Direction for the Future - Successive Acceleration of Positive and Negative Ions Applied to Space Propulsion
DDirection for the Future – Successive Acceleration of Positive andNegative Ions Applied to Space Propulsion
A. Aanesland, J. Bredin, L. Popelier and P. Chabert
Laboratoire de Physique des Plasmas, CNRS – Ecole Polytechnique, France
Abstract
Electrical space thrusters show important advantages for applications in outerspace compared to chemical thrusters, as they allow a longer mission lifetimewith lower weight and propellant consumption. Mature technologies on themarket today accelerate positive ions to generate thrust. The ion beam is neu-tralized by electrons downstream, and this need for an additional neutralizationsystem has some drawbacks related to stability, lifetime and total weight andpower consumption. Many new concepts, to get rid of the neutralizer, havebeen proposed, and the PEGASES ion–ion thruster is one of them. This newthruster concept aims at accelerating both positive and negative ions to gener-ate thrust, such that additional neutralization is redundant. This chapter givesan overview of the concept of electric propulsion and the state of the develop-ment of this new ion–ion thruster. Introduction
In the late 1950s, the golden years of space exploration began. The technology improvements in rocketryafter World War II allowed us to meet the challenge of creating thrust that could overcome Earth’s gravity.Human exploration to the Moon became a reality; and also the huge technology breakthrough led toincreasing numbers of satellites in orbit around the Earth and unmanned missions to our neighbouringplanets, comets, asteroids, etc. The early
Pioneer and
Voyager missions have now reached interstellarspace and still transmit signals back to Earth.Electrical space thrusters show important advantages for applications in outer space compared tochemical thrusters, as they allow a longer mission lifetime with lower weight and propellant consump-tion. The first mission that used electric propulsion as a unique propulsion system was launched in 1989with the
Deep Space I programme. Since then, this technology has become more and more popularwithin the space sector, but still this makes up only a few per cent of the launched missions. The maturetechnologies on the market today rely on accelerating ions either via the Hall current (i.e., Hall thrustersor closed drift thrusters) or via electrostatic grids (i.e., ion thrusters) [1, 2]. Hall thrusters and griddedthrusters have various advantages and drawbacks, and it is mainly political issues that determine whichone is used on a mission. The advantages of gridded thrusters (over Hall thrusters) are lower beam diver-gence, slightly longer lifetime and the fact that they can deliver higher thrust and ion beam velocity. Thecommon point is that positive ions are accelerated from the plasma via electric fields to generate thrust.As the beam is positively charged, electrons are injected into the downstream space to ensure current andcharge neutralization. This need for downstream neutralization is a drawback in all existing systems, as ittakes up space and adds to the weight of the system. It also increases the complexity of the thrusters andconsequently the risk of failure. The
HAYABUSA mission is one example illustrating this problem [3].The PEGASES thruster concept has been proposed to improve electric propulsion and to removethe need for additional neutralization systems [4]. PEGASES is an acronym for ‘plasma propulsion withelectronegative gases’. It belongs to the electrostatic gridded thruster family, but, contrary to classicalsystems, it accelerates alternately positive and negative ions to provide thrust. In this way additionalelectron neutralization is redundant. It is also thought that, since the recombination rate of oppositelycharged ions is much higher than that for electron–ion recombination, the downstream beam or plumewill mainly consist of fast neutrals and a low density of charged particles. a r X i v : . [ phy s i c s . acc - ph ] N ov his chapter will outline the various stages of the PEGASES thruster, from plasma generation tothe formation of an electronegative plasma and the acceleration of both positive and negative ions toprovide thrust. Thrust and specific impulse in space propulsion
