Surface contamination control during plasma etching
`
`H. Miyatake, K. Kawai, N. Fujiwara, M. Yoneda,
`K. Nishioka” and H. Abe
`
`LSI Laboratory, Mitsubishi Electric Corporation,
`4-1 Mizuhara,
`Itami, Hyogo 664,
`Japan
`+Kita-Itami Works, Mitsubishi Electric Corporation,
`4-1 Mizuhara,
`Itami, Hyogo 664, Japan
`
`ABSTRACT
`
`in
`is developed by employing NF 3 gas
`Reactive ion etching (RIE)
`order
`to avoid the fluorocarbon contamination on
`the Si
`surface ex-
`posed to the plasma. A high SiOy etch rate is achieved with magneti-
`cally enhanced RIE
`because of efficient species generation. An aniso-
`tropic etching profile of SiOg is obtained due to the low pressure and
`low temperature operation. The
`reaction layers
`on Si
`surfaces
`are
`investigated by x-ray photoelectron ‘spectroscopy and cross-sectional
`transmission electron microscopy.
`It
`is
`found
`that
`the NFg plasma
`etching is more effective to maintain a clean surface than the CHF 3
`plasma etching.
`In addition the photoresist which is used as
`a mask
`during via-hole etching is easily removed without
`any residues by Oo
`plasma ashing because the fluorocarbon contamination is avoided.
`
`1. INTRODUCTION
`
`surface during plasma etching is
`a Si
`of
`Contamination control
`a high quality device. Fluorocarbon plas-
`required for
`fabrication of
`mas
`have been: studied extensively and
`have
`been used
`for
`etching
`polysilicon and silicon oxide. However, device degradation is caused
`by contamination originating from fluorocarbon deposition during dry
`etching.
`The polymer deposition on
`the silicon surface
`in an NF3
`discharge is minimal compared to etching in a
`fluorocarbon plasma. The
`NF2 gas plasma
`is often used
`for
`surface cleaning treatment after
`conventional
`reactive ion etching (RIE)
`to remove
`the
`fluorocarbon
`film deposited from the reactive gas plasma.%
`In this work,
`an RIE process was developed by employing NF3 gas. A
`high SiQg etch rate was achieved and an anisotropic etching profile
`was
`formed in the NFg plasma with magnetically enhanced RIE (MERIE).
`The
`composition of
`reaction layers
`on Si
`surfaces
`exposed
`to
`the
`plasma was
`investigated by x-ray photoelectron Spectroscopy (XPS). The
`surface quality of
`the Si substrate was also characterized by cross-
`sectional
`transmission electron microscopy (TEM)
`in more detail.
`
`2.EXPERIMENTAL
`
`The MERIE system was used in this study. The wafers were clamped to
`the rf powered electrode. Helium back side cooling was used to main-
`
`0-8194-0724-0/92/84.00
`
`IP Bridge Exhibit 2223
`IP Bridge Exhibit 2223
`TSMC v. Godo Kaisha IP Bridge 1
`TSMCv. Godo KaishaIP Bridge 1
`IPR2017-01843
`IPR2017-01843
`SPIE Vol. 1593 Dry Etch Technology (1991) / 47
`
`

