31] P. O'Hara, \Lasershotsm Peening", The Specialist Repairs & Processes Conference, May 1999,London, England.

[32] M. Steckling, \Lageregelung im Bogensekundenbereich am Beispiel des Mikrosatelliten DLR-TUBSAT", doctoral thesis, VDI progress report, 8/700, Technische Universitat Berlin, Berlin,Germany, 1998.

[33] T.D. Butler, I. Henins, F.C. Jahoda, J. Marshall, R.L. Morse, \Coaxial Snowplow Discharge",Phys. Fluids 12/9, pp. 19041916, 1969.

[34] H.P. Wagner, M. Auweter-Kurtz, \Studie Hyperschalltechnologie (Phase 2), report IRS-2002-P-01, Institut fur Raumfahrtsysteme, Universitat Stuttgart, Germany, 2002.

[35] H.P. Wagner, M. Auweter-Kurtz, \Palsma Impulse Peening of Compressor and Turbine Bladeswith Short-Pulsed Coaxial Plasma Accelerators", 3rd International Symposium on AppliedPlasma Science, July 26, Fairbanks, Alaska 2001.

[36] D. Ballhause, \Entwurf und Realisierung eines instationar gepulsten, koaxialen magnetoplasma-dynamischen Triebwerks", study thesis, IRS-02-S-26, Institut fur Raumfahrtsysteme, UniversitatStuttgart, Germany, 2002.

10