- Email: [email protected]
Incident angle dependence of metal etching using a gas cluster ion beam in acetic acid atmosphere
Incident angle dependence of metal etching using a gas cluster ion beam in acetic acid atmosphere Akihiro Ogawa, Noriaki Toyoda, Isao Yamada PII: DOI: Reference:
S0257-8972(16)30454-6 doi: 10.1016/j.surfcoat.2016.05.070 SCT 21225
To appear in:
Surface & Coatings Technology
Received date: Revised date: Accepted date:
25 December 2015 18 May 2016 23 May 2016
Please cite this article as: Akihiro Ogawa, Noriaki Toyoda, Isao Yamada, Incident angle dependence of metal etching using a gas cluster ion beam in acetic acid atmosphere, Surface & Coatings Technology (2016), doi: 10.1016/j.surfcoat.2016.05.070
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT Incident angle dependence of metal etching using a gas cluster ion beam in acetic acid atmosphere
PT
Akihiro Ogawa*, Noriaki Toyoda. Isao Yamada
RI
Graduate school of engineering, University of Hyogo, 2167 Shosha, Himeji, Hyogo, 671-2280,
SC
Japan
NU
* Corresponding author: Email: [email protected]; Abstract
MA
The dependence of Ru, CoFe, and SiO2 etching with O2-GCIB in an acetic acid atmosphere on the angle of incidence was studied. In the case where an acceleration voltage (Va) of 20 kV was
D
used, the depth to which Ru was etched using O2-GCIB with acetic acid decreased significantly
TE
with increasing incident angle. However, the etch depth produced with an acceleration voltage of
AC CE P
5 kV, with O2-GCIB in acetic acid, did not show a rapid decrease at high incident angles. The etch selectivities of Ru and CoFe to SiO2 at an incident angle of 70o and a Va of 5 kV with acetic acid were 7.6 and 10.4, respectively. This indicates that reactive etching occurred with these metals. Based on XPS and cross-sectional TEM, there was no observable damage to CoFe after O2-GCIB etching with acetic acid at an incident angle of 70o. Keywords: GCIB; acetic acid, STT-MRAM; oblique incidence
Abstract code: (given by the Organizing Committee)
ACCEPTED MANUSCRIPT 1. Introduction Recently, many forms of non-volatile memory have been proposed1-4). Among them, spin
PT
transfer torque magneto-resistive random access memory (STT-MRAM) offers many advantages
RI
such as low power consumption5), high read and write speeds6), and a large number of rewrite cycles. However, there are some issues in the process of fabricating STT-MRAM. Since STT-
SC
MRAM uses etch-resistant materials (Ru, Pt) and the magnetic tunnel junction (MTJ) structure7)
NU
is very sensitive to ion irradiation damage, a novel etching process is required to solve these issues. Currently, plasma etching and ion milling, which have high etch rates, are used for
MA
etching STT-MRAM. However, physical sputtering effects cause irradiation damage via energetic ions8-9) and re-deposition of metals on the sidewall. These problems may cause
TE
D
electrical shorts10) and degradation of device characteristics. We have proposed the gas cluster ion beam (GCIB) as a novel low-damage process, and have
AC CE P
reported that GCIB irradiation with acetic acid at normal incidence can be used to etch various metals such as Ru, Pt, and CoFe.11) It was also reported that magnetic materials irradiated with a glancing angle GCIB experienced less sub-surface damage than materials irradiated at normal incidence. Therefore, the level of irradiation damage on a STT-MRAM sidewall might be small with GCIB irradiation. Moreover, chemical etching occurs via GCIB irradiation in an acetic acid atmosphere. As a result, the thickness of re-deposited metals on the sidewall might be smaller than would be induced by physical sputtering. Thus, we think that it is beneficial to use GCIB for the removal of re-deposited metals from the sidewall. As we stated previously, we have already reported the effect of etching with a GCIB in an acetic acid atmosphere at normal incidence11). However, there is no information regarding the incident angle dependence of metal etching with a GCIB in acetic acid. In this study, the
ACCEPTED MANUSCRIPT dependence of etch effects on the angle of irradiation using an O2-GCIB in acetic acid were investigated as a function of acceleration voltage (Va) and the etch selectivities of CoFe and Ru
PT
to SiO2. The etch depth and modified layer thickness after oblique incidence GCIB were
RI
characterized based on etch depth measurements, XPS depth profiles, and cross-sectional TEM
SC
images.
