Reactive High Impulse Power Magnetron Sputtering

Reactive High Impulse Power Magnetron Sputtering: hysteresis behaviour and process control : 

1 Reactive High Impulse Power Magnetron Sputtering: hysteresis behaviour and process control M. Audronis, V. Bellido-Gonzalez, B. Daniel and D. Monaghan Gencoa Ltd, Physics Road, Speke, L24 9HP, UK 6 July 2010, Sheffield HIPIMS 2010

Structure : 

Structure HIPIMS, Reactive Sputtering / HIPIMS, Hysteresis behaviour Ti-O2, Ti-CO2, Cr-O2, Cr-C2H4, Zn:Al-O2, Al-O2, Target effects, Process control options, Plasma optical emission spectroscopy (OES) plasma analysis, Active feedback R HIPIMS control vs. constant flow, Conclusions. 2

High Power Impulse Magnetron Sputtering (HIPIMS) : 

3 High Power Impulse Magnetron Sputtering (HIPIMS) iPVD method1. Enhanced ionisation of sputtered flux. Ionisation up to ~ 90 % (only a few % in DC discharges). Useful for coating 3D features and/or enhancing properties of certain thin films. An example of U, I pulse waveforms in a HIPIMS discharge. 1 U. Helmersson, M. Lattemann, J. Bohlmark, A.P. Ehiasarian, J.T. Gudmundsson, Thin Solid Films 513 (2006) 1.

Reactive Sputtering : 

4 Reactive Sputtering A target of one chemical composition (e.g. Ti, Al) is sputtered in the presence of a gas (e.g. O2) that will react with the sputtered material to form a coating of a different chemical composition (e.g. compound TiO2, Al2O3). Mechanisms: chemisorption, reactive ion implantation2. Highly unstable – hysteresis effect, immediate transition from ‘metal’ to ‘poisoned’ state, arcing3. Allows relatively high (compared to compound targets) coating deposition rates to be obtained (transition region)4. Requires an active feedback control system to maintain the balance of reactive gas, which ensures stable operation in the ‘transition’ state4. 2 D. Depla, S. Mahieu, R. De Gryse, Thin Solid Films 517 (2009) 2825. 3 W.D. Sproul, D.J. Christie, D.C. Carter, Thin Solid Films 491 (2005) 1. 4 V. Bellido-Gonzalez, B. Daniel, J. Counsell, D.Monaghan, Thin Solid Films 502 (2006) 34.

Reactive (R) HIPIMS process : 

5 Reactive (R) HIPIMS process Exhibits hysteresis effect, sudden transition and arcing6. Same mechanisms: chemisorption, reactive ion implantation7. Shown using Plasma Monitoring (PM), Penning-PM and l-sensor6,7. 6 M. Audronis, V. Bellido-Gonzalez, B. Daniel Surface & Coatings Technology 204 (2010) 2159–2164. 7 M. Audronis, V. Bellido-Gonzalez Thin Solid Films 518 (2010) 1962–1965.

R HIPIMS process : 

R HIPIMS process Influence of pulsing parameters Transition (poisoning) region – nearly identical for different conditions. Depoisoning – slow/fast stages. Different behaviour for different pulsing parameters. 7 M. Audronis, V. Bellido-Gonzalez Thin Solid Films 518 (2010) 1962–1965.

R HIPIMS process aspects : 

R HIPIMS process aspects Target state Figures on the right (Ti-O2, 325 Hz, ~3 kW, clean vs. oxidised target). reactive oxygen ion implantation. Outgassing (e.g. substrates, chamber walls). 7 Target state effects

R HIPIMS process control options : 

8 R HIPIMS process control options Partial pressure mass spectrometry – reasonably quick, expensive. l-sensor (O2) – slow. Pulsing parameters Plasma optical emission monitoring (PM) Inexpensive, exhibits a very fast response time, information on the sputter target state, operation anywhere between the ‘Metal’ and ‘Compound’ states (and beyond if ‘gas’ lines are monitored), Works well when designed for HIPIMS (Gencoa patent pending).

Slide 9: 

Gencoa Speedflo: HIPIMS PM option

Experimental setup : 

10 Experimental setup Target materials: Ti, Cr, Zn:Al, Al. Reactive gases: O2, C2H4, CO2. Chemfilt and Huettinger HIPIMS power supplies. Gencoa Ltd. SpeedfloTM reactive sputtering process controller (HIPIMS PM option). Ocean Optics CCD Spectrometer USB 2000+. Schematic representation of the sputtering system used for experiments.