The thrust T acting on a spacecraft is given by the change in momentum such that T = − d( mv )d t = − v ex d m d t , (1)where v ex is the exhaust velocity and d m/ d t is the rate of change of the total mass ejected. The thrustrequired in space applications can vary from meganewtons to micronewtons. To escape the Earth’sgravitational field, the thrust needs to overcome the weight of the spacecraft. This requires very highthrust, which today can only be reached by chemical rockets that expel a large amount of mass veryquickly (i.e., a large d m/ d t ). However, when in space, the thrust can be much lower. Sometimes, onlymicro-thrust is needed to precisely position space vehicles with respect to one another, and at other timesthe goal is to reach a target with as little propellant as possible but with no strict requirement on themission duration. With low thrust, the spacecraft will not travel as fast, but will eventually reach thetarget. sp [s] (cid:54) m [ k g ] ∆ v = ∆ v = Fig. 1:
Propellant mass required as a function of the specific impulse for two different types of missions
Integrating Eq. (1) gives the well-known rocket equation, ∆ v = v ex ln (cid:18) m m f (cid:19) , (2)where ∆ v is the change in velocity of the space vessel, and m and m f are its initial and final mass,respectively. Hence, ∆ m = m − m f is the required propellant mass. This equation was first derivedby Konstantine E. Tsiolkovsky (1857–1935) and is sometimes referred to as the Tsiolkovsky rocketequation. The specific impulse is defined as I sp = v ex g , (3)and is given in the unit of seconds ( g is the gravitational constant at sea level). Eq. (2) shows that the I sp of a thruster is indirectly a measure of the propellant consumption: the higher I sp , the less propellant is2eeded to reach the same change in velocity. Figure 1 is an example given for an interplanetary missionthat will need a velocity change ∆ v of 0.5 km s − , or for a positioning in low Earth orbit with a velocitychange of 4 km s − . In the first example, to achieve the same manoeuvre one could go from a propellantmass of around 400–800 kg to only 25 kg by choosing electric rather than chemical propulsion. Yet, theduration to achieve the manoeuvre is very different due to the thrust level. To bring 1 kg up to low Earthorbit with an Ariane V rocket costs around 20 000 euros [5]. Hence reducing the propellant consumptiontranslates into real money for the industry. In the example above, a mission to the Moon saves 7–15million euros. For an ion beam, v ex is equivalent to the ion beam velocity v b and d m/ d t can be expressed as the ionflux Γ i out of the thruster with the given mass M i through an effective grid area for ions A i . Assumingthat v b = (cid:112) eV /M i is obtained by accelerating the ions across a voltage difference V , we can expressthe thrust provided by an ion beam as T i = A i Γ i M i v b = A i en s (cid:112) T e V . (4)Hence, the thrust depends on the plasma density at the sheath edge n s , the electron temperature T e , theacceleration voltage V , and the area of active extraction/acceleration A i . When the thruster is operatedat its maximum performance, the current is limited by the space-charge-limited current through the grids,given by the Child–Langmuir relation [6] J CL = 49 ε (cid:18) eM i (cid:19) / V / d ∗ , (5)where d ∗ is the effective space-charge-limited distance between the grids. Assuming that the plasmadensity and hence the flux from the plasma is balanced with the space-charge-limited current, such that Γ i = J CL /e , the maximum thrust can be expressed as T max = 89 ε A i d ∗ V . (6)Here we neglect beam divergence, which would reduce the thrust by cos θ , where θ is the beam di-vergence angle, and we neglect any losses in the grid system. Hence, provided that the plasma is ofsufficient density, the thrust depends only on the grid dimensions and the acceleration voltage. Higherthrust is achieved with larger surfaces and smaller grid distance (or, to be precise, a smaller space-charge-limited distance). As a short remark, note that the mass of the ions does not effect the thrust under theseconditions. However, the mass plays a significant role in the specific impulse, and hence the propellantconsumption.A global (volume-averaged) model of a gridded ion thruster has recently been developed [7].The neutral propellant (xenon gas) is injected into the thruster chamber at a fixed rate and a plasma isgenerated by circulating a radio-frequency (RF) current in an inductive coil. The ions generated in thisplasma are accelerated out of the thruster by a pair of d.c. biased grids. The neutralization downstream isnot treated. Xenon atoms also flow out of the thruster across the grids. The model, based on particle andenergy balance equations, solves for four global variables in the thruster chamber: the plasma density,the electron temperature, the neutral gas (atom) density, and the neutral gas temperature. Figure 2 shows(a) the densities and (b) the thrust as functions of the input RF power, for a gridded system where thegrids are separated by 1 mm. Intuitively, the thrust increases with increasing plasma density, as thecurrent or flux from the plasma increases. What is less intuitive is that the maximum thrust, limitedby the space charge, is reached for rather low RF power and plasma density. The ionization degree ingridded thrusters needs therefore only to be around 5–10%.3 ) b) Fig. 2: (a) Atom and ion densities and (b) corresponding thrust as a function of power. The dotted line indicatesthe Child–Langmuir space-charge limit for the grids used in this calculation ( d = 1 mm). The physics of the PEGASES thruster
