Investigations on two different plasma tools by SEERS, OES and electrical measurements

Transcription

1 Investigations on two different plasma tools by SEERS, OES and electrical measurements Jan Konrad Bauer TU Dresden Institut für Halbleiter- und Mikrosystemtechnik Page 1 Das diesem Bericht zugrundeliegende Vorhaben wurde mit Mitteln des Sächsischen Staatsministeriums für Wirtschaft und Arbeit (Förderkennzeichen 5706) gefördert. Die Verantwortung für den Inhalt dieser Veröffentlichung liegt beim Autor"

2 2.2. Investigations on two different plasma tools SEERS, electrical and OES measurements! Comparison of SEERS, results on different tools! Additional tests with the HERCULES equipment! Analysis of CF 4 processes with SEERS at a parallel plate system! Analysis of CF 4 processes with OES at a parallel plate system! Comparison of CF 4 process data between SEERS and OES! Summary of this chapter Page 2

3 Comparison of SEERS values on different tools Parallel Plate System (ZWG plasma etching tool) chamber diameter: approx. 12 in max. wafer diameter: 4 in max. CCP power: 600W SENTECH SI 600 D (plasma etching system) chamber diameter: approx. 10 in max. wafer diameter: 4 in max. CCP power: 600W max. µ-wave power: 600W Page 3

4 Collision rate [1/s] Comparison of SEERS values on different tools Comparison of SEERS values for Ar at 10Pa and 50sccm with CCP only Upper row: Collision frequency, electron density and DC bias measured at the parallel plate system 1E9 0 1E8 1E Collision rate [1/s] 1E7 Electron density [1/ccm] 1E E6 DC bias [V] 1E Pressure [Pa] E Pressure [Pa] Pressure [Pa] Lower row: Collision frequency, electron density and DC bias measured at the SENTECH system 1E9 0 1E8 1E E7 1E6 Electron density [1/ccm] 1E9 DC bias [V] E Pressure [Pa] E Pressure [Pa] Pressure [Pa] Page 4

5 Comparison of SEERS values on different tools Comparison of SEERS values for N 2 at 10Pa and 50sccm with CCP only Upper row: Collision frequency, electron density and DC bias measured at the parallel plate system 1E9 1E Collision rate [1/s] 1E8 Electron density [1/ccm] DC bias [V] E E E6 1E8 1E Pressure [Pa] E Pressure [Pa] Pressure [Pa] Lower row: Collision frequency, electron density and DC bias measured at the SENTECH system 1E10 0 1E Collision rate [1/s] 1E9 Electron density [1/ccm] E E8 1E7 1E No plasma Pressure [Pa] 500 1E8 1E No plasma 30 Pressure [Pa] 500 DC bias [V] Pressure [Pa] Page 5

6 Comparison of SEERS values on different tools Resumé of HERCULES testing on different tools! Identical recipe parameters in different tools yield different plasma parameters. The tool construction and the HERCULES setup has an inpact on the SEERS algorithm results.! We observed a decreasing electron density at increasing CCP power which might be a hint for a HERCULES problem with the SENTECH SI 600 D! The very high collision frequencies and electron densities under some conditions at low CCP power are currently not well understood. Page 6

7 Additional tests with the HERCULES equipment To understand some of the HERCULES results we have analyzed the CCP voltage. CCP voltage at lower electrode CCP [W] CCP voltage [V] time [ns] Page 7

8 Additional tests with the HERCULES equipment The Fourier spectrum shows the main peak at MHz. The strong intensity at zero results from the DC offset. Harmonics can be found at higher CCP settings, but the condition of a sinusodial rf voltage is approximately fulfilled. CCP Voltage at Lower Electrode Voltage [V] CCP [W] ,56 27,12 40,68 54,24 67,8 81,36 94,92 108,48 122,04 135,6 Frequency [MHz] Page 8

9 Additional tests with the HERCULES equipment The Fourier spectrum of the SEERS current shows strong needle-shaped harmonics. Plasma self resonance frequencies could not be found between 13 and 500 MHz. Fourier Spectrum of SEERS current 000 Intensity = (Isensor) CCP=W CCP=W CCP=W ,0 40,7 81,4 122,0 162,7 203,4 244,1 Frequency [MHz] Page 9