Selection of OES ‘metal’ lines to monitor/control R HIPIMS discharges : 

11 Selection of OES ‘metal’ lines to monitor/control R HIPIMS discharges Plasma OES. Green – HIPIMS of Ti. Red – R HIPIMS of Ti in Ar/O2 atmosphere. The intensity of Ti+ peaks is reduced as the target surface becomes oxidised. Applicable for DC/AC reactive sputtering & HIPIMS. OES spectra as recorded for Ti HIPIMS in Ar (green line) as well for reactive Ti HIPIMS in Ar/O2 atmosphere (red line) with the target fully poisoned. good lines to monitor

Slide 12: 

12 Selection of OES ‘gas’ lines to monitor/control R HIPIMS discharges Green – HIPIMS of Ti. Red – R HIPIMS of Ti in Ar/CO2 atmosphere. The intensity of O2* peak increases as the target surface becomes oxidised. Applicable for DC/AC reactive sputtering & HIPIMS. OES spectra as recorded for Ti HIPIMS in Ar (green line) as well for reactive Ti HIPIMS in Ar/CO2 atmosphere (red line) with the target fully poisoned.

Hysteresis behaviour : 

Hysteresis behaviour for various target material – reactive gas combinations 13

R HIPIMS of Ti-O2: hysteresis loop : 

14 R HIPIMS of Ti-O2: hysteresis loop Oxygen flow ramp from 0 to 30 sccm and associated response of the PM sensor signal during reactive Ti HIPIMS in Ar/O2. Hysteresis plot for R HIPIMS of Ti in Ar/O2 atmosphere @ ~3kW power, 325 Hz, 50 us on time. O2 ramp to characterise the process. Hysteresis loop plot; ‘metal’ PM signal vs. O2 flow (text book example).

Slide 15: 

15 R HIPIMS of Ti-CO2: hysteresis behaviour CO2 gas flow ramp from 0 to 15 sccm and associated response of the PM sensors (monitoring Ti+ and O2* emissions) during reactive Ti HIPIMS in Ar/CO2. Hysteresis plots for R HIPIMS of Ti in Ar/CO2 atmosphere @ ~3kW power, 600 Hz, 50 us on time. CO2 ramp to characterise Ti-CO2 R HIPIMS process. Hysteresis loops obtained for both, ‘metal’ Ti+ and ‘gas’ O2* optical emission signals. Similar to Ti–O2 process.

Slide 16: 

16 R HIPIMS of Cr-O2: hysteresis loop O2 gas flow ramp from 0 to 15 sccm and associated response of the PM sensor (monitoring Cr*) during reactive Cr HIPIMS in Ar/O2. Hysteresis plots for R HIPIMS of Cr in Ar/O2 atmosphere @ ~3kW power, 600 Hz, 50 us on time. O2 ramp to characterise Cr-O2 R HIPIMS process. Process is different as compared to Ti-O2/CO2. Slow/fast target poisoning stages; very sharp transition.

Slide 17: 

17 R HIPIMS of Cr-C2H4: hysteresis loop C2H4 (ethylene) ramp to characterise the process. Hysteresis loop plot - ‘metal’ Cr* PM signal vs. O2 flow. Hysteresis loop is narrower compared to that of metal-oxide systems. C2H4 gas flow ramp and associated response of the PM sensor (monitoring Cr*) during reactive Cr HIPIMS in Ar/C2H4. Hysteresis plots for R HIPIMS of Cr in Ar/C2H4 atmosphere @ ~3kW power, 600 Hz, 50 us on time.

Slide 18: 

18 R HIPIMS of Zn-O2: hysteresis loop O2 gas flow ramp and associated response of the PM sensor (monitoring O2*) during reactive Zn:Al HIPIMS in Ar/O2. Hysteresis plots for R HIPIMS of Zn:Al in Ar/O2 atmosphere @ ~1kW power, 60 Hz, 200 us on time. O2 ramp to characterise Zn:Al-O2 R HIPIMS process. Hysteresis loop plotted for ‘gas’ O2* optical emission signal. Noisy ‘metal’ Zn+ PM signal.