NU
2. Experiment
Details of the GCIB equipment and its characteristics have already been reported in our
MA
previous paper.12) Figure 1 shows the GCIB irradiation equipment used in this study. Neutral oxygen clusters are produced by supersonic expansion of high pressure (2 MPa) oxygen through
TE
D
a nozzle. Neutral oxygen clusters are ionized by electron bombardment in the ionizer. After acceleration, molecular oxygen ions are removed via a permanent magnet, and only oxygen
AC CE P
cluster ions are transported to the target chamber. The average cluster size of O2-GCIB was three thousand molecules per cluster.13) In the target chamber, acetic acid gas is introduced using a needle valve. There is an aperture between the target and ionizing chamber to prevent contamination of the ionizing chamber with acetic acid gas. In the target chamber, O2-GCIB irradiation was carried out at incident angle of from the surface normal. The samples used were Ru, CoFe, and SiO2. Metal films (Ru, CoFe) were prepared via sputter deposition, and SiO2 was a thermally grown, 500 nm thick film. The incident angle ( and acceleration voltage (Va) of O2-GCIB were varied between 0-80o and 5-20 kV, respectively. We used molybdenum steel as a mask. After irradiation, the etch depth was measured using a surface profiler (Dektak 3 / ULVAC). From the etch depth measurement, and Va dependence
ACCEPTED MANUSCRIPT on etch depth and etch selectivity were studied. Ion fluence was varied between 1×1015 and 8×1016 ions / cm2. The partial pressure of acetic acid was 3×10-3 Pa.
PT
In order to characterize the modified layer thickness of the etched surface, X-ray photoelectron
RI
spectroscopy (XPS) (JEOL / JPS-9010MX) depth profile and cross-sectional transmission
SC
electron microscope (TEM) (HITACHI / HF-2200) observations were carried out on 16 nm thick CoFe films. For the XPS depth profile, 500eV Ar+ irradiation was performed on a CoFe sample
NU
to remove the native oxide layer, and subsequently, O2-GCIB irradiation was carried out. After
MA
GCIB etching, the XPS depth profile of peaks from O and Fe were obtained via 500 eV Ar+ ion sputtering. The O2-GCIB irradiation conditions were as follows: The source gas was O2-GCIB
D
and was 0o or 70o. Va was 5-20 kV. Ion fluence was varied between 1×1014 and 1×1015 ions /
AC CE P
TE
cm2. The partial pressure of acetic acid was 3×10-3 Pa.
3. Results and discussion
First, the dependence of the etch depth of Ru was studied, as shown in Fig. 2. was varied between 0o and 80o. Ion fluence was 1×1016 ion/cm2. In the case where was 0o and Va was 20 kV, the etch depth of Ru was doubled when acetic acid was introduced. Since the ion current of O2-GCIB was 1.6 µA, the etch rate of Ru was 13 nm/min. We also varied the ion fluence, but there was no change in etch rate. At 20 kV, the etch depth of Ru decreases significantly with increasing . The etch depth at 70o is almost 15% of the depth after irradiation at 0o, with and without acetic acid. When Va is 5 kV, the etch depth of Ru increases by almost 3 times after introduction of acetic acid, regardless of the value of within the range of 0o to 80o. At 70o, the etch depth reductions from the case of 0o were 81% (with acetic acid) and 97% (without acetic
ACCEPTED MANUSCRIPT acid), respectively. These results indicate that physical sputtering is dominant at a Va of 20 kV1416)
, and reactive sputtering is dominant at a Va of 5 kV.17)
PT
Next, material dependence was studied. Figure 3 shows the etch depths of CoFe, Ru, and
RI
SiO2 after etching with an O2-GCIB, with and without acetic acid, at of 0o and 70o. Va is 5 kV
SC
or 20 kV. The ion fluence is 1×1016 ions/cm2. The etch depths of CoFe and Ru with an O2-GCIB and acetic acid are about two or three times higher than those achieved without acetic acid, either
NU
at a Va of 5 kV or a Va of 20 kV. The etch rate of CoFe using a 20 kV O2-GCIB in acetic acid is 12 nm/min. In contrast, the etch depth of SiO2 does not show a clear increase with or without
MA
acetic acid under any conditions.
D
Table I shows the etch selectivities of CoFe and Ru / SiO2 at of 70o, with and without
TE
acetic acid. In the case of O2-GCIB irradiation without acetic acid, the etch selectivity is almost 1
AC CE P
regardless of Va. However, in those cases where acetic acid is used, the etch selectivities of CoFe/SiO2 increases to 5.4-10.4. In addition, the etch selectivities of Ru/SiO2 and CoFe/SiO2 increases from 7.6-10.4 at a Va of 5 kV. Since reactive etching of CoFe and Ru is dominant at 5 kV, the etch selectivities of these metals to SiO2 increased. SiO2 can be used as a mask material. Next, the dependence of CoFe modified layer and oxide layer thicknesses after O2-GCIB etching with acetic acid were characterized via a XPS depth profile. Figure 4 shows the oxide layer thickness on CoFe after irradiation using an O2-GCIB in various conditions. The oxide layer depth is defined as the thickness at which the oxygen peak disappears in XPS. The oxide layer thickness on pristine CoFe is 1.6 nm. When a Va of 20 kV is used with a of 0o, the oxide layer thickness is almost 11 nm without the acetic acid treatment. When acetic acid is added, it decreases to 7 nm. However, if the metal is irradiated with 20 kV O2-GCIB at 70o, the oxide
ACCEPTED MANUSCRIPT layer is almost identical to that of pristine CoFe. When Va is set to 5 kV, the thickness of the CoFe oxide layer is smaller than when the metal is irradiated at 20 kV. The thickness of the
PT
CoFe oxide layer without the acetic acid treatment at 0o is 6.5 nm. After addition of acetic acid, it
RI
decreases to 2.16 nm. When irradiated at an angle of incidence of 70o, without acetic acid, the CoFe oxide layer is 2.7 nm thick. When acetic acid is added, the thickness decreases to 1.08 nm.