`

`temperature. Clean (100) Si wafers were etched
`tain a constant wafer
`with 700 W of
`rf
`power applied using NF2
`or CHF.
`for
`60
`sec. The
`chamber pressure was
`50 mTorr
`and the electrode temperature was 5°C.
`The other process parameters such as gas
`flow (40 scem)
`and magnetic
`field strength (90 G) were held constant
`in this experiment.
`The surface of
`the wafers was analyzed using XPS. The XPS spectra
`were excited with Mg Kg x-ray at
`10
`kV. For
`TEM observation (011)
`cross-sectional
`specimens were prepared. The observation of
`surface
`profile imaging was carried out with a high-resolution electron micro-
`scope operated at 200 kV.
`thermal-
`Si09 etch rates were measured on samples that consisted of
`ly grown SiQyg
`layer
`on
`a silicon substrate with a photoresist mask.
`The SiQg profiles
`etched by NF,
`gas were observed using scanning
`electron microscopy (SEM). The chamber pressure was varied from 20 to
`200 mTorr
`and the electrode temperature was varied from -50 to 20°C.
`Finally,
`the surface residues of
`the samples after via-hole etching at
`low pressure (20 mTorr) and low temperature (-50°C) were investigated
`by
`SEM. The
`samples were observed after
`the
`resist
`removal
`by Og
`plasma ashing.
`
`3.RESULTS AND DISCUSSION
`
`sample etched by NFg and CHFg gas are shown
`the Si
`XPS spectra of
`in Fig.l. The intensities of
`the Si 2s and 2p peaks are strong for
`the
`sample etched in NFg. For
`the sample etched in CHF,
`they are weak. For
`the NFg etched sample peaks due to O Auger, O 1s, F Auger and F 1s are
`observed,
`and the C ls peak is very smal!. For
`the CHF, etched sample
`the peaks due
`to oxygen appear
`to be smaller, while the peaks due
`to
`fluorine and the C ls peak is much larger. This results can be simply
`explained by the formation of
`a
`fluorocarbon layer
`on
`the Si
`surface
`after etching in CHF,g. On
`the other hand
`there is
`an oxide
`layer
`which contains F atoms on the surface exposed in NFa.
`In Fig.2 cross-sectional TEM images of NFg2 and CHF 3 plasma-exposed
`Si surface are shown.
`It
`shows
`(200)
`lattice planes parallel
`to the Si
`surface. The
`(200) plane spacing is 0.27 nm. For
`the NF3z etched sample
`there are no defects in evidence and the Si surface is smooth within a
`few monolayers
`[Fig.2(a)]. Extensive defects are found in the surface
`layer after etching in CHFg
`[Fig.2 (b)]. Small amorphouslike regions
`and a high density of planer defects are observed. They are heavily
`decorated by
`impurities, possibly H,
`C or F.> Figure 2{a)
`shows
`the
`a
`presence of
`2
`nm thick amorphous
`film (indicated by arrows) which
`was
`found by XPS
`to contain mostly oxygen and
`thus
`represents
`the
`native oxide on the surface exposed in NF3. As
`shown in Fig-.2(b)
`a
`3
`nm thick amorphous
`film (also indicated by
`arrows)
`on
`the
`surface
`etched by CHF,
`is observed. This
`film was
`identified to be
`a
`fluoro-
`carbon film by XPS.
`Figure 3
`shows
`the etch rates of SiO,
`in NF,
`as
`a
`function of
`pressure.
`It
`is noted that
`the SiO, etch rate does not
`change very
`much with pressure. A high SiO»5 etch rate is achieved with the MERIE
`because of efficient species generation. Figure 4
`shows
`the etch rates
`of SiO,
`as
`a
`function of
`temperature. The SiO, etch rate slightly
`
`48 / SPIE Vol. 1593 Dry Etch Technology (1991)
`
`

`

`keeps nearly constant
`
`it
`but
`temperature
`lowering
`increases with
`within the temperature range from -50 to 20°C.
`by Os
`removal
`resist
`SEM micrographs
`of
`the
`samples after
`the
`the NFg pressure
`ashing are shown in Fig.5. For
`the sample etched at
`of
`200 mTorr
`and
`the
`temperature of
`5 °C,
`the
`sidewall profile is
`slightly bowed with a positive taper
`[Fig.5{a)]. As
`the pressure is
`decreased to 20 mTorr,
`the bowed
`feature of
`the sidewall! disappears
`and the profile exhibits a slightly positive slope [Fig.5(b)]. For
`the
`sample etched at
`the temperature of
`-50°C and the NFx. pressure of
`20
`mTorr,
`the straight
`sidewall
`is produced [Fig.5(c)]. An anisotropic
`etching profile is
`formed due
`to
`low pressure and
`low temperature
`operation. Etching at higher pressures tends towards chemical process-
`es where ion energies are lower and the density of reactive species is
`higher. Etching at
`lower pressures emphasizes physical processes and
`etching at
`lower
`temperatures further enhances them. Figure 5(d)
`shows
`the cross-sectional view of the sample etched in CHF, at
`the tempera-
`ture of
`-50°C and the pressure of
`20 mTorr. The
`tapered sidewall
`is
`formed. The
`tapered etching profile is attained by
`the
`simultaneous
`progress of etching and deposition. The deposition of
`a
`fluorocarbon
`film on the sidewall
`induces the tapered profile formation at
`the low
`temperature. © During SiO, etching in NF3
`the
`polymer
`film is
`not
`deposited
`on
`the
`sidewall
`for
`lack
`of deposition gas
`such
`as
`in
`hydrocarbon.
`For
`the
`sample
`etched
`NF3g
`on
`the
`same
`etching
`condition as CHF 3.
`therefore,
`the straight sidewall
`is obtained.
`the
`Figure 6
`shows
`the profile of
`the via-hole etched in NFg at
`temperature of
`-50°C and the pressure of
`20 mTorr. The
`samples were
`treated in HF solution before the plasma etching in NFg. Undercutting
`by the HF treatment
`is observed at
`the interface between the photore-
`sist
`and the SiO,
`film. The surface residues of
`the samples etched in
`NFz and CHF, after the resist
`removal by Oy ashing are shown in Fig.7.
`For
`the sample etched in CHF
`the residual
`films still
`remain around
`the via-holes
`[Fig.7(b)].
`The
`fluorocarbon film which contains Al
`atoms sputtered during overetching was
`redeposited on the sidewall and
`the surface of
`the photoresist. On
`the sample etched in NF3
`there are
`no
`residues
`[Fig.7(a)],
`because
`the
`fluorocarbon contamination
`is
`avoided.
`
`4.CONCLUSION
`
`substrate induced during plasma
`surface modification in Si
`The
`etching was
`studied by XPS
`and cross-sectional
`TEM.
`For
`pure NF 2
`etching gas,
`a native oxide film 2
`nm thick grown on
`the Si
`surface
`was observed and there were no extended surface defects.
`It was
`found
`that
`the NF3g plasma etching was effective to maintain a clean surface
`as compared to the sample exposed to the CHF plasma.
`In addition,
`the
`photoresist used during via-hole etching was easily removed without
`any residues by Oo ashing because the fluorocarbon contamination was
`avoided. A clean RIE process was developed by employing the NF g gas. A
`high SiO g etch rate was achieved and an anisotropic etching profile
`was obtained in the NFg plasma with the MERIE.
`
`SPIE Vol. 1593 Dry Etch Technology (1991)/ 49
`
`