Slide 19: 

19 R HIPIMS of Al-O2: hysteresis loop O2 gas flow ramp from 0 to 15 sccm and associated response of the PM sensor (monitoring Al*) during reactive Al HIPIMS in Ar/O2. Hysteresis plots for R HIPIMS of Al in Ar/O2 atmosphere @ ~2.5kW power, 325 Hz, 50 us on time. O2 ramp – severe arcing below 60% PM signal. Slow ramp – same. Limited operation window (100 - 60%), unless the power supply arc handling circuitry is improved. Plasma ‘Off’

Constant flow process behaviuor : 

Constant flow process behaviuor for various target material – reactive gas combinations 20

Slide 21: 

21 The change of Penning-PEM signal (O2 778nm) during constant gas flow process controlled reactive HIPIMS of Ti in Ar/O2 atmosphere. R HIPIMS of Ti-O2: constant O2 flow O2 partial pressure increases with time. Target is drifting towards more ‘poisoned’ state. Constant flow does not provide stable R HIPIMS process. The change of PEM signal (Ti+ 330nm) during constant gas flow process controlled reactive HIPIMS of Ti in Ar/O2 atmosphere.

Slide 22: 

22 R HIPIMS control of Al-O2 and Cr-C2H2: constant flow No stable R HIPIMS process obtained for other systems, such as Al-O2 and Cr-C2H4. The change of PM signal (Al* 395nm) during constant gas flow process controlled reactive HIPIMS of Al in Ar/O2 atmosphere. Plasma ‘Off’ The change of PM signal (Cr* 520nm) during constant gas flow process controlled reactive HIPIMS of Cr in Ar/C2H4 atmosphere. Al-O2 Cr-C2H4

Active feedback process control using Speedflo : 

Active feedback process control using Speedflo for various target material – reactive gas combinations 23

Slide 24: 

24 R HIPIMS of Ti-O2: process control Accurate R HIPIMS Ti-O process control. Excellent long term process stability. R HIPIMS coating process control at 50% PM set-point during a >3 hour long deposition run. R HIPIMS control at different set-points using the PM-based reactive sputtering control technology (@ ~3kW power, 325 Hz, 50 us on time). (long term stability)

Slide 25: 

25 R HIPIMS of Ti-CO2: process control R HIPIMS process control at different set-points using Ti+ ‘metal’ signal. Ti-CO2 system ~3kW power, 600 Hz, 50 us on time. R HIPIMS process control at different set-points using O2* ‘gas’ signal. Ti-CO2 system ~3kW power, 600 Hz, 50 us on time. Accurate R HIPIMS Ti-CO2 process control. Smoother signal obtained using ‘gas’ O2* signal. Noisiness of the optical signal is HIPIMS power supply dependent. ‘metal’ line ‘gas’ line

Slide 26: 

26 R HIPIMS of Cr-O2 & Cr-C2H4: process control R HIPIMS process control at different set-points using Cr* ‘metal’ signal. Cr-O2 system ~3kW power, 600 Hz, 50 us on time. Accurate R HIPIMS Cr-O2 & Cr-C2H4 process control. noisy Cr ‘metal’ signal (arcing). Cr-O2 Al-O2 R HIPIMS process control at different set-points using Cr* ‘metal’ signal. Cr-C2H4 system ~3 kW power, 600 Hz, 50 us on time. Cr-C2H4

Slide 27: 

27 R HIPIMS of Zn-O2 & Al-O2 : process control Accurate R HIPIMS Zn:Al-O2 process control for production of TCO coatings. Al-O control above 60% only (due to severe arcing below the 60% PM signal mark). R HIPIMS Zn:Al-O2 process control using O2* ‘gas’ signal. ~1kW power, 60 Hz, 200 us on time. R HIPIMS process control at different set-points using Al* ‘metal’ signal. Al-O2 system ~2.5kW power, 325 Hz, 50 us on time. Severe arcing! Zn-O2 Al-O2

Slide 28: 

28 O2 ramps at different speeds

Slide 29: 

29 Concealed portion of the hysteresis loop Conventional hysteresis plot at different O2 ramp speeds

Slide 30: 

30 Sensor vs. Time plot: full extent of R HIPIMS hysteresis revealed

Conclusions : 

31 Conclusions R HIPIMS hysteresis behaviour has been demonstrated for various material and reactive gas combinations. R HIPIMS hysteresis behaviour is to certain extent material/gas dependent (same as in reactive DC and AC processes). PM-based active feedback control technology using Speedflo has been demonstrated to provide accurate process control for different target material and reactive gas combinations. Spedflo provides excellent long term process stability. Both ‘metal’ and ‘gas’ optical plasma emissions are adequate to control R HIPIMS. Constant reactive gas flow method does not lead to stable deposition process. Arc handling capability of HIPIMS power supplies requires improvements.

Thank you for your attention! : 

32 Thank you for your attention!