A NEW TECHNIQUE FOR ULTRAFAST VELOCITY DISTRIBUTION MEASUREMENTS OF ATOMIC SPECIES BY POST-IONIZATION LÁSER INDUCED FLUORESCENCE ( PILIF )

A new method for single shot velocity distribution measurement of metallic impurities of relevance for studies involving continuous sources, such as limiter experiments in fusion devises or sputtering experiments, based in the combination of Resonant Enhanced Multiphoton Ionization (REMPI) and Laser Induced Fluorescence (LIF) is proposed. High ionization yield and good time resolution are expected according to the numerical simulation of the experiment that has been run for several atomic species. Other possible applications of REMPI to plasma edge physics and to conventional techniques for velocity distribution measurements are briefly addressed.


1.INTR0DUCTI0N
Velocity distributions, of great relevance in many fields of physics and chemical physics, can be readily measured by Time of Flight ( TOF ) techniques "*, provided that free colusión conditions exist in the región between the source and the detector.The accuracy of the method is restricted by the finite sharpness of the gating function and the finite dimensions of the detection volume, so that long flight distances are needed for good resolution, in detriment of the signal to noise ratio.On the other hand, Doppler shifted LIF spectroscopy has been widely used for this purpose 1 , although the full sean on the láser wavelength required for the velocity distribution measurement implies a continuous character of the experiment and narrowing the láser bandwidth is made on expenses of pulse energy.
In plasma fusión research, velocity distribution of neutrals are important not only for the evaluation of impurity fluxes but also to determine the mechanism responsible for their ejection 2. The continuous character of the flow of sputtered partióles and the impossibility to chop it make time of flight (TOF) techniques not applicable to in situ velocity measurements in fusión plasma experiments so that Doppler shifted excitation in LIF detection of neutrals is the only method extensively used until now for this purpose^-In general these measurements require many, reproductible , plasma discharges and new methods based on fast scanning of the dye láser frecueney during a single discharge have been developed as alternative 4 .In any case velocity resolution has to be gained on expenses of the signal versus noise ratio and correction for the láser power at each wavelength is always needed due to the low saturation parameter required for these experiments .
In the present work a new method for single shot, in situ velocity distribution measurements based on REMPI in combination with Láser Induced Fluorescence (LIF) , of relevance for impurity flux determinations in fusión plasma research is proposed, altogether with some others applications.

PILIF EXPERIMENT
The proposed experiment consist on crossing two láser beams (the ionizing and probing ones) in the scattering volume and to record the time evolution of the LIF signal after complete ionization has taken place.After the detection volume has been depleted of neutral atoms by the REMPI process an spatial hole in terms of neutral density is formed.As the sputtered atoms start to fill it, the density in the observation volume will continuously increase and therefore the LIF signal when used as a density diagnostic.i.e.bandwith greater than the Doppler profile.This last requirement restricts the proposed experiment to those atomic systems where metastable levéis do not act as a sink of the laser-populated level , as in 3 level systems .
Assuming a well collimated atomic beam with a given velocity distribution f(v) and a scattering volume with dimensión parallel to the travelling direction / the density after a given time after ionization is given by : o ñ where n(0) stands for the density of neutrals before ionization takes place.The accurancy of the velocity distribution obtained with method will be limited by that of the scattering volume dimensions and its resolution by the dimensions themselves and the minimum sampling interval, ultimately limited by the lifetime of the excited level, providing that ionization takes place ¡n a large extend during the ionizing pulse.
The velocity distribution function,f(v),can be reconstructed from the time evolution of the LIF signal by applying the expression : where v in the case of a point source, far away from the scattering volume, ¡s simply given by //t.
The most appealing way to carry out the experiment would be by using the same experimental set-up as for LIF detection.This in many instances consists of a dye láser pumped by an excimer one, typically XeCI at 308 nm, so that a 4.03 eV high power photon source is readily available .This photon energy combined with that of the pumping photon is able to bring all the neutral atoms to the ionization continuum in most of the metáis typically monitored in limiter experiments.In order to completely ionize all the atoms ¡n the sampling volume , the right spiitting of the excimer power into direct (ionizing) beam and pumping one has to be made.That in principie will depend on the particular atomic system under consideration, but a simple calcuiation based on the rate equations for a three level system plus ionization 5 shows that if, for example, a 0.5 Jul/pulse XeCI láser and a ratio 1:1 between the power used to pump the dye láser and that of the ionizing beam are used, so that a modest 50 jaJui/pulse UV radiation is obtained after frecuency doubling ,still high enough to satúrate the resonant transition (S=150 for Be,S=100 for Fe.focussing in 2x2mm^ ), ionization will take place to a 100% in a time shorter than the láser pulse (15 to 20 nsec), even if the excimer radiation is focussed in an área several times larger ,thus minimizing alignment problems.As an example, Fig 1 shows the results for the Fe atom assuming an excimer láser square pulse of 20 nanoseconds and an ionization cross section, for the excited atom, of 10' 18 cm 2 • The conditions used for the simulation are similar to that used for REMPI detection of Fe atoms in sputtering experiments 6 .As it can be seen a high degree of ionization is expected even-at relatively low saturation parameter and modest excimer láser power.
Sampling of the LIF signal in a relevant time scale could be achieved by optically delaying successive reflections of the probing láser or by using a long pulse dye láser (several hundred nanoseconds ) and crossing the excimer at the beginning of the pulse, thus obtaining a continuous LIF signal.The first scheme implies long distances ( optical delay > pulse duration ) and correction for the láser divergence, so that the second scheme will be more feasible.Due to the short time required by the TOF experiment (see below ) synchronization of the two lasers should not be critical.

