APPLIED PHYSICS LETTERS
VOLUME 80, NUMBER 10
11 MARCH 2002
Effect of N2 O plasma treatment on the stabilization of water absorption
in fluorinated silicon-oxide thin films fabricated by
electron-cyclotron-resonance plasma-enhanced
chemical-vapor deposition
S. P. Kim and S. K. Choia)
Department of Materials Science and Engineering, Korea Advanced Institute of Science and Technology,
373-1 Kusung-Dong, Yusung-Gu, Taejon 305-701, Korea
Youngsoo Park
Materials and Devices Laboratory, Samsung Advanced Institute of Technology, Suwon 440-600, Korea
Ilsub Chung
School of Electrical and Computer Engineering, Sungkyunkwan University, Suwon, 440-746, Korea
共Received 31 August 2001; accepted for publication 9 January 2002兲
The variation of residual stress with the water absorption was reduced drastically by the N2 O plasma
treatment for fluorinated silicon-oxide thin films. Fourier transformed infrared spectroscopy analysis
showed that the film was oxidized by the plasma treatment. It was also determined that the oxidation
occurred on the film surface from the P-etch rate and x-ray photoelectron spectroscopy analysis. The
experimental results show that the stabilization results from the oxidation of the surface by the N2 O
plasma treatment. © 2002 American Institute of Physics. 关DOI: 10.1063/1.1458528兴
In a multilevel interconnection technology, propagation
delay due to parasitic capacitance is one of the important
issues. The delay becomes more serious in advanced integrated circuits as interconnect dimensions are scaled down.
To reduce the delay, various materials with low dielectric
constant have been investigated.1 Fluorinated silicon-oxide
共SiOF兲 is one of the attractive materials for an intermetal
dielectric layer because it has a low dielectric constant and
the process is similar to that of SiO2 which is currently used
as intermetal dielectrics.2 The relative dielectric constant of
SiOF film is lower 共3.0–3.7兲 than that of SiO2 共⬎4.2兲.1,2
However, the incorporation of fluorine results in the instability of the film properties. Films with higher fluorine
concentration easily absorb water and it degrades the electrical and mechanical properties of films.3– 8 It has been accepted that the water absorption is caused by the interaction
between SiuF bond and water. There have been many
works on the stabilization of SiOF films to the water
absorption.6 – 8 Two methods were suggested to suppress the
water absorption. One is capping the SiOF film with oxide or
nitride which is stable to the water.6,7 The other is a plasma
treatment.8,9 Although the latter is a very simple process, its
effect is noticeable. Lee and Park reported that O2 plasma
treatment enhanced the stability of SiOF films to the water
absorption.8 They expected that the stabilization was due to
the densification of the film by the ion bombardment effect
or oxidation. In the case of SiO2 films, the effects of O2 and
N2 O plasma treatments on the electrical properties were also
studied.9,10 The breakdown fields increased and the interfacestate densities decreased after the plasma treatment. They
explained that the passivation of silicon dangling bonds and
linking reaction of SiuOH bonds caused the
a兲
Electronic mail: [email protected]
improvement.10 However, there is little experimental evidence which supports these explanations.
In the present work, the effect of N2 O plasma treatment
on the water absorption and the origin of stabilization in
SiOF films were studied. In order to characterize the water
absorption behavior, the variation of residual stress was observed as a function of storage time. The change of chemical
bonding structure with the N2 O plasma treatment time was
also analyzed using Fourier transformed infrared 共FTIR兲
spectroscopy. The elemental composition depth profile was
obtained from the x-ray photoelectron spectroscopy 共XPS兲
analysis combined with Ar⫹ sputtering. The etching rate was
measured using P-etch solution 关 15HF(49%):10HNO3
(70%):300H2 O兴 . The water absorption was reduced drastically after the N2 O plasma treatment. The results of FTIR,
P-etch rate, and XPS analyses in our study implied that the
stabilization was due to the oxidation of the surface.
1100-Å-thick SiOF films were deposited on chemically
cleaned 4 in., 具100典, Si wafers using electron-cyclotronresonance 共ECR兲 plasma-enhanced chemical-vapor deposition 共PECVD兲. The flow rates of SiF4 , Ar, and N2 O were
fixed at 5 sccm, 5 sccm, and 30 sccm, respectively. The
pressure was kept at 2 mTorr and microwave power was 600
W during the deposition. The substrate was not heated intentionally. After the deposition process, the flow of SiF4 and Ar
was stopped. Then, the films were exposed to N2 O plasma
with the same pressure and power. The exposure time was
varied from 0 to 4 min. SiO2 films were also prepared for a
comparison with the SiOF films. For SiO2 films, SiH4 was
used instead of SiF4 and other conditions were fixed. FTIR
analysis was performed using a Bio Rad FTS 6000 system.
The residual stress was measured by a Flexus stress measurement system 共FLX 2320兲. It has been reported that the residual stress varies sensitively with the water absorption for
SiO2 and SiOF films deposited by PECVD.5,11,12 According
0003-6951/2002/80(10)/1728/3/$19.00
1728
© 2002 American Institute of Physics
Downloaded 19 Jul 2009 to 143.248.110.216. Redistribution subject to AIP license or copyright; see http://apl.aip.org/apl/copyright.jsp
Kim et al.
