Heat, moisture and chemical resistance on low dielectric constant (low-k) film using diethoxymethylsilane (DEMS) prepared by plasma enhanced chemical vapor deposition

Heat, moisture and chemical resistance on low dielectric constant

(low-k) film using Diethoxymethylsilane (DEMS) prepared

by plasma enhanced chemical vapor deposition

Y.L. Cheng

*, Y.L. Wang

, J.K. Lan

, G.J. Hwang

, M.L. O’Neil

, C.F. Chen

College of Science and Engineering, National University of Tainan, Taiwan b

Air Product and Chemical, Inc., USA

c_{Department of Materials Science and Engineering, National Chiao-Tung University, Hsin-Chu, Taiwan} Available online 16 August 2005

Abstract

Resistance of low dielectric constant (low-k) dielectrics, deposited using Diethoxymethylsilane (DEMS) precursor and helium (He) carrier gas with or without oxygen (O2) reaction gas, against heat, moisture stress and chemical treatment is clarified. The low dielectric constant organosilicate glass (OSG) films deposited using DEMS and O2is shown to be the most reliable: the dielectric constant are stable even after a heating test at 700-C and a pressure cooker test (PCT) for 168 h. This stability is high enough to ensure the low-k properties throughout fabricating multilevel interconnects and long-term reliability after the fabrication. This is due to the stability of Si – CH3bonds and more Si – C – Si – bonds, which has high degree of cross-linking. However, the degradation of the dielectric constant occurs after O2plasma ashing process. The nitrogen plasma treatment is proposed to prevent the damage from O2 attack in the low-k films deposited using DEMS precursor.

D 2005 Elsevier B.V. All rights reserved.

Keywords: Low dielectric constant; Diethoxymethylsilane; Organosilicate glass; Thermal stability

1. Introduction

As minimum device features shrink below 180 nm, the increase in propagation delay, the resistance and capacitance delay (RC) of the interconnect has become a limiting factor in ultra-large scale integration (ULSI) device performance. Since RC delay is a product of the resistance in the metal interconnect (R) and the capacitance between the metal line (C), incorporating copper (Cu) wiring and low-k dielectric to

replace the conventional AlCu/SiO2into interconnect

tech-nology can effectively reduce the RC delay[1 – 3].

Various low dielectric constant materials have been proposed to decrease the time delay caused by capacitance. Recently, organosilicate glass (OSG), deposited by Plasma-Enhanced Chemical Vapor Deposition (PECVD) using various organo-precursors, such as Methylsilane (MS),

Tetramethylsilanetetrasiloxane (TOMCATS) and Diethoxy-methylsilane (DEMS), is the most promising low-k dielectric

candidate[4 – 8]. Among these precursors, DEMS precursor

is a strong candidate based on the excellent film properties.

DEMS (H – Si(CH3)(OC2H5)2)-based low-k films not only

have a lower dielectric constant (k = 2.8 – 3.0) but also have a greater hardness (higher cross link) as it contains an O/Si

ratio of 2:1 in the precursor produced optimum films[9].

However, during the interconnect fabrication process, thermal cycle and photoresist stripping are the indispensable steps. Therefore, the physical (thickness and refractive index) and electrical properties (dielectric constant and leakage current) of the low-k interconnect dielectric are needed to be resistance against heat, moisture, and chemical stress in order to prevent the degradation during the interconnect fabrication process. Furthermore, to ensure the high reliability performance of the high-speed ULSI, the moisture resistance is essential to avert moisture penetrating

from the outside the package [10 – 12].

0257-8972/$ - see front matterD 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.surfcoat.2005.07.014

* Corresponding author. Tel.: +886 6 505 9688; fax: +886 6 5051262. E-mail address: ylwang@tsmc.com (Y.L. Cheng).

In this work, the stability of the low-k films deposited using DEMS precursor against heat, moisture and chemical stresses is clarified. The low-k films prepared using DEMS

or DEMS/O2reactant gas are submitted to reliability tests to

distinguish the stability divergence for the effect of the

addition of O2. Furthermore, the electrical measurements

and material analyses have also been used to evaluate the low-k film before and after the reliability tests.