Our team at Laboratoire de Physique des Plasmas (LPP) is developing a new thruster for space propul-sion. This thruster is called PEGASES (for ‘plasma propulsion with electronegative gases’) and belongsto the electric thrusters family. As for classical thrusters, the aim is to generate thrust and high specificimpulse by the acceleration of charged particles. The innovative idea behind the PEGASES thruster is togenerate thrust by both positive and negative ions, which has many advantages over existing technolo-gies, such as eliminating the additional neutralization system and decreasing the weight and size of thethruster.
A+ e- A+A-A-RF powerA2
MagneticfilteringPlasma core Ion-Ion plasma AccelerationRecombination
A+ A-
Fig. 3:
Illustration of the PEGASES thruster, where positive and negative ions are accelerated alternately from anion–ion plasma to generate thrust.
The PEGASES concept is illustrated in Fig. 3 and can be described by three almost independentstages:
Stage 1 is the plasma generation stage, where RF power is coupled to the plasma such that electrons areheated and cause ionization. This stage is important for the power efficiency of the thruster.
Stage 2 is the formation of an ion–ion plasma a certain distance from the ionization in Stage 1. For thiswe use a magnetic filter to confine and cool down the electrons. The hot electrons upstream of thebarrier lead to high ionization and production of positive ions, while the cold electrons downstreamof the barrier lead to efficient electron attachment and production of negative ions (for this to occurwe need to use electronegative gases such as halogen-containing gases, e.g., O , SF or I ). As aresult, positive and negative ions exist downstream with negligible contribution of electrons. Thisdownstream region is described as an ion–ion plasma. Stage 3 is the acceleration stage, where an alternate acceleration of positive and negative ions (from the4on–ion region in Stage 2) will provide the thrust for the spacecraft. The acceleration in PEGASESis based on classical gridded thrusters, where the acceleration is obtained by creating an electricfield between two or more grids. In PEGASES the acceleration field is changed in time by applyinga square waveform to the plasma grid; this allows consecutive bursts of positive and negative ionbeams.
RF plasmas are very often used in the semiconductor industry for etching and deposition, and a detaileddescription of the physics of these plasma discharges can be found elsewhere [8]. Briefly, the electro-magnetic field from the RF power supply will couple to the electrons in the plasma and transfer energyto them. As illustrated in Fig. 4, there are a variety of methods to couple this energy to the electrons. Forexample, in Fig. 4(a) the RF voltage (from an electrode or a coil) couples capacitively to the electronswithin an oscillating sheath. In Fig. 4(b) the RF current flowing in a coil induces an RF current in theplasma. The induced electric field transfers the energy to the electrons and consequently the field decaysover a distance of a few centimetres in the plasma, called the skin depth. In Fig. 4(c) an additional staticmagnetic field allows an electromagnetic wave to propagate in the plasma and in this case transfers theenergy to the electrons in the volume.The plasma source in the PEGASES thruster is a purely inductively coupled plasma (ICP) withoutcapacitive coupling, symmetrically driven at 4 MHz. The planar inductor is separated from the plasmaby a thin (3 mm) ceramic window and encapsulated in a ferrite to reduce losses and to enhance the ICPcoupling to the plasma [9, 10]. The RF power is fed to the inductor via an impedance matching networkusing a low-loss transmission-line transformer and air variable capacitors in symmetrical (push–pull)configuration. The symmetrical drive of the ICP inductor practically eliminates capacitive coupling tothe plasma, resulting in negligible plasma RF potential. The power efficiency in such systems can reachup to 90% or more [9] and is very important in space applications where every gram and every unit ofelectric power counts.