10 Additional tests with the HERCULES equipment The enveloping curves of the Fourier spectrum look like the SEERS spectra in the small window displayed by older versions of HERC/PL. Enveloping curves of I(SEERS current) 000 Intensity = (Isensor) CCP=W CCP=W CCP=W CCP=W CCP=W CCP=W ,0 40,7 81,4 122,0 162,7 203,4 244,1 Frequency [MHz] Page 10

11 Additional tests with the HERCULES equipment Page 11! The HERCULES results on our etching tools differ from the results described by Michael Klick et. al. for the same gases.! It appears that tool construction parameters are the reason for the differences in our HERCULES results.! CCP spectrum analysis shows that the sinusodial rf voltage condition is fulfilled, therefore the different SEERS values must also be caused by mechanical parameters inside the etching tool.! Fourier spectrum analysis of the SEERS current do not show plasma self resonance frequencies independent from 13,56 MHz, so the electron density in the plasma cannot be calculated directly to verify the SEERS values. " Michael Klick, Wolfgang Rehak and Marita Kammeyer, Plasma Diagnostics in rf Discharges Using Nonlinear and Resonance Effects, Jpn. J. Appl. Phys. Vol. 36 (1997)

12 Analysis of CF 4 processes with SEERS at ZWG! This series of experiments was meant to show the usability of HERCULES in a semi-realistic etching process! The following experiments were run with a constant flow of mixed gases. The experiment series were made from different CCP power settings and variious gas mixtures of CF 4 and O 2. 1% of Ar was added for Optical Emission Spectroscopy (OES) purposes. This low portion of Ar has no significant impact on the HERCULES parameters: Collision frequency and electron density. Page 12

13 Analysis of CF 4 processes with SEERS at ZWG Etching process CF4(1-x%)+O2(x%)+Ar(1%) 3,50E+08 3,00E+08 Collision freq. [1/s] 2,50E+08 2,00E+08 1,50E+08 1,00E+08 0% O2 20% O2 40% O2 60% O2 80% O2 % O2 10% O2 90% O2 5,00E+07 0,00E Page 13

14 Analysis of CF 4 processes with SEERS at ZWG In general, the SEERS electron density is increasing with higher CCP power. However, the different trends at high O 2 levels have to be further examined. Etching process CF4(1-x%)+O2(x%)+Ar(1%) 1,20E+09 1,00E+09 Electron density [1/ccm] 8,00E+08 6,00E+08 4,00E+08 0% O2 20% O2 40% O2 60% O2 80% O2 % O2 10% O2 90% O2 Page 14 2,00E+08 0,00E

15 Analysis of CF 4 processes with SEERS at ZWG The DC bias seems to be a good indicator for the CCP power. The gas mixture has only minor impact. Etching process CF4(1-x%)+O2(x%)+Ar(1%) DC bias [V] % O2 20% O2 40% O2 60% O2 80% O2 % O2 10% O2 90% O Page 15

16 Analysis of CF 4 processes with SEERS at ZWG Collision Frequency vs. Gas Mixture 3,50E+08 3,00E+08 2,50E+08 Collision freq. [1/s] 2,00E+08 1,50E+08 CCP=W CCP=W CCP=W CCP=W CCP=500W 1,00E+08 5,00E+07 0,00E Gas mixture: F (1-x)sccm + O x sccm + Ar 1 sccm Page 16

17 Analysis of CF 4 processes with SEERS at ZWG Electron Density vs. Gas Mixture 1,20E+09 1,00E+09 Electron density [1/ccm] 8,00E+08 6,00E+08 4,00E+08 CCP=W CCP=W CCP=W CCP=W CCP=500W 2,00E+08 0,00E Gas mixture: F (1-x)sccm + O x sccm + Ar 1 sccm Page 17

18 Analysis of CF 4 processes with SEERS at ZWG The DC bias voltage determined in the match box is nearly independent from the gas mixture. Bias Voltage vs. Gas Mixture DC bias [V] CCP=W CCP=W CCP=W CCP=W CCP=500W Gas mixture: F (1-x)sccm + O x sccm + Ar 1 sccm Page 18

19 Analysis of CF 4 processes with SEERS at ZWG The electron density versus collision freq. ratio is nearly independent of rf power but depends on the gas mixture. Electron Density vs. Collision Freq. 1,20E+09 1,00E+09 Electron density [1/ccm] 8,00E+08 6,00E+08 4,00E+08 0% O2 10% O2 20% O2 40% O2 60% O2 80% O2 90% O2 % O2 Page 19 2,00E+08 0,00E+00 0,00E+00 5,00E+07 1,00E+08 1,50E+08 2,00E+08 2,50E+08 3,00E+08 3,50E+08 Collision freq. [1/s]