SC
When we used oblique incidence O2-GCIB at 5 kV with acetic acid, the oxide layer became
NU
thinner than that of pristine CoFe. Therefore, irradiation with 5 kV O2-GCIB at incident angle of
MA
70o, and addition of acetic acid produced a thin oxide layer. Next, cross-sectional TEM and electron diffraction measurements were carried out to
D
observe cross sectional structure and crystalline structure. Ion fluence was 3×1015 ions / cm2.
TE
Figure 5 (a) shows a cross-sectional TEM image of a CoFe film after irradiation with O2-GCIB at a Va of 5 kV and of 0o, with acetic acid. Figure 5 (b) shows a CoFe film irradiated under the
AC CE P
same conditions, but with of 70o, and with acetic acid. Neither of the samples was heavily damaged. However, there is a very thin (~1 nm) modified layer on the sample irradiated at 0o. There is no such modified layer on the sample modified at 70o. Figure 5 (c) shows an electron diffraction pattern near the surface (at the white point in Fig 5 (b)). There is no difference in the diffraction pattern between the middle and near surface points, which indicates that there was no damage to the crystalline structure on the CoFe surface. From these results, it can be said that there was not a layer of modified material, or an oxide layer after 5 kV O2-GCIB etching with an incident angle of 70o.
4. Conclusion
ACCEPTED MANUSCRIPT The dependence of Ru, CoFe and SiO2 etching with an O2-GCIB in an acetic acid atmosphere on angle of incidence was studied. The dependence of CoFe, Ru, and SiO2 etch depth
PT
on angle of incidence was investigated in for several different values of Va. The etch depth produced by 5 kV O2-GCIB with acetic acid did not show a rapid decrease at high incident
RI
angles. The etch selectivities of Ru and CoFe to SiO2 at an incident angle of 70o, and Va of 5 kV
SC
with acetic acid were 7.6 and 10.4. Moreover, there was no modified layer after etching at an
NU
incident angle of 70o with acetic acid. It was found that oblique O2-GCIB irradiation with acetic
AC CE P
TE
D
MA
acid can be used as low-damage method to etch metals on sidewalls for STT-MRAM.
ACCEPTED MANUSCRIPT
TableⅠ. Etch selectivities of CoFe / SiO2 and Ru / SiO2 at of 70o with and without acetic acid.
With acetic acid
5 kV
0.6
7.6
20 kV
1.6
7.5
NU MA D TE AC CE P
CoFe
Without acetic acid
With acetic acid
RI
Without acetic acid
0.7
10.4
1.4
5.4
SC
Acceleration voltage
PT
Ru
ACCEPTED MANUSCRIPT
Sample
PT
Target chamber
NU
MA
Ionizing chamber
SC
RI
D
Source chamber
AC CE P
TE
Fig 1. Schematic diagram of the GCIB irradiation equipment
ACCEPTED MANUSCRIPT
180
PT
Va=20kV with AcOH
140
60 Va=5kV
40 with AcOH
Va=5kV without AcOH
20 0
10 20 30 40 50 60 70 80
MA
0
SC
100 Va=20kV 80 without AcOH
RI
120
NU
Etching depth (D) (nm)
O2-GCIB on Ru 2 1e16 ions/cm
Ru
160
Incident angle(degree)
D
Fig 2. Dependence of Ru etch depth, using O2-GCIB, on angle of incidence, with and
AC CE P
TE
without acetic acid. (Va: 5, 20 kV, Ion fluence: 1×1016 ions/cm2)
ACCEPTED MANUSCRIPT
200 160
20kV
RI SC
120
Gas :O2-GCIB Ion fluence : 1×1016 (ions/cm2)
NU
80 40 CoFe SiO2 SiO2 Ru CoFe Ru 0
=70o
SiO2 Ru CoFe SiO2
=0o
=70o
D
=0o
MA
Etching depth(nm)
Ru
PT
CoFe
5kV
TE
Fig 3. Etch depths of CoFe, Ru, and SiO2, using O2-GCIB with acetic acid at 0 and 70 degrees
AC CE P
(Va: 5 kV and 20 kV, Ion fluence: 1×1016 ions/cm2)
8
PT
20kV
O2-GCIB 20kV:1E14(ions/cm2)
O2-GCIB 5kV:1E15(ions/cm2)
RI
10
without AcOH
5kV
with AcOH
without AcOH
SC
12
NU
6 4
without with AcOH with AcOH AcOH
MA
Pristine 2 CoFe
0
D
Oxide thickness of CoFe [nm]
ACCEPTED MANUSCRIPT
AC CE P
TE
=0o
=70o =0o
without with AcOH AcOH
=70o
Fig 4. The oxide layer thickness of CoFe, irradiated by O2-GCIB, with and without acetic acid, measured with XPS. (Va: 5 kV and 20 kV, Ion fluence: 1×1014 and 1×1015 ions/cm2,
ACCEPTED MANUSCRIPT
θ=0o
θ=70o Au
A
CoFe 16nm
CoFe 16nm
PT
B
Au
✕ ✕
Ta 15nm
SC
RI
Ta 15nm
(b)
NU
(a)
MA
Fig 5. TEM figures of CoFe irradiated with O2-GCIB at a Va of 5 kV,a (b) . and electron diffraction patterns at point A (near surface) and B (at the center).