I ~ B ~ I ind ~ Skin depth
I V
Sheath
E ~ ~ ~
I ~ B EM wave
Capacitive E-mode Inductive H-mode Wave W-mode B ~ Volume a) b) c)
Fig. 4:
Illustration of RF power discharges operated in the (a) capacitive, (b) inductive and (c) wave modes
The plasma in the PEGASES thruster contains both positively and negatively charged ions. To createnegative ions, the gas has to be electronegative, allowing electrons to attach to neutral species. Thisprocess can occur in the volume or on surfaces with low work function. In the negative ion sources forfusion, negative hydrogen ions are for example produced on a caesium surface [11, 12].In most other experiments with negative ions, negative ion production occurs in volume by whatwe call dissociative attachment collisions. Electrons are typically too energetic to attach directly to aneutral, so in the process either the excess energy leads to the neutral molecule breaking up and/or it dis-sipates into vibrational and rotational energies. Electronegative gases are therefore molecular gases andtypically formed by halogens (group 17 in the periodic table) such as Cl and I or halogen-containingmolecules such as SF and CF . Although not halogens, O and H are also electronegative gases.5he ionization and attachment rate coefficients are typically increasing and decreasing functions of theelectron temperature, where the typical reactions are given in the form [13] AB x + e → AB + x + 2e (7) AB x + e → AB − y + B z (8)for ionization and dissociative attachment. Hence, for most electronegative gases, ionization dominatesfor high electron temperatures whereas attachment occurs for low temperatures. The aim for this stage is to segregate the plasma into two regions: (i) the plasma core, where electronshave a rather high temperature for efficient ionization; and (ii) a downstream ion–ion plasma, where theelectron density can be neglected and the plasma dynamics is controlled by ions. As seen above, toacquire this segregation, we need to control the electron temperature in the plasma. Commonly, in low-temperature plasmas, the electron temperature T e is governed by the ionization balance (electron creationand loss processes) and is a function of the kind of gas and the product pL , where p is the gas pressure and L is the characteristic size of the plasma [14]. The specific mechanism of electron heating seen in Fig.4 and the value of discharge power have a minor influence on the electron temperature. In negative ionsources, the electron cooling is usually achieved with magnetic filters (a localized transverse magneticfield) placed in front of the ion extracting aperture. This technique is also used in the PEGASES thruster,where a localized magnetic barrier is generated by a set of permanent neodymium magnets forming alocalized Gaussian magnetic field perpendicular to the extraction axis. The illustration of the PEGASESprototype with the magnetic field lines is shown in Fig. 5(a). We have recently shown that the gradientand strength of the magnetic field strongly affect the electron cooling, and the position of the minimumelectron temperature is achieved at the maximum magnetic field strength [10]. We have also shown thatthe ion–ion plasma is also generated around the maximum field region [15]. As an example, the electrontemperature measured in argon, and the measured densities in SF are shown in Fig. 5. Close to themaximum magnetic field, the electron density is three orders of magnitude lower than the ion densities.The ion densities in this region remain high, showing that ion transport across the magnetic field is notaffected by the magnetic field. For an RF power of only 120 W, the ion density reaches × m − inthe ion–ion region. RF coil Distributed gas injection Ceramic window
Pumping surface Ferrite material Teflon Aluminum body
12 cm xz yxMagnets Flange T e ( e V ) B ( G ) !" (( ) * ' + , - ! . / ' ' ! " ' + / ' ' a) b) T e ( e V ) B ( G ) T e ( e V ) B ( G ) !" (( ) * ' + , - ! . / ' ' ! " ' + / ' ' a) b) T e ( e V ) B ( G ) B=0 B=245 G x (cm) n ( m − ) n + n − n EEPF n e a) a) c) b) Fig. 5:
Illustration of RF power discharges operated in the (a) capacitive, (b) inductive and (c) wave modes