2. Experimental 2.1. Material prepared

All thin film deposition was performed on an Applied Materials Producer system with a 200 mm Producer chamber. The thin films were deposited on p-type (100) silicon substrates by radio frequency (13.56 MHz) PECVD with Diethoxymethylsilane (DEMS, CVD precursor) carried to the reaction chamber in the vapor phase by inert helium (He) gas. The chamber pressure and RF power were maintained at 800 Pa and 700 W, respectively, throughout the deposition process. The deposition temperature and He flow were kept

400-C, and 150 sccm, respectively. The DEMS flow rate was

fixed at 1500 mg and oxygen (O2) flow was varied from 0 to

250 sccm (herein the O2/DEMS ratio was 0 – 0.175).

2.2. Reliability test

To determine the thermal stability, the films were annealed for 1 h in a nitrogen ambient at temperatures

ranging from 400 to 800 -C. Moreover, to mimic the

thermal stresses encountered during Cu interconnects

fabrication process; thermal annealing at 425-C in nitrogen

ambient for 1 h was performed 7 times. For the humidity

test, a pressure cooker test (PCT) was carried out at 120-C,

100% relative humidity, and 2 atmosphere pressure for 168

h. To test the impact of O2plasma on the film properties, the

as-deposited DEMS-based films were exposed to O2plasma

environment in a cathode-coupled rf asher. The pressure and RF power were 10 mTorr and 200 W, respectively, and the process time was set at 60 s.

2.3. Analysis method

The low-k films were analyzed for thickness and refractive index (RI, at 633 nm) by reflectometer (SCI

FilmTek) and/or ellipsometer (Nano-Spec\9100) before

and after stressing. The thickness change is defined herein as

Thickness change %ð Þ ¼THKstressing THKAs:dep:

THKAs:dep:

 100%

where THKstressing and THKAs.dep. represent the measured

thickness after stressing and as-deposition, respectively.

Fourier transform infrared (FT-IR) measurements were operated in the absorbance mode on a Bio-Rad Win-IR PRO FT-IR spectrometer with wave number ranging from 400 to

4000 cm 1. FT-IR was performed at a resolution of 4 cm 1,

each spectrum being the average signal over 32 scans with the background corrected to a silicon reference. The dielectric constant (k) and leakage current density were measured by a SSM Inc. mercury probe cyclic voltammeter (CV) system at 1 MHz frequency. The k value was obtained from the average of 9 sites measurement.

3. Results and discussions

The dependence of the stress behavior of the as-deposited DEMS-based low-k films as a function of the

O2/DEMS ratio was investigated. The intrinsic stress of

DEMS-based low-k films becomes more tensile as the O2

flow rate is increased. Furthermore, thermal stability had

been evaluated by stress/temperature analysis.Fig. 1shows

the stress hysterisis curves of the low-k films with

polymer-ization of DEMS and oxidation of DEMS and O2. It reveals

that the stress shift between the first thermal cycle and the second cycle was suppressed for the low-k films prepared

by DEMS and O2. Minimal change in stress is observed for

(a)

(b)

Fig. 1. The stress-hysterisis of DEMS-based low-k films prepared by (a) O2/ DEMS = 0; (b) O2/DEMS = 0.05.

low-k films deposited using DEMS and O2. The shift

magnitude is 2.0E7 dyn/cm2, significantly smaller than that

with pure DEMS deposition (3.0E07 dyn/cm2). This result

implies that the low-k films deposited using DEMS and O2

have better thermal stability.

In the Cu integration processes, there are at least 7 – 8 layers interconnect dielectric deposition with the deposition

temperature of around 400 -C. The dependence of the

heating cycles on the degradation of the thickness and dielectric constant of DEMS-based low-k films with various

O2/DEMS ratios are compared in Fig. 2(a) and (b),

respectively. For DEMS-based low-k films, the thickness

remains constant after performing 425-C annealing cycling,

independent of O2/DEMS ratios shown in Fig. 2(a). This

implies that low-k films deposited using DEMS precursor has a higher bonding strength when subjected to the heating tests related to other OSG films prepared by other precursors [13 – 15]. On the other hand, it can be seen from Fig. 2(b) that the dielectric constant change of low-k films with

DEMS/O2 was found to behave differently to that with

DEMS only. The dielectric constant of the low-k films deposited only using DEMS gradually increase with

increasing the heating cycles of 425-C. In contrast, as O2

was incorporated into the reaction, the dielectric constant of DEMS-based low-k films almost retains constant, even after

the seven cycles of the 425 -C heating test. To further

investigate the heating resistance of DEMS-based low-k films, different heating temperatures, ranging from 400 to