AC CE P
TE
D
(Ion fluence: 3×1015 ions / cm2.)
ACCEPTED MANUSCRIPT References
PT
1) G. Muller, T. Happ, M. Kund, G. Y. Lee, N. Nagel, and R. Sezi: IEDM Tech. Dig., 2004, p.567. 2) Z. Diao, Z. Li, S. Wang, Y. Ding, A. Panchula, E. Chen, L.C. Wang, and Y. Huai: J. Phys.: Condens. Matter, 19, 2007, 165209.
4) Y. Huai: AAPPS Bulletin, 18(6), 2008, p.33.
SC
RI
3) K. Yakushiji, K. Noma, T. Saruya, H. Kubota, A. Fukushima, T. Nagahama, S.Yuasa, and K. Ando: Appl. Phys. Express, 3, 2010, 053003.
NU
5) K.L. Wang, J.G. Alzate, and P.K. Amiri: J. Phys. D: Appl. Phys., 46, 2013, 074003.
MA
6) J.J. Nowak, R. P. Robertazzi, J. Z. Sun, G. Hu, David W. Abraham, P. L. Trouilloud, S. Brown, M. C. Gaidis, E. J. O’Sullivan, W. J. Gallagher, and D. C. Worledge: IEEE Magn. Lett., 2, 2011.
D
7) S. Ikeda, J. Hayakawa, Y. Ashizawa, Y. M. Lee, K. Miura, H. Hasegawa, M. Tsunoda, F. Matsukura, and H. Ohno: Appl. Phys. Lett., 93, 2008, 082508.
TE
8) J.C. Read, P.M. Braganca, N. Robertson, and J.R. Childress: APL Mater., 2, 2014, 046109.
AC CE P
9) K. Song and K. Lee: J. Appl. Phys., 118, 2015, 053912. 10) S.M. Hwang, A. Adalberto Garay, W. In Lee, and C. Won Chung: Thin Solid Films, 587, 2015, p.28-33. 11) A. Yamaguchi, R. Hinoura, N. Toyoda, K. Hara, and I. Yamada: Jpn. J. Appl. Phys, 52, 2013, 05EB05-2. 12) N. Toyoda and I. Yamada: IEEE Trans. Plasma. Sci. 36, 2008, 1471. 13) I. Yamada, J. Matsuo, Z. Insepov, D. Takeuchi, M. Akizuki, and N. Toyoda: J. Vac. Sci. Technol. A, 14, 1996, 781-785. 14) N. Toyoda, H. Kitani, N. Hagiwara, T. Aoki, J. Matsuo, and I. Yamada: Mater. Chem. Phys. 54, 1998, 262. 15) Z. Insepov and I. Yamada: Nucl. Instrum. Methods Phys. Res., Sect. B, 99, 1995, 248. 16) Z. Insepov and I. Yamada: Nucl. Instrum. Methods Phys. Res., Sect. B, 121, 1997, 44.
ACCEPTED MANUSCRIPT
AC CE P
TE
D
MA
NU
SC
RI
PT
17) N. Toyoda, N. Hagiwara, J. Matsuo, and I. Yamada: Nucl. Instrum. Methods Phys. Res., Sect. B, 148, 1999, 639.
ACCEPTED MANUSCRIPT
Cover page
PT
Research highlights
RI
► The dependence of O2-GCIB etching of Ru, CoFe, and SiO2, with acetic acid, on the angle of incidence was studied with the goal of removing re-deposited metals from the sidewall.
SC
► Physical sputtering is dominant at a Va of 20 kV and reactive sputtering is dominant at Va 5
NU
kV.