The PEGASES thruster is a gridded thruster based on the same principles as classical gridded acceler-ation systems. However, the originality of this thruster is that positive and negative ions are alternatelyaccelerated from an ion–ion plasma using the same grids. To achieve this, the first grid in contact with6he plasma is biased with alternate voltage waveforms and the second grid is grounded [16]. The wallsin the thruster are floating such that the potential of the plasma follows the plasma grid potential, andhence the electric field between the two grids will change direction during one bias period. A simplifiedillustration of this concept is shown in Fig. 6, where (a) shows one hole in the grid system with idealizedion trajectories, and (b) shows the potential distribution in the plasma and across the grids for the positiveand negative bias period applied to the first grid. Note that possible variations of the plasma potentialwithin the bulk plasma are not shown here. sheath grid 1 meniscusfloating boundaries grid 2+V g1 P o t en t i a l ( V ) Biased waveform ion beam packets + ion-ion plasma - -V g1 + - a)b) Fig. 6:
Illustration of RF power discharges operated in the (a) capacitive, (b) inductive and (c) wave modes
The acceleration grids in any ion source are designed and adapted to the acceleration voltage andthe available plasma parameters (density and temperature), to achieve the desired beam properties withrespect to thrust, beam current, divergence, etc. Square waveforms are therefore best suited. Particle-in-cell simulations have shown that the ratio of extracted positive and negative ion fluxes from the plasmadepends strongly on the electronegativity, particularly when the negative ion density n i − is much higherthan the electron density n e [17]. When negative ions are extracted from the plasma, we should thereforeexpect co-extracted electrons. In this case, the period over which the negative charges are acceleratedshould be shorter than the corresponding period for positive ions in order to compensate both the chargedensity and the current (i.e., the duty cycle can be optimized). This asymmetry could even be pushedtowards a system of accelerating positive ions and electrons. For an optimized system, the accelerationvoltages for the positive and negative charges should also be adapted independently to compensate for thedifference in their effective temperatures. The most suitable waveform is therefore a square waveformwith adjustable duty cycle and voltage offset.The upper and lower frequency requirements for the applied voltage should take into account:(i) both the ion plasma frequency (so the ions can react to the variations) and the transit time of the ionsthrough the grids; and (ii) the potential barrier created in the downstream space from the space chargeof the single beam packet or envelope, respectively. Analytical models developed in one dimension,which completely neglect any secondary electron emission or other external sources of space-chargeneutralization, predict an operating frequency in the lower megahertz region [18].7 .4 First laboratory tests in PEGASES II Successive positive and negative ion beams have been measured downstream of the alternately biasedgrids in the PEGASES II thruster. The grids are placed in the ion–ion plasma region and the bias fre-quency is only 1 kHz biased. Figure 7(a) shows a typical result of time-resolved ion energy distributionfunctions (IEDFs) measured downstream of the acceleration stage. The red and blue curves are obtainedduring the positive and negative bias periods, respectively. Both the positive and negative ions are mea-sured with a constant amplitude and energy during their respective bias period. However, their meanenergies are not the same. Figure 7(b) shows the IEDFs measured for various positions downstream ofthe grounded grid. These measurements are continuous while the grid is alternately biased. The peak ionenergies remain constant downstream of the grid, but interestingly the positive peak amplitude decreaseswhile the negative ion peak increases as a function of position. − − −
50 0 50 100 150012345 x 10 − V D [V] I E D F − − −
50 0 50 100 15000.20.40.60.81 x 10 − V D [V] I E D F a) b) Fig. 7: (a) IEDFs measured during the bias period, where the red and blue curves are measured during the positiveand negative bias period, respectively. (b) IEDF measured continuously for various positions downstream of thegrounded grid. The acceleration voltage is here V = ± V at a bias frequency of 1 kHz.