800 -C, were performed. The influence of the heating

temperatures on the change in the thickness and the dielectric constant of DEMS-based low-k films are shown inFig. 3(a) and (b), respectively. Similar to other OSG

low-k films using other precursors [14,15], the dielectric

constant of the low-k films deposited using DEMS or

DEMS/O2 maintain a stable value (k = 2.8 – 3.2) at

temper-atures up to 600 -C. Moreover, the dielectric constant of

low-k films deposited only using DEMS degrade to 3.6 as

the heating temperature is increased to 700 -C and sharply

increases to 5.0 as the temperature increases to 800-C. On

the other hand, the dielectric constant of low-k films

deposited using DEMS/O2 does not degrade until the

annealing temperature is increased to 800-C. This indicates

that DEMS-based low-k films with O2 as an oxidant gas

have a superior thermal resistance compared to those deposited only using DEMS precursor. This result seems to imply that low-k films deposited using DEMS and

DEMS/O2have different bonding structures, which exhibits

1.05 1.025 1 0.975 0.95 As. dep. 1st ₄nd ₇nd Thermal Cycle (425ºC) THK thermal / THK As. dep. O2/DEMS=0 O2/DEMS=0.05 O2/DEMS=0.1 O2/DEMS=0.175 O2/DEMS=0 O2/DEMS=0.05 O2/DEMS=0.1 O2/DEMS=0.175

(a)

(b)

Fig. 2. The film properties change of DEMS-based low-k films after 7 times 425-C thermal annealing (a) thickness; (b) dielectric constant.

2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 k value (1 MHz) O2/DEMS=0 O2/DEMS=0.05 O2/DEMS=0.175

(b)

(a)

Fig. 3. The film properties of DEMS-based low-k films after thermal tests with temperature ranging from 400 to 800-C (a) thickness; (b) dielectric constant.

a different thermal resistance. To further investigate the difference in the bonding structure for these low-k films, FTIR and X-ray Photoemission spectroscopy (XPS)

analy-ses were conducted depicted in the pervious report [9,13].

The low-k films deposited using only DEMS precursor have

more – CH3 terminal bonds and less C – Si4xHx (x < 2)

bonds. In contrast, the deposited low-k films contain more

C – Si4xHx (x < 2) bonds to form a cross-linking structure

and more Si – H terminated bonds as O2gas was added to

the reaction. The speculated schematics of the low-k film chemical bonding structures deposited using (I) DEMS, (II)

DEMS/O2are shown inFig. 4. Furthermore, it is worth to

note that the dielectric constant of DEMS-based low-k films

deposited using DEMS/O2 further decrease after the

annealing process with temperature below 600 -C. The

decreasing dielectric constant could be attributed to the desorption of the small amount water in the film and rearrangement of the amorphous structure during annealing. It was observed from FTIR spectra that no significant

change in the concentration of the C – H bond and Si – CH3

bond was observed, which are thermally stable up to 600 -C. We believe that the decreasing dielectric constant of

low-k films deposited by DEMS/O2 was a result of the

formation of open ring structures in the films caused by the

loss of water and CHxorganic materials during annealing at

400 – 600 -C. Since the thermal resistance of Si–CHx– Si

bonds is lower than Si – CH3, the Si – CHx– Si bonds were

converted to CHx organic materials and desorped during

400 – 600 -C annealing. This is positive to reduction the

dielectric constant, but is negative to the thickness stability.

As the annealing temperature is increased to 700-C, the Si–

CH3absorbance peak dramatically decreases and there is an

increase in the oxide character of the film, and a loss of

methyl groups after the annealing process as shown in Fig.

5. Interestingly, the enhancement in Si – O characteristics is

less for the low-k films deposited using DEMS/O2 film,

which also showed much smaller increases in the dielectric

constant. This suggests that the Si – CH3 bonding did not

decompose until the thermal temperature above 700 -C. In

addition, low-k films deposited only using DEMS, where Si

is bonded with multi-methyl group ( – CH3) bonding would

degrade easily during the higher temperature annealing. The change in the dielectric constant of DEMS-based

low-k films under the moisture stress test in terms of O2/

DEMS ratios is shown in Fig. 6. It indicates a slightly

increase in the refractive index and the dielectric constant after a 168 h PCT for two different deposition conditions. Compared to low-k films deposited using 3MS as precursor [13], low-k films deposited using DEMS as precursor shows a better moisture resistance, implying that this film has a

