plasma flow and plasma expansion around 3d objects in metal
Transcrição
plasma flow and plasma expansion around 3d objects in metal
PLASMA FLOW AND PLASMA EXPANSION AROUND 3D OBJECTS IN METAL PLASMA IMMERSION ION IMPLANTATION Darina Manova & Stephan Mändl Leibniz-Institut für Oberflächenmodifizierung 1 Motivation Leibniz-Institut für Oberflächenmodifizierung 2 Motivation Visualisation of Water Flow from Dynamic Sand Dunes Leibniz-Institut für Oberflächenmodifizierung 3 Motivation Plasma Flow around Objects & Interaction with Expanding Plasma Sheath How to Measure? Analytic methods SIMS – very sensitive method, absolute calibration is necessary RBS – absolute calibration, low sensitivity Spectroscopic Ellipsometry – precise method for simple systems: interference fringes from transparent layer, extinction from adsorbing layer Colour – fast and global measurement, but very simplistic interpretation Leibniz-Institut für Oberflächenmodifizierung 4 Table of Contents Motivation Generation of Supersonic (Metal) Plasma Flow Experimental Set-up Sampling of plasma flow at surfaces Results Influence of background pressure Influence of high voltage Conclusions Leibniz-Institut für Oberflächenmodifizierung 5 MePIII: Vacuum Arc Vacuum Arc Self sustaining, high current, low voltage gas discharge Local heating (cathode spots) + thermal emission of electrons + ejection of metal atoms Plasma density near cathode ~ 1020 – 1024 cm-3 Properties of arc plasma Small voltage drop between cathode and anode Ion flux parallel to electron flux from the cathode Fully ionised plasma, charge states 1+ – 3+ Initial kinetic energy of ions 10 – 100 eV Supersonic flow Leibniz-Institut für Oberflächenmodifizierung 6 HD vs. MHD A. Anders, Surf. Coat. Technol., 136 (2001) T. Arnold, Ph.D. thesis Leibniz-Institut für Oberflächenmodifizierung 7 Diffusionsprozesse Hydrodynamics Re ≡ ρuL η transition from laminar to turbulent flow: Reynolds number Magnetohydrodynamics τ R µ 0 L2 / η µ 0 v A L S≡ = = τA L / vA η time scale for diffusion vs. convection: Lundquist number for vacuum arc S≈1 but resistive MHD dominant in plasma sheath regions N 1/ 3 << λc << LH mean free path similar to system dimensions in our case ⇒ transition between single particle and collective motion picture Leibniz-Institut für Oberflächenmodifizierung 8 MePIII&D and MHD Deposition vs. Implantation small & large sheath • Addition of high voltage pulses change sampling volume of plasma flux • Pulse length and pulse voltage as free parameters • Modification of Child-Langmuir law with non-stationary initial plasma Leibniz-Institut für Oberflächenmodifizierung P.V. Akimov et al., Physics of Plasmas 8 (2001) 3788 9 Experimental Set-up 30 mm 60 mm “Tail” Cathode Front side top Back side Gas inlet Trigger • Vacuum arc with simple shield as filter RF plasma source • Cathode materials: Al and Ti 39 cm • Ar flow: 0, 15, 35 sccm • Background pressure: 10-2, 0.9, 1.8 µbar • High voltage Pulses (f = 3 kHz): 0 to 10 kV • Pulse length: 2.5 to 50 µs • Substrates: Si and SiO2/Si 10 cm Cathode Pumping system Sample 10 cm Filter (shield) Leibniz-Institut für Oberflächenmodifizierung Pulse generator 10 MHD Flow 800 800 600 600 4 2 -5 Al: 10 mbar 0 kV 5 kV 10 kV 5 10 15 20 25 Distance from Edge (mm) 400 200 0 0 Al_m Al_e Ti_m Ti_e 30 Thickness (nm) Thickness (nm) 18 -2 Area Density (10 cm ) 6 1.8 µbar Al_0kV Al_10kV Ti_0kV Ti_10kV 400 200 0 0 5 -2 10 µbar 0.9 µbar 1.8 µbar 10 15 20 25 Distance from center (mm) Front Side • Minor pressure dependence as primary shield far away from substrate • Thinner layer near edge caused by off-normal incidence and correspondingly higher sputter rate • Mapping of sheath width for different materials complicated by different self-sputter yields and charge state distribution Leibniz-Institut für Oberflächenmodifizierung 11 Back Side top inside outside • Schematic set-up of coupons on the back side of the sample holder bottom Leibniz-Institut für Oberflächenmodifizierung 12 MHD Flow Background pressure 10-2 µbar Al 0.9 µbar 1.8 µbar 10-2 µbar Ti 0.9 