► The thickness of the oxide layer on CoFe that has been irradiated an angle of 70o at Va 5 kV
AC CE P
TE
D
MA
with acetic acid is lower than that of pristine CoFe.
S0257-8972(16)30454-6 doi: 10.1016/j.surfcoat.2016.05.070 SCT 21225
To appear in:
Surface & Coatings Technology
Received date: Revised date: Accepted date:
25 December 2015 18 May 2016 23 May 2016
Please cite this article as: Akihiro Ogawa, Noriaki Toyoda, Isao Yamada, Incident angle dependence of metal etching using a gas cluster ion beam in acetic acid atmosphere, Surface & Coatings Technology (2016), doi: 10.1016/j.surfcoat.2016.05.070
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT Incident angle dependence of metal etching using a gas cluster ion beam in acetic acid atmosphere
PT
Akihiro Ogawa*, Noriaki Toyoda. Isao Yamada
RI
Graduate school of engineering, University of Hyogo, 2167 Shosha, Himeji, Hyogo, 671-2280,
SC
Japan
NU
* Corresponding author: Email: [email protected]; Abstract
MA
The dependence of Ru, CoFe, and SiO2 etching with O2-GCIB in an acetic acid atmosphere on the angle of incidence was studied. In the case where an acceleration voltage (Va) of 20 kV was
D
used, the depth to which Ru was etched using O2-GCIB with acetic acid decreased significantly
TE
with increasing incident angle. However, the etch depth produced with an acceleration voltage of
AC CE P
5 kV, with O2-GCIB in acetic acid, did not show a rapid decrease at high incident angles. The etch selectivities of Ru and CoFe to SiO2 at an incident angle of 70o and a Va of 5 kV with acetic acid were 7.6 and 10.4, respectively. This indicates that reactive etching occurred with these metals. Based on XPS and cross-sectional TEM, there was no observable damage to CoFe after O2-GCIB etching with acetic acid at an incident angle of 70o. Keywords: GCIB; acetic acid, STT-MRAM; oblique incidence
Abstract code: (given by the Organizing Committee)
ACCEPTED MANUSCRIPT 1. Introduction Recently, many forms of non-volatile memory have been proposed1-4). Among them, spin
PT
transfer torque magneto-resistive random access memory (STT-MRAM) offers many advantages
RI
such as low power consumption5), high read and write speeds6), and a large number of rewrite cycles. However, there are some issues in the process of fabricating STT-MRAM. Since STT-
SC
MRAM uses etch-resistant materials (Ru, Pt) and the magnetic tunnel junction (MTJ) structure7)
NU
is very sensitive to ion irradiation damage, a novel etching process is required to solve these issues. Currently, plasma etching and ion milling, which have high etch rates, are used for
MA
etching STT-MRAM. However, physical sputtering effects cause irradiation damage via energetic ions8-9) and re-deposition of metals on the sidewall. These problems may cause
TE
D
electrical shorts10) and degradation of device characteristics. We have proposed the gas cluster ion beam (GCIB) as a novel low-damage process, and have
AC CE P
reported that GCIB irradiation with acetic acid at normal incidence can be used to etch various metals such as Ru, Pt, and CoFe.11) It was also reported that magnetic materials irradiated with a glancing angle GCIB experienced less sub-surface damage than materials irradiated at normal incidence. Therefore, the level of irradiation damage on a STT-MRAM sidewall might be small with GCIB irradiation. Moreover, chemical etching occurs via GCIB irradiation in an acetic acid atmosphere. As a result, the thickness of re-deposited metals on the sidewall might be smaller than would be induced by physical sputtering. Thus, we think that it is beneficial to use GCIB for the removal of re-deposited metals from the sidewall. As we stated previously, we have already reported the effect of etching with a GCIB in an acetic acid atmosphere at normal incidence11). However, there is no information regarding the incident angle dependence of metal etching with a GCIB in acetic acid. In this study, the
ACCEPTED MANUSCRIPT dependence of etch effects on the angle of irradiation using an O2-GCIB in acetic acid were investigated as a function of acceleration voltage (Va) and the etch selectivities of CoFe and Ru
PT
to SiO2. The etch depth and modified layer thickness after oblique incidence GCIB were
RI
characterized based on etch depth measurements, XPS depth profiles, and cross-sectional TEM
SC
images.