MODEL CALCULATION FOR EXTENDED SOURCES
The time evolution predicted by Eq. 1 is not directly applicable to extended sources as one has in limiter experiments.Convolution over all the emitter área and the different paths across the sampling volume depending on geometry, as well as attenuation through the plasma edge has to be taken into account.
The results of the calcuiation for a Be bar limiter, where a 0.6 cm diameter hole is used to look at the scattering volume, a prism of 2x2x4 mm placed at 1 cm from the limiter in this simulation (see Fig. 2) are displayed ¡n Fig. 3. Plasma edge temperature and density profiles as well as ionization and excitation rate constants are the same as in Ref 7 .No contribution to the refilling of the hole due to CX or electrón recombination is considered as they are expected to be not fast enough to effectively compete with the direct flux of sputtered atoms in the relevant time scale .A cosine distribution for sputtered partióles is assumed.
As it can be seen in Fig 3a, discrimination between sputtering (Thompson model ) and thermal distributions should be obvious even in a short time after the ionizing pulse (t=0).It must be recalled at this point that accurately measuring the Doppler profile for a thermal distribution requires an extremely narrow láser bandwidth, not always available .A higher resolution in this case, by PILIF will be easily achieved by simply enlarging /.A factor of two in the binding energy will be also distinguishable after several tens of nanosecond.Fig 3b shows the velocity distributions for these two cases (Eb= 3.32 eV and 1.66 eV respectively) as they would be measured by Doppler shift in the same geometry and plasma edge conditions.

EXPERIMENTAL SET-UP
The available experimental set-up, where the proposed experiment will be undertaken, has been previously described 8 .Basically, a DC Glow Discharge in an inert gas ( He, Ar) is produced , the stainless steel chamber acting as the cathode ( = 17% Cr ).The sputtered Cr atoms are detected by LIF at 429 nm ( two level system ) by using an excimer pumped ( XeCI ) dye láser ( Lambda Physik LPX 205¡-ibid.FL 3002E ).The available bandwidth can be reduced from 0.2 to 0.04 cm-1 by an intracavity etalon , so that a velocity resolution of = 0.5 Km/s can be obtained for the standard Doppler shifted LIF experiment ( see fig.3b) , providing that saturation of the transition is avoided ( S= 0.6 kW/cnr 2 s'" 1 ).Splitting of the excimer radiation at 308 nm will still allow 200 mJ/ pulse for the ionizing beam and > 20 mJ /pulse for pumping of the transition, so that the required conditions for the REMPI-LIF experiment can be easily fulfilled, even without focussing of either láser beam .A flashlamp-pumped dye láser ( 1 J/ pulse, pulse duration = 1 jas, bandwidth = 1 nm ), presently under construction, will be used for " long time" excitation under saturation conditions.The LIF time evolution will be recorded in a fast digital oscilloscope ( Tektronix DSA 601,1 GS/s ) and the data transferred to a PC.Synchronization of the excimer and the dye láser beams will be achieved by optically delaying part of the 429 nm radiation used to trigger the excimer láser via a fast photodiode.

OTHER EXPERIMENTS
The good spatial and temporal resolution of láser diagnostics could be used in the REMPI experiments to probé the plasma edge .Although no detailed calculations have been performed yet ,the screening properties of the plasma edge could,in principie , be tested without perturbing other plasma parameters in combination with neutral atomic beam diagnostics.Besides, REMPI could be used to créate a highly located.timeresolved high intensity pulse of single charged ions which propagation through the plasma, followed by optical methods,could yield information concerning particle transpon, among others.
In atomic beams experiments, the proposed ionization scheme could be used to optically chop the beam with an extremely narrow equivalent gate function , thus improving the velocity resolution in the conventional TOF experiment.Fig. 2 : Geometry assumed in the model calculation for a Be limiter .The interaction with the plasma in assumed to take place within a rectangle of 2 cm wide and 4 cm high.See text for the rest of the parameters used.¡ "Nueva técnica para la determinación ultrará-" pida de distribución de velocidades de especies ¡ atómicas por fluorescencia inducida por láser en ¡ la post-ionización".] "Nueva técnica para la determinación ultrará-1 pida de distribución de velocidades de especies i atómicas por fluorescencia inducida por láser en !la post-ionización".( Se presenta un nuevo método para la medida de distribución de velocidades de impurezas metálicas, de relevancia en estudios que Impliquen fuentes continuas i tales como los experimentos de interacción plasma-limitador en plasmas de fusión o , experimentos de spultering, basada en la combinación de las técnicas de Ionización • Mullifotónica Resonante y de Fluorescencia Inducida por Láser.De acuerdo con las i simulaciones hechas para diversas especies atómicas cabe espeiar un alto grado de ( ionización de las especies atómicas en estudio y una alta resolución temporal en la • medida de su distribución de velocidades.También se proponen otras posibles , aplicaciones del proceso de Ionización Multifotónica Resonante a estudios en el bordo de '