Figure 8(a) shows the ion mean energies as a function of the applied acceleration voltage for con-tinuous and alternate acceleration operated in a mixture of argon and SF (about 20% SF ). The positiveions have energies equivalent to the acceleration voltage V . Some variations are obtained between eachdataset, possibly due to slight differences in the ion–ion plasma conditions and therefore a change in thesheath voltage in front of the grids. The negative ions are, on the contrary, measured with lower energiesthan expected from the applied acceleration voltage. Similar results are obtained with planar probes.Figure 7(b) shows that the beam energy does not depend on the downstream position, indicating that thebipolar beam is space-charge-neutralized downstream.The difference in the positive and negative ion energies can be understood as follows. In plasmaimmersed ion implantation (PIII) sudden negative bias pulses are applied to a surface such that positiveions are implanted in the material [19]. Several authors have seen that implanting into a dielectric sub-strate results in a significant voltage build-up in the wafer, reducing the effective implant energy [20, 21].In our case, a film composed of fluorine and sulfur can deposit on the acceleration grids. A dielectricfilm is observed by eye on the grids after some hours of operation in mixed Ar/SF and after some tensof minutes in pure SF . This film is also detectable by measuring a decrease in the current on the grids.In the presence of a dielectric layer on the plasma grid, a certain voltage will be dropped in the dielectricsuch that the effective acceleration voltage is [21] V eff = V − eδQε ε r , (9)8 −
100 0 100 200 − − [V] V b [ V ] − −
100 0 100 200 − − − [V] V b [ V ] a) b) Fig. 8:
Ion mean energy as a function of the acceleration voltage. (a) Circles and diamonds are obtained for alter-nate and continuous acceleration, respectively, and measured with the Retarding Field Energy Analyser (RFEA).Triangles are measured with a planar probe in continuous mode. (b) Increasing deposition of a fluorine–sulfur filmon the acceleration grid. Diamonds are the initial conditions, triangles after about 5 min and diamonds after 30 minof operation, and circles after cleaning the grids. where δ is the thickness of the dielectric film and Q is the surface charge. For a quick estimate using δ ∼ µ m, Q ∼ × and ε r = 3 . as for pure sulfur, the voltage drop in the dielectric is about50 V, which is of the order of the measured energy loss for negative ions. Figure 8(b) shows the beamenergies as a function of V in pure SF after about 5 min and 30 min of operation, i.e., the deposit on thegrids became thicker and thicker. The negative ions are not affected much with increasing film thickness.However, the positive ion energies are similar to the negative ion energies after some time of operation.It seems likely that it is the dielectric film that is responsible for the lower beam energies. However, it isstill uncertain why the negative ions are affected more than the positive ions in the initial conditions. Thismight be due to a change in the dielectric layer formed as a function of the bias voltage or a change inhow this dielectric is charged when applying a positive or negative bias. Further experiments are neededbefore any definite conclusion can be drawn. Conclusion
The PEGASES thruster is a new and promising gridded ion thruster where, in contrast to classical grid-ded thrusters, both positive and negative ions are accelerated alternately from the same source. Thisconcept makes the additional neutralization system redundant and provides therefore many advantagesover existing thrusters. The summary given here provides a brief overview of the state of the art in thedevelopment of this thruster. Although there is still a long way to go before this thruster is flying inspace, we show that an ion–ion plasma is efficiently produced downstream of a localized magnetic filter,such that the degree of ionization required in a gridded thruster is met. The ion–ion plasma source cantherefore provide high enough densities for gridded ion thrusters operating with an acceleration of 500 Vor so, and can therefore provide similar thrust and I sp as classical gridded thrusters. We have also shownthe first laboratory evidence that alternate acceleration of positive and negative ions can be achieved. Acknowledgements
This work was funded by EADS Astrium and by Agence Nationale de la Recherche (ANR) under con-tract ANR-11-BS09-040. The authors would like to thank V. Godyak, J.-P. Booth and G. Hagelaar foruseful discussions and collaboration. 9 eferences [1] H.R. Kaufman, Technology of closed-drift thrusters.