NU
2. Experiment
Details of the GCIB equipment and its characteristics have already been reported in our
MA
previous paper.12) Figure 1 shows the GCIB irradiation equipment used in this study. Neutral oxygen clusters are produced by supersonic expansion of high pressure (2 MPa) oxygen through
TE
D
a nozzle. Neutral oxygen clusters are ionized by electron bombardment in the ionizer. After acceleration, molecular oxygen ions are removed via a permanent magnet, and only oxygen
AC CE P
cluster ions are transported to the target chamber. The average cluster size of O2-GCIB was three thousand molecules per cluster.13) In the target chamber, acetic acid gas is introduced using a needle valve. There is an aperture between the target and ionizing chamber to prevent contamination of the ionizing chamber with acetic acid gas. In the target chamber, O2-GCIB irradiation was carried out at incident angle of from the surface normal. The samples used were Ru, CoFe, and SiO2. Metal films (Ru, CoFe) were prepared via sputter deposition, and SiO2 was a thermally grown, 500 nm thick film. The incident angle ( and acceleration voltage (Va) of O2-GCIB were varied between 0-80o and 5-20 kV, respectively. We used molybdenum steel as a mask. After irradiation, the etch depth was measured using a surface profiler (Dektak 3 / ULVAC). From the etch depth measurement, and Va dependence
ACCEPTED MANUSCRIPT on etch depth and etch selectivity were studied. Ion fluence was varied between 1×1015 and 8×1016 ions / cm2. The partial pressure of acetic acid was 3×10-3 Pa.
PT
In order to characterize the modified layer thickness of the etched surface, X-ray photoelectron
RI
spectroscopy (XPS) (JEOL / JPS-9010MX) depth profile and cross-sectional transmission
SC
electron microscope (TEM) (HITACHI / HF-2200) observations were carried out on 16 nm thick CoFe films. For the XPS depth profile, 500eV Ar+ irradiation was performed on a CoFe sample
NU
to remove the native oxide layer, and subsequently, O2-GCIB irradiation was carried out. After
MA
GCIB etching, the XPS depth profile of peaks from O and Fe were obtained via 500 eV Ar+ ion sputtering. The O2-GCIB irradiation conditions were as follows: The source gas was O2-GCIB
D
and was 0o or 70o. Va was 5-20 kV. Ion fluence was varied between 1×1014 and 1×1015 ions /
AC CE P
TE
cm2. The partial pressure of acetic acid was 3×10-3 Pa.
3. Results and discussion
First, the dependence of the etch depth of Ru was studied, as shown in Fig. 2. was varied between 0o and 80o. Ion fluence was 1×1016 ion/cm2. In the case where was 0o and Va was 20 kV, the etch depth of Ru was doubled when acetic acid was introduced. Since the ion current of O2-GCIB was 1.6 µA, the etch rate of Ru was 13 nm/min. We also varied the ion fluence, but there was no change in etch rate. At 20 kV, the etch depth of Ru decreases significantly with increasing . The etch depth at 70o is almost 15% of the depth after irradiation at 0o, with and without acetic acid. When Va is 5 kV, the etch depth of Ru increases by almost 3 times after introduction of acetic acid, regardless of the value of within the range of 0o to 80o. At 70o, the etch depth reductions from the case of 0o were 81% (with acetic acid) and 97% (without acetic
ACCEPTED MANUSCRIPT acid), respectively. These results indicate that physical sputtering is dominant at a Va of 20 kV1416)
, and reactive sputtering is dominant at a Va of 5 kV.17)
PT
Next, material dependence was studied. Figure 3 shows the etch depths of CoFe, Ru, and
RI
SiO2 after etching with an O2-GCIB, with and without acetic acid, at of 0o and 70o. Va is 5 kV
SC
or 20 kV. The ion fluence is 1×1016 ions/cm2. The etch depths of CoFe and Ru with an O2-GCIB and acetic acid are about two or three times higher than those achieved without acetic acid, either
NU
at a Va of 5 kV or a Va of 20 kV. The etch rate of CoFe using a 20 kV O2-GCIB in acetic acid is 12 nm/min. In contrast, the etch depth of SiO2 does not show a clear increase with or without
MA
acetic acid under any conditions.
D
Table I shows the etch selectivities of CoFe and Ru / SiO2 at of 70o, with and without
TE
acetic acid. In the case of O2-GCIB irradiation without acetic acid, the etch selectivity is almost 1
AC CE P
regardless of Va. However, in those cases where acetic acid is used, the etch selectivities of CoFe/SiO2 increases to 5.4-10.4. In addition, the etch selectivities of Ru/SiO2 and CoFe/SiO2 increases from 7.6-10.4 at a Va of 5 kV. Since reactive etching of CoFe and Ru is dominant at 5 kV, the etch selectivities of these metals to SiO2 increased. SiO2 can be used as a mask material. Next, the dependence of CoFe modified layer and oxide layer thicknesses after O2-GCIB etching with acetic acid were characterized via a XPS depth profile. Figure 4 shows the oxide layer thickness on CoFe after irradiation using an O2-GCIB in various conditions. The oxide layer depth is defined as the thickness at which the oxygen peak disappears in XPS. The oxide layer thickness on pristine CoFe is 1.6 nm. When a Va of 20 kV is used with a of 0o, the oxide layer thickness is almost 11 nm without the acetic acid treatment. When acetic acid is added, it decreases to 7 nm. However, if the metal is irradiated with 20 kV O2-GCIB at 70o, the oxide
ACCEPTED MANUSCRIPT layer is almost identical to that of pristine CoFe. When Va is set to 5 kV, the thickness of the CoFe oxide layer is smaller than when the metal is irradiated at 20 kV. The thickness of the
PT
CoFe oxide layer without the acetic acid treatment at 0o is 6.5 nm. After addition of acetic acid, it
RI
decreases to 2.16 nm. When irradiated at an angle of incidence of 70o, without acetic acid, the CoFe oxide layer is 2.7 nm thick. When acetic acid is added, the thickness decreases to 1.08 nm.