20 Analysis of CF 4 processes with OES at ZWG SpectraMax OES! 1024*128 px CCD matrix! Monochromator with 3 different gratings! GPIB Interface for control and measurement Optical Emission Spectroscopy! OES does not disturb the plasma process! OES is immune to HF problems! Deposited layers on the viewport disturb OES measurements! OES can deliver independend values to verify SEERS results Page 20

21 Analysis of CF 4 processes with OES at ZWG The intensity of atomic Fluorine increases with higher CCP and shows an interesting maximum at about 50% oxygen. The remaining F intensity at % O 2 results from desorption off the chamber wall. Etching process CF4(1-x%)+O2(x%)+Ar(1%) 1 0,9 0,8 Intensity ratio I F685/I Ar750 [1] 0,7 0,6 0,5 0,4 0,3 0% O2 20% O2 40% O2 60% O2 80% O2 % O2 10% O2 90% O2 0,2 0, Page 21

22 I F685 / I Ar750 intensity ratio Analysis of CF 4 processes with OES at ZWG The intensity of atomic Fluorine increases with higher CCP and shows an interesting maximum at about 50% oxygen. The remaining F intensity at % O 2 results from desorption off the chamber wall. 1,0 The x-axis shows the oxygen portion in the plasma. The fluor portion is % - x-value. 0, , O2 portion [%] Page 22

23 Analysis of CF 4 processes with OES at ZWG This diagramme shows the OES dependence of CCP to the oxygen concentration. Etching process CF4(1-x%)+O2(x%)+Ar(1%) 1,E+01 1,E+01 Intensity ratio I O777/I Ar750 [1] 1,000E+01 8,000E+00 6,000E+00 4,000E+00 0% O2 20% O2 40% O2 60% O2 80% O2 % O2 10% O2 90% O2 2,000E+00 0,000E Page 23

24 Analysis of CF 4 processes with OES at ZWG This diagramme shows the OES dependence of CCP to the oxygen concentration. Changing CCP between and 500W makes only little difference in the intensity ratio The x-axis shows the oxygen portion in the plasma. The fluor portion is % - x-value I O777 / I Ar750 intensity ratio O2 portion [%] Page 24

25 Comparison between SEERS and OES This diagramme shows the Fluorine intensity ratio I685/I703 as a function of the collision rate. Collision Frequency = f(f685/f703) 2,5 Intensity ratio I F685/ I F703 [1] 2 1,5 1 0,5 0%O2 10% 20% 40% 60% 80% 90% 0 0,00E+00 5,00E+07 1,00E+08 1,50E+08 2,00E+08 2,50E+08 Collision freq. [1/s] Page 25

26 Comparison between SEERS and OES The CCP power increases in the marked direction. The corner of the orange approximation line lies between and W. Collision Frequency = f(f685/f703) 2,5 Intensity ratio I F685/ I F703 [1] 2 1,5 1 0,5 0%O2 10% 20% 40% 60% 80% 90% 0 0,00E+00 5,00E+07 1,00E+08 1,50E+08 2,00E+08 2,50E+08 Collision freq. [1/s] Page 26

27 Comparison between SEERS and OES The direct comparison of CCP power and and the F685/F703 OES intensity ratio does not show this pronounced corner. Intensity ratio I F685/I F 703 = f(ccp) 2,25 2 Intensity ratio I F685/I F 703 [1] 1,75 1,5 1,25 1 0% O2 10% 20% 40% 60% 80% 90% 0,75 0, Page 27

28 Summary and Conclusion Page 28! Experimental HERCULES results of pure gases cannot be used on different etching tools without knowing the specific impact of tool characteristics on SEERS.! The collision frequency versus electron density ratio is a sentive measure for CF 4 and O 2 gas mixtures. The settings of the MFCs and the SEERS values have been verified with OES.! The OES measurements show a maximum in the concentration of atomic Fluorine when the O 2 fraction is at about 50%. HERCULES shows no representation for this effect.! CCP power settings below W cause a different trend for the collision frequency which is up to now not well understood.! Further experiments are necessary to understand the impact of the plasma system specifics on the SEERS model.