Time resolved LIF studies on sputtered atoms velocity distribution function 13 th FLTPD Ludovic de Poucques, Mikaël Desecures, Abderzak El Farsy, Jamal Bougdira Institut Jean Lamour, NANCY, FRANCE 13/05/2019
Outline Introduction (S2-S8) Magnetron sputtering, HiPIMS, objectives of our work. Experimental setup (S9-S18) : TR-TDLIF (Time Resolved-Tunable Diode Laser Induced Fluorescence) Optical arrangement, time resolution Velocity distribution function of sputtered W atoms (S19-S32) Analysis of the TR-TDLIF signal : calculation of Flux velocity distribution function (FVDF) of energetic and thermalized populations. Study of Ar/He gas mixture effect on the transport of W atoms. Velocity distribution function of sputtered Ti atoms (S33-S38) Analysis of the TR-TDLIF signal : highlights an intermediate regime of transport between ballistic (energetic atoms) and diffusive (thermalized atoms) ones, named quasi-diffusive regime of transport (quasi-thermalized atoms). Ludovic de Poucques –13 th FLTPD – 13/05/2019 1
Introduction Magnetron sputtering Conventional magnetron sputtering (dc-MS or rf-MS) : - Developed since the 70’s for microelectronics applications - Nowadays established and widely used method for thin films growth (thin films of materials such as metals, oxides, nitrides, ceramics, etc) Deposition on Sputtering of the the substrate target (W or Ti) n + + n + n Plasma ions Sputtered atoms due to : V � � E � B E : cathode voltage Racetrack on B � the target B : permanent magnets Ludovic de Poucques –13 th FLTPD – 13/05/2019 2
Introduction Magnetron sputtering Conventional magnetron sputtering (dc-MS or rf-MS) : - In dc-MS or rf-MS : ionization degree of sputtered atoms is weak (~1%). Sputtered atoms remain mainly neutral atoms (between target and substrate) (ionized sputtered atoms do not contribute to the deposition) The high level of expectations regarding new applications (e. g. deposition on complex shape : 3D) requires to ionize sputtered neutral atoms (difficult to control : trajectories and energies) HiPIMS : High Power Impulse Magnetron Sputtering (has emerged in the late 90’s) Ludovic de Poucques –13 th FLTPD – 13/05/2019 3
Introduction HiPIMS - high power (a few kW.cm -2 ) during short pulses (tens of µs) to avoid the cathode overheating and arc formation (dc-MS : a few 10 W.cm -2 ). - Plasma density : 10 12 - 10 13 cm -3 (dc-MS : ~10 11 cm -3 ). - Ionization degree of the sputtered particles : usually ≥ 50 % (dc-MS : ~1 %). Deposition on complex shape Cu Si Ion energy and trajectory can be controlled by polarizing the substrate, while neutrals are difficult to control. However, a significant fraction of sputtered neutral atoms remains and may influence thin film properties. Ludovic de Poucques –13 th FLTPD – 13/05/2019 4
Introduction - V Substrate Number of collisions Thermalized Energetic atoms pressure × distance Magnetron atoms Anode cathode Transport : function of “p × d”, ε : a few 0.01 eV gas mixture, σ ( ε ), ionization ε : several eV (Thomson’s distribution) Balistical transport: Diffusive transport: Energetic atoms (EN) Thermalized atoms (TH) collisions Objectives Therefore, the knowledge of the properties of incoming film-forming neutral species (study of atoms transport) is required for a better understanding of the HiPIMS deposition process …. One approach to study sputtered atoms transport is the experiment (TR-Laser diagnostics) Ludovic de Poucques –13 th FLTPD – 13/05/2019 5
Introduction Objectives Literature : some TR laser measurements on neutral sputtered atoms in HiPIMS Ex. 1 : Palmucci M. et al., JAP 114, 113302 (2013) TR-LIF measurements with a dye laser (pumped by Nd:YAG laser at 532 nm) : 305 ≤ λ ≤ 330 nm (DCM dye) LIF : space-resolved measurements In this work : distribution of v // is considered And the laser linewidth is ~0.8 pm (~Doppler broadening) Diode Laser (high spectral resolution : < 0.01 pm) Ludovic de Poucques –13 th FLTPD – 13/05/2019 6
Introduction Objectives Literature : some TR laser measurements on neutral sputtered atoms in HiPIMS Ti Ex. 2 : Sushkov V. et al., PSST 22, 015002 (2013) TR-TDLAS measurements with a single mode DL (Toptica Photonics DL 100) : λ = 398.18 nm z=7 cm LAS : information is averaged along laser beam which is not suitable to probe an inhomogenous plasma like in magnetron discharge (close to the target and/or at very low pressure) Most of laser experiments : distribution of v // is considered (v r ) Ludovic de Poucques –13 th FLTPD – 13/05/2019 7