surface hydrophobic property as a result of more – CH3

Si Si Si CHn O CH3 O Si CH3 CH3 CH3 O O O O Si CH3 Si CH3 Si Si CHn CHn CH3 O O CH3 O O Si Si Si CHn O CH3 O Si CH3 CH3 O O O O Si CH3 Si CH3 Si Si CHn CHn CH3 O O O O CH3 CH3 CH3 Si Si Si CHn O CH3 O Si CH3 CH3 CH3 O O O O Si CH3 Si CH3 Si Si CHn CHn CH3 O O CH3 O O Si Si Si CHn O CH3 O Si CH3 CH3 O O O O Si CH3 Si CH3 Si CHn CHn CH3 O O O O CH3 CH3 CH3 Si Si CH O CH O Si CH O O Si CH CH O Si Si CHn CH3 CHn O Si CH3 CH_n O O O O O Si Si Si Si CH_n CHn O H O H H Si CH O CH O Si CH O O Si CH CH O Si Si CHn CH3 CHn O Si CH3 CH_n O O O O O Si Si Si Si CH_n CHn O H O H H

(a)

(b)

Si

Fig. 4. Schematics of the chemical bonding structure (a) DEMS; (b) DEMS + O2. Post-alloy Pre-alloy Post-alloy Pre-alloy Si-O Si-O Wavenumber(cm-1₎ C-H Si-H C-H Si-H Si-CH₃ 4000 3500 3000 2500 2000 1500 1000 500 4000 3500 3000 2500 2000 1500 1000 500 0.20 0.16 0.12 0.08 0.04 0.00 0.20 0.16 0.12 0.08 0.04 0.00 -0.04 Absorbance Absorbance

(a)

(b)

Fig. 5. FTIR spectrum change of DEMS-based low-k films after 700-C thermal annealing (a) DEMS; (b) DEMS + O2.

terminal bonds. Additionally, the dielectric constant of low-k films deposited only using DEMS precursor increases to 2.93 from 2.86, which is slightly lower than that prepared by

DEMS/O2with a dielectric constant of 2.96. More surface

hydrophobic – CH3 bonds in low-k films prepared using

DEMS precursor hinders the moisture penetrate into the film, causing this divergence. More interesting to note, the stress of the low-k films tends to decline to neutral value after the moisture test. Furthermore, Thermal Desorption

Spectrum (TDS) analysis indicated that the H2O peak was

observed at about 250-C and the desorption amount of H2O

was greater for the low-k films prepared using DEMS/O2

reaction. These results imply that the deposited low-k film still contains moisture in terms of physical absorption and low-k films produced only using DEMS precursor have better moisture resistance.

In the interconnect integration fabrication, the ILD layer etching and photo-resist stripping processes are

indispen-sable steps. Fig. 7shows the etching rate of DEMS-based

low-k film as a function of O2/DEMS ratios, performed

using Ar/C5F8/N2 gas. The etching rate of DEMS-based

low-k films slightly increases with increasing the O2/DEMS

ratio. The higher etching rate for DEMS-based low-k films

with a higher O2flow rate is suspected to porosity structure

in the low-k film; that is, the film is less dense (refractive index is lower for DEMS-based low-k films with a higher

O2 flow rate). On the other hand, the increase of the

dielectric constant for post-etching DEMS-based low-k films is less than 6%. This indicates that low-k film

deposited by DEMS and O2 gas can efficiently withstand

the treatment of the etching chemical gas.

In conventional photo-resist stripping step process, O2

plasma ashing is commonly implemented because of its better efficiency in removing the polymer. However, all porous low-k films would suffer a degradation of the

dielectric constant after exposing O2plasma ashing process

[16]. Therefore, the effect of O2 plasma ashing on low-k

films prepared by DEMS or DEMS/O2was investigated in

this study. Fig. 8 shows the change of thickness and the

dielectric constant of DEMS-based low-k film as a

function of the O2/DEMS ratio. The thickness reduction

can be negligible since the maximum thickness reduction is about 2.5%, which occurred on the low-k films deposited only using DEMS precursor. Additionally, the thickness reduction is less 1% for low-k films deposited

using DEMS/O2 reactant. A plausible explanation is that

terminated methyl groups in DEMS-based low-k films are

oxidized in the O2 plasma condition, in line with the

following Eqs. (1) – (3):  Si  CH3þ H3C Si  þOY  Si  OH þ OH  Si  þCO þ H2O ð1Þ Si  OH þ OH  Si Y  Si  O  Si  þH2O ð2Þ 250 255 260 265 270 275 280 285 290 295 300 0 0.025 0.05 0.075 0.1 0.125 0.175 O₂/DEMS ratio

Dry Etching Rate (nm/min.)