SC
When we used oblique incidence O2-GCIB at 5 kV with acetic acid, the oxide layer became
NU
thinner than that of pristine CoFe. Therefore, irradiation with 5 kV O2-GCIB at incident angle of
MA
70o, and addition of acetic acid produced a thin oxide layer. Next, cross-sectional TEM and electron diffraction measurements were carried out to
D
observe cross sectional structure and crystalline structure. Ion fluence was 3×1015 ions / cm2.
TE
Figure 5 (a) shows a cross-sectional TEM image of a CoFe film after irradiation with O2-GCIB at a Va of 5 kV and of 0o, with acetic acid. Figure 5 (b) shows a CoFe film irradiated under the
AC CE P
same conditions, but with of 70o, and with acetic acid. Neither of the samples was heavily damaged. However, there is a very thin (~1 nm) modified layer on the sample irradiated at 0o. There is no such modified layer on the sample modified at 70o. Figure 5 (c) shows an electron diffraction pattern near the surface (at the white point in Fig 5 (b)). There is no difference in the diffraction pattern between the middle and near surface points, which indicates that there was no damage to the crystalline structure on the CoFe surface. From these results, it can be said that there was not a layer of modified material, or an oxide layer after 5 kV O2-GCIB etching with an incident angle of 70o.
4. Conclusion
ACCEPTED MANUSCRIPT The dependence of Ru, CoFe and SiO2 etching with an O2-GCIB in an acetic acid atmosphere on angle of incidence was studied. The dependence of CoFe, Ru, and SiO2 etch depth
PT
on angle of incidence was investigated in for several different values of Va. The etch depth produced by 5 kV O2-GCIB with acetic acid did not show a rapid decrease at high incident
RI
angles. The etch selectivities of Ru and CoFe to SiO2 at an incident angle of 70o, and Va of 5 kV
SC
with acetic acid were 7.6 and 10.4. Moreover, there was no modified layer after etching at an
NU
incident angle of 70o with acetic acid. It was found that oblique O2-GCIB irradiation with acetic
AC CE P
TE
D
MA
acid can be used as low-damage method to etch metals on sidewalls for STT-MRAM.
ACCEPTED MANUSCRIPT
TableⅠ. Etch selectivities of CoFe / SiO2 and Ru / SiO2 at of 70o with and without acetic acid.
With acetic acid
5 kV
0.6
7.6
20 kV
1.6
7.5
NU MA D TE AC CE P
CoFe
Without acetic acid
With acetic acid
RI
Without acetic acid
0.7
10.4
1.4
5.4
SC
Acceleration voltage
PT
Ru
ACCEPTED MANUSCRIPT
Sample
PT
Target chamber
NU
MA
Ionizing chamber
SC
RI
D
Source chamber
AC CE P
TE
Fig 1. Schematic diagram of the GCIB irradiation equipment
ACCEPTED MANUSCRIPT
180
PT
Va=20kV with AcOH
140
60 Va=5kV
40 with AcOH
Va=5kV without AcOH
20 0
10 20 30 40 50 60 70 80
MA
0
SC
100 Va=20kV 80 without AcOH
RI
120
NU
Etching depth (D) (nm)
O2-GCIB on Ru 2 1e16 ions/cm
Ru
160
Incident angle(degree)
D
Fig 2. Dependence of Ru etch depth, using O2-GCIB, on angle of incidence, with and
AC CE P
TE
without acetic acid. (Va: 5, 20 kV, Ion fluence: 1×1016 ions/cm2)
ACCEPTED MANUSCRIPT
200 160
20kV
RI SC
120
Gas :O2-GCIB Ion fluence : 1×1016 (ions/cm2)
NU
80 40 CoFe SiO2 SiO2 Ru CoFe Ru 0
=70o
SiO2 Ru CoFe SiO2
=0o
=70o
D
=0o
MA
Etching depth(nm)
Ru
PT
CoFe
5kV
TE
Fig 3. Etch depths of CoFe, Ru, and SiO2, using O2-GCIB with acetic acid at 0 and 70 degrees
AC CE P
(Va: 5 kV and 20 kV, Ion fluence: 1×1016 ions/cm2)
8
PT
20kV
O2-GCIB 20kV:1E14(ions/cm2)
O2-GCIB 5kV:1E15(ions/cm2)
RI
10
without AcOH
5kV
with AcOH
without AcOH
SC
12
NU
6 4
without with AcOH with AcOH AcOH
MA
Pristine 2 CoFe
0
D
Oxide thickness of CoFe [nm]
ACCEPTED MANUSCRIPT
AC CE P
TE
=0o
=70o =0o
without with AcOH AcOH
=70o
Fig 4. The oxide layer thickness of CoFe, irradiated by O2-GCIB, with and without acetic acid, measured with XPS. (Va: 5 kV and 20 kV, Ion fluence: 1×1014 and 1×1015 ions/cm2,
ACCEPTED MANUSCRIPT
θ=0o
θ=70o Au
A
CoFe 16nm
CoFe 16nm
PT
B
Au
✕ ✕
Ta 15nm
SC
RI
Ta 15nm
(b)
NU
(a)
MA
Fig 5. TEM figures of CoFe irradiated with O2-GCIB at a Va of 5 kV,a (b) . and electron diffraction patterns at point A (near surface) and B (at the center).