µbar 1.8 µbar Leibniz-Institut für Oberflächenmodifizierung 13 Back Side 0 5 10 15 20 25 30 0 50% 75% 10 15 20 25 30 80 140 9,5 120 8 60 40 6 8,5 40 8,0 Ti (µbar) 25% -2 10 0.9 1.8 7,5 20 50% 75% 20 7,0 0 0 5 10 15 20 25 30 0 0 5 10 Position (mm) 15 20 25 30 Position (mm) • Colour can be used as a substitute for layer thickness • Reasonable agreement with layer thickness for Ti for thin films as long as film is still partially transparent • Optical properties of Al (i.e. n + k) apparently dependent of thickness • Gradient of film thickness directly comparable to mean free path (conversion factor still to be determined) Leibniz-Institut für Oberflächenmodifizierung 14 Thickness (nm) 80 60 9,0 Color Index 100 Thickness (nm) Al µbar 25% -2 10 0.9 1.8 10 Color Index 5 10,0 12 Total Current 80 60 40 0.9 µbar A) 10-2 µbar Total Current (m Ti 80 1.8 µbar 80 60 60 40 40 10 lse 15 2 Le 0 25 ng 3 th 0 4 (µ 0 5 s) 0 20 7. 1 .5 2 .0 2. 5 Volta 3.0 3.5 4. 0 g e (k V) 4.5 5.0 2. 5 0 5 0 100 Al 10-2 µbar 0.9 µbar 100 0 5 20 Pu 20 100 1.8 µbar 80 80 80 60 60 40 40 20 20 0 0 60 40 20 0 Leibniz-Institut für Oberflächenmodifizierung 15 Back Side I: small pulse length Ti, 10 kV, 0.01 µbar 225 45 200 40 175 35 150 30 125 25 100 20 75 15 50 10 25 5 0 0 0 5 10 15 20 25 30 Ti Thickness (nm) SiO2 Thickness (nm) Ti, 10 kV, 1.8 µbar Ti, 3,5 kV, 0.01 µbar • Deposition concentrated at high pressures on tail region • Only sputter removal of oxide from back side Position (mm) Leibniz-Institut für Oberflächenmodifizierung 16 Back Side: SE 180 180 150 150 120 120 90 Exp 60 Sim Psi Delta 30 ψ, ∆(degree) Winkel (Grad) Exp Sim1 Sim2 Sim3 Psi Delta 0.1 nm Ti 5.1 nm Interface 141.3 nm SiO2 Si substrate 90 1.6 nm 0.0 nm 0.0 nm 0.0 nm 152.8 nm 153.5 nm 60 30 196.4 nm SiO2 Si substrate 0 400 500 600 700 800 0 Wellenlänge (nm) 400 500 600 700 800 Wavelength (nm) • Surfaces subjected to sputtering by ion bombardment need modelling with a graded interface (i.e. implanted layer) for reasonable results (10 kV, 10-2 µbar series) Leibniz-Institut für Oberflächenmodifizierung 17 Back Side II: large pulse length 1.5 kV, 50 µs 2.5 kV, 50 µs 3.5 kV, 50 µs 3.5 kV, 15 µs • Peculiar behaviour at long pulse lengths and intermediate voltages can be traces to highly localized increase in ion flux • Initial oxide of 200 nm is completely removed at 3.5 kV and 50 µs 5 kV, 50 µs Leibniz-Institut für Oberflächenmodifizierung • Maximum depth could reach 500 nm 18 Discussion & Conclusions • (Plasma) flow around obstacles is still a very modern problem • Colour visualisations can be a very helpful tool • Congruence of different analytic results from the same samples is not always achieved as each method may measure something different • Metal plasmas originating from a cathodic arc can be guided by simple shields • The fraction of lost flow depends on pressure and voltage (50 – 75%) • MePIIID will lead to highly inhomogeneous deposition and implantation distributions; especially around low symmetry objects depending on pressure, voltage and pulse length • Combinatorial materials science could be based on this approach to vary energy flux and deposition rate independently from each other • Prediction of actual distributions requires much more work • Sheath expansion in 3D geometry is a very complex function, in contrast to 1D geometries where a monotonous increase with time and voltage is observed. Leibniz-Institut für Oberflächenmodifizierung 19 Comparison to PIII Geometries 30 mm 2.9 ×1017cm-2 60 mm 30 mm 2.7 × 1017 2.2 × cm-2 O O B O O O 90 mm 1017 cm-2 100 mm 3D samples Flat samples influenced by substrate holder Steel, 1.4301, PIII, 10 kV, front side, back side before and after corrosion test Leibniz-Institut für Oberflächenmodifizierung 20 Summary & Conclusions Interplay of plasma flow, plasma sheath and geometry can create very strange and highly localized effects Leibniz-Institut für Oberflächenmodifizierung 21 Danksagung Johanna Lutz Susann Heinrich Sabine Schirmer Katharina Scholze David Haldan Leibniz-Institut für Oberflächenmodifizierung 22