0 1 2 3 4 5 6 7 k increase (%)

Fig. 7. The dry etching rate and dielectric constant change of DEMS-based low-k films as a function of the O2/DEMS ratio.

0 0.5 1 1.5 2 2.5 3 0 0.025 0.05 0.075 0.1 0.125 0.175 O2/DEMS ratio Thickness Reduction (%) 0 5 10 15 20 25 30 35 k increase (%)

Fig. 8. The thickness and dielectric constant change of DEMS-based low-k films after O2plasma ashing as a function of the O2/DEMS ratio. 1.00 1.50 2.00 2.50 3.00 3.50 4.00 As. Dep. As. Dep. PCT Test PCT Test O2/DEMS=0 O2/DEMS=0.05 1.00 1.50 2.00 2.50 3.00 3.50 4.00 k value (1 MHz) As. Dep. As. Dep. PCT Test PCT Test O2/DEMS=0 O2/DEMS=0.05

Fig. 6. The dielectric constant change of DEMS-based low-k film after 168 h PCT test.

 Si  CHn C  Si  þOY  Si  O  Si  þCO

The forming – Si – O – Si – bonds are almost as dense as – Si – C – Si – as the bond length of Si – O is similar to that of

Si – C bonds. As a result, the thickness change under O2

oxidation can be negligible, which may account for the lower

reduction in the DEMS/O2reaction because the low-k films

deposited using DEMS/O2contains more Si – C – Si bonds.

In contrast to the thickness reduction, the degradation in the dielectric constant of low-k films deposited using

DEMS/O2gas becomes serious as O2flow rate is increased.

On the other hand, low-k films deposited only using DEMS gas show a lower change in the dielectric constant. This is expected to the increased Si – H bonds in the deposited

low-k films with a higher O2flow rate. Exception of the above

oxidation, O2plasma treatment has also been found to result

in dangling bonds arising from the enhanced breaking of Si – H bonds by oxygen radicals. The reaction of Si – H and oxygen radicals can be expressed as the following Eq. (4). The resulting OH bonding increases the dielectric constant.

Si  H þ OY  Si  OH ð4Þ

Therefore, low-k films deposited using DEMS/O2gas are

needed to further improve its O2plasma ashing resistance

when we consider it as the inter-layer dielectric.

To solve the dielectric constant degradation issue for the deposited low-k films, the dielectric constant of

DEMS-based low-k films with post O2 plasma treatment was

measured after dry etching method, as shown inFig. 9. The

purpose of dry etching is to remove the dense layer, which is

oxidized on the top of low-k film during the O2 plasma

ashing process. As can be seen, after the dry etch, the dielectric constant of the low-k films deposited using DEMS

or DEMS/O2 reduced to about 3.3, which is close to the

original as-deposition value. It clearly shows that the low-k

film inside is not degraded during the O2ashing process.

SEM also displays the distinct two-layers inside the low-k

film after the O2plasma ashing. The depth of the oxidation

film increases with increasing O2exposure time. This dense

surface can be removed by sputter etching method prior to the metal deposition in the practical fabrication process. As

a result, control of the O2 plasma time and pre-sputter

etching process is a feasible method in recovering the low-k film dielectric constant. Another approach to alleviate

dielectric constant degradation is using NH3or N2plasma

treatment on the as-deposited low-k films. This N2plasma

treatment would induce the N atom doping in low-k films and replace the Si – H bonds to form a thin Si – N layer on the film surface. The dielectric constant of low-k films

prepared by DEMS/O2 is slightly increased from 2.78 to

2.86 after N2plasma treatment. The new forming layer can

effectively impede the O2plasma ashing attack for the low-k

films.Fig. 10shows the leakage current density at 2 MV/cm

and dielectric constant of the N2plasma treated low-k films

before and after being exposed to the O2plasma treatment.

The leakage current density remains at a stable value of

about 1.40E 8 A/cm2

at 2 MV/cm, which is slightly lower

than the as-deposited low-k film (1.65E 8 A/cm2

at 2 MV/ cm) dut to the formation of Si – N bonds. In addition, the

dielectric constant of N2 plasma-treated low-k films

deposited with 0.05 ratios of O2/DEMS maintains the stable

value when compared to the films without N2 plasma

treatment. This result indicates this functional group

produced by N2 plasma treatment ensures that the low-k

films have a superior electrical performance to against O2

plasma ashing damage.