AC CE P
TE
D
(Ion fluence: 3×1015 ions / cm2.)
ACCEPTED MANUSCRIPT References
PT
1) G. Muller, T. Happ, M. Kund, G. Y. Lee, N. Nagel, and R. Sezi: IEDM Tech. Dig., 2004, p.567. 2) Z. Diao, Z. Li, S. Wang, Y. Ding, A. Panchula, E. Chen, L.C. Wang, and Y. Huai: J. Phys.: Condens. Matter, 19, 2007, 165209.
4) Y. Huai: AAPPS Bulletin, 18(6), 2008, p.33.
SC
RI
3) K. Yakushiji, K. Noma, T. Saruya, H. Kubota, A. Fukushima, T. Nagahama, S.Yuasa, and K. Ando: Appl. Phys. Express, 3, 2010, 053003.
NU
5) K.L. Wang, J.G. Alzate, and P.K. Amiri: J. Phys. D: Appl. Phys., 46, 2013, 074003.
MA
6) J.J. Nowak, R. P. Robertazzi, J. Z. Sun, G. Hu, David W. Abraham, P. L. Trouilloud, S. Brown, M. C. Gaidis, E. J. O’Sullivan, W. J. Gallagher, and D. C. Worledge: IEEE Magn. Lett., 2, 2011.
D
7) S. Ikeda, J. Hayakawa, Y. Ashizawa, Y. M. Lee, K. Miura, H. Hasegawa, M. Tsunoda, F. Matsukura, and H. Ohno: Appl. Phys. Lett., 93, 2008, 082508.
TE
8) J.C. Read, P.M. Braganca, N. Robertson, and J.R. Childress: APL Mater., 2, 2014, 046109.
AC CE P
9) K. Song and K. Lee: J. Appl. Phys., 118, 2015, 053912. 10) S.M. Hwang, A. Adalberto Garay, W. In Lee, and C. Won Chung: Thin Solid Films, 587, 2015, p.28-33. 11) A. Yamaguchi, R. Hinoura, N. Toyoda, K. Hara, and I. Yamada: Jpn. J. Appl. Phys, 52, 2013, 05EB05-2. 12) N. Toyoda and I. Yamada: IEEE Trans. Plasma. Sci. 36, 2008, 1471. 13) I. Yamada, J. Matsuo, Z. Insepov, D. Takeuchi, M. Akizuki, and N. Toyoda: J. Vac. Sci. Technol. A, 14, 1996, 781-785. 14) N. Toyoda, H. Kitani, N. Hagiwara, T. Aoki, J. Matsuo, and I. Yamada: Mater. Chem. Phys. 54, 1998, 262. 15) Z. Insepov and I. Yamada: Nucl. Instrum. Methods Phys. Res., Sect. B, 99, 1995, 248. 16) Z. Insepov and I. Yamada: Nucl. Instrum. Methods Phys. Res., Sect. B, 121, 1997, 44.
ACCEPTED MANUSCRIPT
AC CE P
TE
D
MA
NU
SC
RI
PT
17) N. Toyoda, N. Hagiwara, J. Matsuo, and I. Yamada: Nucl. Instrum. Methods Phys. Res., Sect. B, 148, 1999, 639.
ACCEPTED MANUSCRIPT
Cover page
PT
Research highlights
RI
► The dependence of O2-GCIB etching of Ru, CoFe, and SiO2, with acetic acid, on the angle of incidence was studied with the goal of removing re-deposited metals from the sidewall.
SC
► Physical sputtering is dominant at a Va of 20 kV and reactive sputtering is dominant at Va 5
NU
kV.
► The thickness of the oxide layer on CoFe that has been irradiated an angle of 70o at Va 5 kV
AC CE P
TE
D
MA
with acetic acid is lower than that of pristine CoFe.