AIAA J. (1985) 78–87.[2] D.M. Goebel and I. Katz, Fundamentals of Electric Propulsion: Ion and Hall Thrusters (JPL SpaceScience and Technology Series, Wiley, Hoboken, NJ, 2008).[3] N. Kazutaka and H. Kuninaka,
Trans. Jpn. Soc. Aeronaut. Space Sci., Aerosp. Technol. Jpn. (2012) Tb-1.[4] A. Aanesland, A. Meige, and P. Chabert, Electric propulsion using ion–ion plasmas. J. Phys.: Conf.Ser. (2009) 012009.[5] R. Janovsky and D.E. Koelle, Development and transportation costs of space launch systems. Pre-sented at the DGLR/CEAS European Air and Space Conference (2007).[6] H.R. Kaufman, Technology and applications of broad-beam ion sources used in sputtering. Part I.Ion source technology.
J. Vac. Sci. Technol. (1982) 725.[7] P. Chabert, J. Arancibia Monreal, J. Bredin, L. Popelier, and A. Aanesland, Global model ofa gridded-ion thruster powered by a radiofrequency inductive coil. Phys. Plasmas (7) (2012)073512.[8] P. Chabert and N.S.J. Braithwaite, Physics of Radio-Frequency Plasmas (Cambridge UniversityPress, Cambridge, 2011).[9] V.A. Godyak, Electrical and plasma parameters of ICP with high coupling efficiency.
PlasmaSources Sci. Technol. (2011) 025004.[10] A. Aanesland, J. Bredin, P. Chabert, and V.A. Godyak, Electron energy distribution function andplasma parameters across magnetic filters. Appl. Phys. Lett. (2012) 044102.[11] Yu.I. Belchenko, G.I. Dimov, and V.G. Dudnikov, A powerful injector of neutrals with a surface-plasma source of negative ions.
Nucl. Fusion (1) (1974) 113.[12] M. Bacal, Physics aspects of negative ion sources. Nucl. Fusion (2006) S250–S259.[13] D. Rapp and D.D. Briglia, Total cross sections for ionization and attachment in gases by electronimpact. II. Negative-ion formation. J. Chem. Phys. (1965) 1480.[14] Lieberman, M. A., and Lichtenberg, A. J., Principles of Plasma Discharges and Materials Process-ing. (2nd ed.). Wiley-Interscience (2005).[15] J. Bredin, P. Chabert, and A. Aanesland, Langmuir probe analysis of highly electronegative plas-mas. Appl. Phys. Lett. (2013) 154107.[16] A. Aanesland, P. Chabert, M. Irzyk, and S. Mazouffre, Method and device for forming a plasmabeam. Patent WO 2012/042143 A1 (2012).[17] N. Oudini, J.-L. Raimbault, P. Chabert, A. Meige, and A. Aanesland, Particle-in-cell simulationof an electronegative plasma under direct current bias studied in a large range of electronegativity.
Phys. Plasmas (2013) 043501.[18] A. Aanesland, Low temperature plasmas exposed to external magnetic and electric fields. Habilita-tion à Diriger des Recherches, Université Pierre et Marie Curie, 2013.[19] M.A. Lieberman, Model of plasma immersion ion implantation. J. Appl. Phys. (1989) 2926.[20] N. Cheung, Plasma immersion ion implantation for semiconductor processing. Mater. Chem. Phys. (1996) 132–139.[21] R.K.Y. Fu, Influence of thickness and dielectric properties on implantation efficacy in plasma im-mersion ion implantation of insulators. J. Appl. Phys.95