Introduction Objectives Our work : characterize sputtered neutral atoms (W, Ti) transport in HiPIMS process. W Ti Axial laser beam knowing that the transport of atoms is from the target toward the substrate : TR-TDLIF (velocity component ⊥ to the target v z ). TR- axial VDF (z, t), z = the distance from the target. calibration by means of TR-TDLAS to get absolute values (in conditions where the sputtered vapor can be considered as homogeneous : TH, p, z, t). Ludovic de Poucques –13 th FLTPD – 13/05/2019 8
Outline Introduction (S2-S8) Magnetron sputtering, HiPIMS, objectives of our work. Experimental setup (S9-S18) : TR-TDLIF (Time Resolved-Tunable Diode Laser Induced Fluorescence) Optical arrangement, time resolution Velocity distribution function of sputtered W atoms (S19-S32) Analysis of the TR-TDLIF signal : calculation of Flux velocity distribution function (FVDF) of energetic and thermalized populations. Study of Ar/He gas mixture effect on the transport of W atoms. Velocity distribution function of sputtered Ti atoms (S33-S38) Analysis of the TR-TDLIF signal : highlights an intermediate regime of transport between ballistic (energetic atoms) and diffusive (thermalized atoms) ones, named quasi-diffusive regime of transport (quasi-thermalized atoms). Ludovic de Poucques –13 th FLTPD – 13/05/2019
Experimental setup Optical arrangement Ludovic de Poucques –13 th FLTPD – 13/05/2019 9
Experimental setup Optical arrangement DL beam is split into two parts with a beam splitter Ludovic de Poucques –13 th FLTPD – 13/05/2019 10
Experimental setup Optical arrangement The 1 st beam (20 %=4 mW) is guided to a Fabry–Pérot interferometer ( ∆ ν =1 GHz free spectral range) to perform a calibration of the laser wavelength scan ( ∆λ laser (t)). Ludovic de Poucques –13 th FLTPD – 13/05/2019 11
Experimental setup Optical arrangement The 2 nd beam (80 %=16 mW) : - launched into the discharge chamber - oriented normally to the target surface for v z measurements - towards the racetrack center (R 0 =1,3 cm), where sputtering of atoms is mostly localized Ludovic de Poucques –13 th FLTPD – 13/05/2019 12
Experimental setup Optical arrangement Doppler’s relation : ∆ λ � λ ����� � λ � � � � c λ � λ � V z =0 : λ laser = λ 0 = λ transition V z >0 : λ laser > λ 0 (toward the substrate) V z <0 : λ laser < λ 0 (backscattering) Ludovic de Poucques –13 th FLTPD – 13/05/2019 13
Experimental setup Detection system + laser beam section Optical arrangement Probed volume ≈ 3 mm 3 Reliable comparisons of VDF(z) : Detection system is fixed Magnetron cathode is moved by means of a bellows+linear translator system Ludovic de Poucques –13 th FLTPD – 13/05/2019 14
Experimental setup Time resolution Asymmetric current Current ramp (5 Hz) λ 1 λ 2 λ 3 λ 180 Wavelength: Time Pulse number: Ludovic de Poucques –13 th FLTPD – 13/05/2019 15
Experimental setup Time resolution Asymmetric current Current ramp (5 Hz) λ 1 λ 2 λ 3 λ 180 Wavelength: Time Pulse number: ∆ λ during one HiPIMS period � 1 ms ~ 14 180 � 0.077 pm Doppler velocity increment is only ∆ v z ~60 m/s for Ti (30 m/s for W) HiPIMS period n ° 1: TR-TDLIF at λ laser = λ 1 = λ 1 λ laser = λ 1 +0.077pm = λ 2 2: 3: λ laser = λ 1 +2 × 0.077pm = λ 3 180: λ laser = λ 1 + 179 × 0.077pm = λ 180 Ludovic de Poucques –13 th FLTPD – 13/05/2019 15
Experimental setup Time resolution Asymmetric current Current ramp (5 Hz) λ 1 λ 2 λ 3 λ 180 Wavelength: Time Pulse number: TR-TDLIF signal Signal de TD-LIF 0 40 80 120 160 Time ( µ s) Time (ms) Ludovic de Poucques –13 th FLTPD – 13/05/2019 16
Experimental setup Time resolution Asymmetric current Current ramp (5 Hz) λ 1 λ 2 λ 3 λ 180 Wavelength: Time Pulse number: λ 56 λ 55 λ 57 TR-TDLIF signal Signal de TD-LIF 55 56 57 Time (ms) 0 40 80 120 160 Time ( µ s) Time (ms) Ludovic de Poucques –13 th FLTPD – 13/05/2019 17
Experimental setup Time resolution Asymmetric current Current ramp (5 Hz) λ 1 λ 2 λ 3 λ 180 Wavelength: Time Temporal Pulse number: evolution λ ++ � λ � � � � ,-°++/ c λ � λ 56 λ 55 λ 57 TR-TDLIF signal Signal de TD-LIF λ +0 � λ � � � � ,-°+0/ c λ � v z (n ° 56)=v z (n ° 55)+ ∆ v z 55 56 57 Time (ms) AVDF(0 ≤ t ≤ ~1 ms) 0 40 80 120 160 Time ( µ s) Time (ms) Ludovic de Poucques –13 th FLTPD – 13/05/2019 18
Recommend
More recommend