`

`5. REFERENCES
`
`1. Y. H. Lee, G. S. Oehriein and C. Ranson, “RIE-Induced Damage and
`Contamination in Silicon,” Radiation Effects
`and Defects
`in Solids,
`vol.111&@112, pp.221-232, 1989.
`2. H. Cerva, E. G. Mohr
`and H. Oppolzer, “Transmission Electron
`Microscope Study of Lattice Damage
`and Polymer Coating Formed after
`Reactive Jon Etching of Si09,” J. Vac. Sci. Technol., vol.B5, no.2,
`pp-590-593, 1987.
`“Removal of RIE Induced
`3. T. Akimoto, K. Kasama and M. Sakamoto,
`Damage Layer Using NFa/Q, Chemical Dry Etching,” Proc. of 10th Sympo-
`3°
`<2
`sium on Dry Process, pp.92-97, 1988.
`Ito, M. Yoneda and K. Nishioka, “Si Sur-
`4. T. Ogawa, K. Kawai, H.
`face Cleaning Using NFg after Glow Plasma
`and Deep UV Irradiation,”
`Proc. of
`llth Symposium on Dry Process, pp.94-99, 1989.
`5. S.
`J.
`Jeng and G. S. Oehrlein, “Microstructural Studies of Reac-
`-
`tive lon Etched Silicon,’
`Appl. Phys. Lett., vol.50, no.26, pp.1912-
`1914, 1987.
`and H.
`I. Hasegawa
`6. T. Ohiwa, K. Horioka, T. Arikado,
`“SiOg Tapered etching employing Magnetron Discharge,” Proc.
`Symposium on Dry Process, pp.105-109, 1990.
`
`Okano,
`of 12th
`
`
`
`INTENSITY(a.u.)
`
`Si 2s
`
`Si 2p
`
`F Auger
`
`1000
`
`800
`
`600
`
`400
`
`200
`
`0
`
`Fig.l. XPS spectra of NF 3 and CHF, etched Si samples.
`
`BINDING ENERGY(eV)
`
`50 / SPIE Vol. 1593 Dry Etch Technology (1991)
`
`

`

`
`
`(b)
`
`Fig.2. TEM cross sections of Si surface etched in (a) NF, and (b) CHF 3
`plasmas.
`
`SPIE Vol. 1593 Dry Etch Technology (1991) / 51
`
`

`

`5 °C.
`
`in a
`Fig.3. SiO, etch rates
`NF, plasma
`as
`a
`function of
`chamber
`pressure.
`The
`elec-
`trode temperature was
`
`
`
`ETCHRATE(nm/min)
`
`200
`
`10
`
`100
`
`PRESSURE (mTorr)
`
`400 —o —
`
`o~
`
`200
`
`Oo
`
`in a
`SiQg etch rates
`Fig.4.
`NFz plasma
`as
`a
`function of
`electrode
`temperature.
`The
`chamber pressure was 20 mTorr.
`
`-50
`
`0
`
`50
`
`TEMPERATURE(°C)
`
`=~
`£ca
`Lu
`
`-xr
`
`Ok
`
`-
`uu
`
`52 / SPIE Vol. 1893 Dry Etch Technology (1991)
`
`

`

`
`
`
`(a)
`
`|
`
`(b)
`
`(c)
`
`|
`
`(d)
`
`samples were
`SEM cross sections of SiO g sidewall! profiles. The
`Fig.5.
`etched at
`(a)
`the NF. pressure of 200 mTorr
`and the temperature of
`5
`°C,
`(b)
`the NFz pressure of
`20 mTorr
`and the temperature of 5°C,
`(c)
`the NFz pressure of 20 mTorr and the temperature of
`-50°C, and (d)
`the
`CHF, pressure of 20 mTorr and the temperature of
`-50°C.
`
`SPIE Vol. 1593 Dry Etch Technology (1991}/ 53
`
`

`

`
`
`
`SEM cross section of via-hole etched in NFg plasma.
`
`Fig.6.
`
`SEM micrographs of surface residues after the resist
`Fig.7.
`Oy plasma ashing. The
`samples were etched in
`(a) NF 3
`and
`plasma.
`
`removal by
`(b) CHF3
`
`54 / SPIE Vol. 1593 Dry Etch Technology (1991)
`
`

`

`DRY ETCH TECHNOLOGY
`
`Volume 1593
`
`SESSION 2
`
`Dry Etch Requirements
`for Advanced Photoresist
`
`Chair
`G. Ken Herb
`AT&TBell Laboratories
`
`