1. Surface and volume production of negative ions in a low-pressure plasma E. Stoffels, W.W. Stoffels, V.M. Kroutilina*, H.-E. Wagner* and J. Meichsner*, Department of Physics, Eindhoven University of Technology *Institute for Physics, University of Greifswald Negative ions in low-pressure plasmas: formation, confinement, extraction Negative ion detection by mass spectrometry Where (and how) are the ions produced? Implications for plasma chemistry

2. Volume production of negative ions (example: low-pressure oxygen plasma) For O - : dissociative attachment to (excited) O 2 : O 2 + e  O + O - { O 2 ( ), O 2 (a), O 2 (b) etc.} For O 2 - : non-dissociative attachment: O 2 + e + X  O 2 - + X charge transfer: O - + O 2 (a)  O + O 2 - dissociative attachment to O 3 : O 3 + e  O 2 - + O All these processes are inefficient at low pressures!

3. Some facts in radio-frequency plasmas in oxygen Electron/ion density measurements: Negative ion density 10 16 m -3 Electron density 10 15 m -3 Main ionO - Substantial O 2 - density:10-20% Modelling of ion density: calculated < measured(about factor 2) less molecular ions expected This is also valid for other electronegative gases Do we miss an important formation process?

4. Extraction of negative ions in RF plasmas Ions are heavy and immobile they “feel” average potential Ions have little energy (room temperature) Positive ions are accelerated by the sheath field Negative ions are confined in the glow No negative ion extraction during normal plasma operation! RFgnd X V

5. Mass spectrometry of negative ions problem: extraction In DC discharges - no problem (extraction to a mass spectrometer through an orifice in the anode) In RF discharges - sheath field must be (locally) cancelled Experimental tricks: –positively biased extraction orifice (*disturbs plasma) –pulsed plasma & detection in the afterglow (*no information about “active” discharge) However... We detect negative ions in RF without these tricks! What is the origin of these ions?

6. Negative ions in O 2 RF plasma energy spectra by QMS Direct negative ion fluxes recorded by mass spectrometer O - detected energy spectra obtained ions arrive at low energies (at most a few eV) potential barrier between glow and grounded electrode = 40 V O - from the glow does not have enough energy to pass it! RF gnd HIDEN QMS glow

7. Negative ions in O 2 RF plasma dependence on plasma parameters Where are the ions created? –not in the plasma glow –surface? sheath? Surface conversion (O +  O - ) unlikely –no forward velocity to the surface –ions would not reach the QMS Signals of O - increase with increasing power –high energy particles (positive ions, neutrals) important? –plasma chemistry? Signals increase with pressure –background gas important? –sheath process?

8. Most sheath-produced ions are directed to the glow. We detect only a small fraction. Is this an important ion formation channel? What does it mean for plasma chemistry? Case study: Take a DC plasma At the anode: ions from the glow At the cathode: sheath ions? -V0 cathode fall 0-V anode fall QMS (At reference -V)

9. DC glow in oxygen extraction at the anode Both O - and O 2 - are observed Typical energy spectra: O - energy  eV(anode fall) –ions accelerated in the column –not well thermalised O 2 - energy  eV(anode fall) –charge transfer collisions: O 2 - (E) + O 2 (th)  O 2 - (th) + O 2 (E) –efficient thermalisation –energy loss in the anode fall region O-O- O2-O2-

10. DC glow at low pressure Normally [O - ] > [O 2 - ] At p < 0,05 mbar O 2 - is dominant High-energy ions observed! V(anode) thermal ions (glow) “cathode” ions V - acceleration High-energy tail ions created in the cathode fall low pressures - less collisions these ions arrive at the anode!

11. DC glow in oxygen cathode side O - signal at 0,2 mbar: Ion production takes place close to the cathode Can we extract negative ions at the cathode side? QMS frame at a reference voltage = V(cathode) Large signals observed! Energy spectra of O - have a large high-energy tail O - velocity towards the cathode? O 2 - ions have energy  0

12. DC glow in oxygen cathode side Both O - and O 2 - signals increase with increasing pressure and DC voltage Signals are larger than in the RF plasma Cathode voltages higher than sheath voltages in RF High energy particles essential in negative ion generation! O- signal at 0,15

13. Possible mechanisms surface/RF sheath/DC cathode fall surface conversion of O 2 + (O + ) into O 2 - (O - ) : unlikely (at normal incidence of X +, X - is reflected) formation of fast (excited) neutrals O 2 + (high E) + O 2  O 2 + + O 2 (high E) and dissociative attachment O 2 (high E) + e  O - + O : unlikely (low electron density in the sheath) ion pair formation: X + (high E) + O 2  O + + O - (explains high energy of O - ) for O 2 - - charge exchange between excited and ground state molecules: O 2 * + O 2  O 2 + + O 2 - Problem: lack of cross-section data on ion-induced processes

14. Conclusions Negative ions can escape from low-pressure RF plasma (observed in O 2 as well as CF 4 plasmas) Ions are generated in the sheath region Sheath chemistry is very rich High energy particles are involved in ion production Most likely mechanism: ion pair formation upon high- energy positive ion impact Rough estimations: if  > 10 -22 m -2, O - production by ion pair formation in the sheath comparable with electron attachment in the glow! Sheath process may be the major formation channel for some molecular ions (O 2 - ) at low pressures (attachment does not work)