Fig. 3 :
Fig. 3 : Results from the model calculation : a) : Time evolution of the LIF signal.Thompson model for Eb (Be) =3.32 eV (top) and Eb=1.66 (middle) and thermal distribution (bottom) forT=2000 K. b) : Velocity distribution along the detection line of sight for the Eb=3.32 and Eb= 1.66 éV cases for the same geometry as above .

1 A
i A new method (or single shol velocity distribution measurement of metallic 1 impurilies of relevance for studies involving continuous sources, duch as limiler i experimente in fusión devices or spultering experiments , based in the combination of i Resonant Enhanced Multiphoton lonization (REMPI) and Láser Induced Fluorescence ¡ (LIF) is proposed.High ionization yield and good time resolution are expected 1 according lo the numerical simulation of the experiment that has been run for i several atomic species.Other possible applications of REMPI to plasma edge physics | and to conventional techniques for velocity distribution measurements are briefly .igacionesEnergéticas!, Medioambientales y Tecnológicaa.I Instituto de Investigación Básica.-Madrid.¡ "A new technique for ultrafast velocity dis-1 tribution measurements of atomic species by I post-ionization láser induced fluorescente ¡ (PILIF)".1 TASARES, F.L. (1992) 10 pp.;3 figs.;8refs. 1 new method for single shot velocity distribution measurement of metallic 1 impurilies of relevance for studies involving continuous sources, áuch as limiter i experimenls ¡n fusión devices or sputtering experiments , based in the combination of ¡ Resonant Enhanced Multiphoton lonization (REMPI) and Láser Induced Fluorescence 1 (LIF) is proposed.High ionization yield and good time resolution are expected i according to the numerical slmulation of the experiment that has been run for ¡ several atomic species.Other possible applications of REMPI to plasma edge physics 1 and to conventional techniques for velocity distribution measurements are briefly Energét iciis , Medioambient ales y Tecnológicas.Instituto de Investigación Básica.-Madrid."A new technique for ultrafast velocity dis-¡ tribution measurements of atomic species by ' post-ionization láser induced fluorescente < (PILIF)".¡ .TABARES, P.L. (1992) 10 pp.;3 figs.;Brefs.( !A new method for single shot velocity distribution measurement of metallic i impurities of relevance for studies involving continuous sources, áuch as limiter i experimenls in fusión devices or sputtering experiments , based In the combinalion of ' Resonant Enhanced Multiphoton lonization (REMPI) and Láser Induced Fluorescence i (LIF) is proposed.High ¡onization yield and good time resolution are expecled i according to the numerical simulation of the experiment that has been run for !several atomic species.Other possible applications of REMPI to plasma edge physics ' and to conventional techniques for velocity distribution measurements are briefly i addressed.¡ CIEMAT-686 ¡ Centro de Investigaciones Energéticas, Medioambientalea y Tecnológicas.' Instituto de Investigación Básica.-Madrid.' "A new technique for ultrafast velocity dis-! tribution measurements of atomic species by \ post-ionization láser induced fluorescente • (PILIF)".!TABARES, P.L. (1992) 10 pp.;3 figs.;flrefs.' A new method for single shot velocity distribution measurement of metallic ¡ impurities of relevance for sludies involving continuous sources, áuch as limiter ' experiments in fusión devices or sputtering experiments , based In the combinalion of i Resonant Enhanced Multiphoton lonization (REMPI) and Láser Induced Fluorescence , (LIF) ¡s proposed.High ionization yield and good time resolution are expected ¡ according to the numerical simulation of the experiment that has been run for ' several atomic species.Other possible applications of REMPI to plasma edge physics i and to conventional techniques for velocity distribution measurements are briefly , nridrosr.od-'« DOE CLASSIFICATION AND DESCRIPTORS: 700100.Time-of-fligh ¡ Method.Fluorescence Spectroscopy.Multi-photon Processes.i lonization.Simulation.Measuring ¡Methods.Velocity.¡ DOE CLASSIFICATION AND DESCRIPTORS: 700100.Time-of-t'ligh i Method.Fluorescence Spectroscopy.Multi-photon Processes. 1 Ionization.Simulation.Measuring Methods.Velocity. 1 DOE CLASSIFICATION AND DESCRIPTORS: 700100.Time-of-fligh ¡ Method.Fluorescence Spectroscopy.Multi-photon Processes., Ionization.Simulation.Measuring Methods.Velocity.• DOE CLASSIFICATION AND DESCRIPTORS: 700100.Time-o£-£1igh ¡ Method.Fluorescence Spectroscopy.Multi-photon Processes.Ionization.Simulation.Measuring Methods.Velocity.