4. Conclusion

The resistance of the organo-silicate glass low-k film

deposited using DEMS and various O2flow against heat,

moisture stress and chemical test for was investigated. Low dielectric constant organo-silica-glass film deposited using

DEMS and O2 is shown to be the most reliable. The

dielectric constants are stable even after a heating test at 700 -C and a pressure cooler test for 168 h, and are superior to other PECVD low-k films deposited using other precursors.

1.00 1.50 2.00 2.50 3.00 3.50 4.00 4.50 5.00 Ashing Dry Etching O2/DEMS=0 k value (1 MHz) O2/DEMS=0.05 As. Dep. Value

Ashing

Dry Etching

Fig. 9. The dielectric constant change of DEMS-based low-k films by removing the surface layer.

1 1.5 2 2.5 3 3.5 4 4.5 5

As. Dep. Nitrogen Treated Oxygen Ashing

k value (1 MHz) 1.3 1.35 1.4 1.45 1.5 1.55 1.6 Leakage Current ( x 10 -8 A/cm 2)

Fig. 10. The change in dielectric constant and leakage current at 2 MV/cm of the nitrogen treated low-k films after O2plasma ashing.

This excellent stability ensures the low-k film deposited using DEMS is suitable for application as multilevel interconnects, showing long-term reliability after fabrication.

However, the O2plasma ashing process leads to a dielectric

degradation in deposited low-k films during photoresist

removal processing. A N2plasma treatment is proposed as a

method of preventing the damage from an O2plasma attack

on the low-k films deposited using DEMS/O2gas.

Acknowledgement

The authors gratefully acknowledge the financial support of National Science Council (NSC) of Taiwan for this research project under Contract no. NSC92-2120-E-006-003.

References

[1] L.M. Han, J.S. Pan, S.M. Chen, N. Balasubramanian, J. Shi, L.S. Wong, P.D. Foo, J. Electrochem. Soc. 148 (2001) F148.

[2] P.T. Liu, T.C. Chang, H. Su, Y.S. Mor, Y.L. Yang, H. Chung, J. Hou, S.M. Sze, J. Electrochem. Soc. 148 (2001) F115.

[3] V. Ligatchev, T.K.S. Wong, B. Liu, Rusli, J. Appl. Phys. 92 (2002) 2172.

[4] A. Grill, V. Patel, J. Appl. Phys. 85 (1999) 3314.

[5] Q. Wu, K.K. Gleason, J. Vac. Sci. Technol., A, Vac. Surf. Films 21 (2003) 388.

[6] V. Ligatchev, T.K.S. Wong, B. Liu, Rusli, J. Appl. Phys. 92 (2002) 4605.

[7] Y. Shioya, Y. Kotake, T. Ishimaru, T. Masubuchi, H. Ikakura, K. Maeda, J. Electrochem. Soc. 150 (2003) F1.

[8] P. Gonon, A. sylvestre, H. Meynen, L.V. Cotthem, J. Electrochem. Soc. 150 (2003) F47.

[9] M.L. O’Neill, R.N. Vritis, A.S. Lukas, J.L. Vincent, B.K. Peterson, M.D. Bitner, E.J. Karwacki, in proceedings of 2000 VLSI.

[10] T. Furusawa, D. Ryuzaki, R. Yoneyama, Y. Homma, K. Hinode, J. Electrochem. Soc. 4 (2001) G31.

[11] T.C. Chang, Y.S. Mor, P.T. Liu, T.M. Tsai, C.W. Chen, Y.J. Mei, S.M. Sze, J. Electrochem. Soc. 149 (2003) F81.

[12] T. Furusawa, D. Ryuzaki, R. Yoneyama, Y. Homma, K. Hinode, J. Electrochem. Soc. 148 (2001) F175.

[13] Y.L. Cheng, Y.L. Wang, Y.L. Wu, C.L. Liu, J.K. Lan, M.L. O’Neill, C.P. Liu, C. Ay, M.S. Feng, Thin Solid Films 447 (2004) 681. [14] D.J. Thomas, Y.P. Song, K. Powell, Solid State Technol. 15 (2001)

107.

[15] C.B. Labelle, K.K. Gleason, J. Vac. Sci. Technol., A, Vac. Surf. Films 17 (1999) 445.

[16] K. Yonekura, S. Sakamori, K. Goto, M. Matsuura, N. Fujiwara, M. Yoned, J. Vac. Sci. Technol., B 22 (2004) 548.