Appl. Phys. Lett., Vol. 80, No. 10, 11 March 2002
1729
TABLE I. Variations of initial stress, ( ␴ ⫺ ␴ i )/ ␴ i 共%兲 and I Si—F /(I Si—O
⫹I Si—F) 共%兲 with the plasma treatment time.
Plasma
treatment
time 共min兲
0
1
2
3
4
SiO2
Initial stress, ␴ i
共MPa兲
( ␴ ⫺ ␴ i )/ ␴ i 共%兲
after 5 h in
45% RH air
I Si—F /
(I Si—O⫹I Si—F)
共%兲
⫺70.0
⫺115.5
⫺115.4
⫺93.0
⫺90.8
⫺270.0
100.2
23.8
9.2
5.9
⫺2.6
0.4
10.9
9.4
8.3
6.9
6.3
0
to Park et al., the magnitude of the compressive stress increases with the water absorption for SiO2 films deposited by
PECVD.11 Haque, Naseem, and Brown insisted that three
factors effect the variation of the residual stress in the case of
SiO2 films; water adsorption on the surface, SiuOH formation, and the interaction between adsorbed water dipoles in
the columnar structure.12 We have previously reported that
the SiOF films fabricated by ECR PECVD also show an
increase of compressive stress with the water absorption,
which results from the physical adsorption of water on the
film surface.5 In this study, the amount of the water absorption was estimated by measuring the variation of the residual
stress in films with the storage time in room air. The etching
rate was determined in P-etch solution. All the analyses were
carried out immediately after the plasma treatment process.
All the films were under compressive stress conditions
共Table I兲. The compressive residual stress increases with the
storage time in room air 关25 °C and 45% relative humidity
共RH兲兴, however that of SiO2 film does not change 共⬃⫺270
MPa兲. It is inferred that the SiO2 films deposited in this study
are very stable to water. Chang et al. also observed that ECR
SiO2 even deposited at room temperature did not absorb water and swell.6 Figure 1 shows the increment of compressive
stress with respect to an initial stress value ( ␴ i ), ( ␴
⫺ ␴ i )/ ␴ i 共%兲, as a function of storage time. The increment
decreases significantly with increasing the plasma treatment
time. Finally, the change of residual stress becomes negligible. These results indicate that the N2 O plasma treatment
effectively suppresses the water absorption of SiOF films.
The inhibition of the water absorption by the plasma treatment has been reported previously, however an origin is not
FIG. 2. Variations of FTIR spectra, Si—O and Si—F bond, with the N2 O
plasma treatment time.
known. In order to clarify the reason of stabilization, the
change of chemical bonding structure was analyzed using
FTIR.
The variations of SiuO and SiuF peaks with the
plasma treatment time are plotted in Fig. 2. The intensity of
SiuO stretching mode 共⬃1070 cm⫺1兲 increases, while that
of SiuF stretching mode 共⬃945 cm⫺1兲 decreases with the
treatment time. The integrated intensities of SiuO and
SiuF stretching peaks (I SiuO ,I SiuF) were used to estimate
the fluorine content in the film. I SiuF /(I SiuF⫹I SiuO) ratios
are also summarized in Table I. The result implies that the
plasma treatment removes the SiuF bonds and enhances the
formation of SiuO bonds. It is thought that the energetic
oxygen ions or radicals react with the films. It has been
known that the water absorption in SiOF films is due to the
interaction between the water molecules and SiuF
bonds.3,4,7 Thus, it is supposed that the reduction of SiuF
bond and formation of SiuO bond might have caused the
stabilization of SiOF films. Researchers have already presented similar explanations about the effect of plasma treatment on SiO2 and SiOF films.8 –10 They explained that the
surface state was changed after the treatment because of the
ion bombardment or the oxidation. However, there is little
experimental evidence that supports the change of surface
clearly.
FIG. 1. Increment of compressive stress with storage time in 45% RH air for
FIG. 3. Variation of film thickness with etching time in P-etch solution of
SiOF and SiO2 films with different N2 O plasma treatment time.
N2 O plasma treated films.
Downloaded 19 Jul 2009 to 143.248.110.216. Redistribution subject to AIP license or copyright; see http://apl.aip.org/apl/copyright.jsp
1730
Kim et al.
Appl. Phys. Lett., Vol. 80, No. 10, 11 March 2002
FIG. 4. XPS depth profile results and cross sectional SEM images of SiOF
films before 关共a兲 and 共b兲兴 and after 关共c兲 and 共d兲兴 N2 O plasma treatment for 4
min.
The film thickness was measured as a function of etching time in P-etch solution and the results are shown in Fig.
3. It has been well known that P-etch rate of silicon oxide
film is very sensitive to the film density. The etching rate of
SiO2 in this study is ⬃3 Å s, which is a similar value for
thermally grown SiO2 共⬃2 Å/s兲.13 In the case of SiOF film
without the N2 O plasma treatment, the etch rate is ⬃40 Å/s.
Figure 3 indicates that the etching rate of the surface region
共⬃200 Å兲 is quite different from that of the bulk region for
SiOF films after the N2 O plasma treatment. Although, the
etching rate of bulk region remains nearly constant even after
the N2 O plasma treatment, the etching rate of surface region
decreases with increasing the plasma treatment time. It finally reaches to the value of SiO2 film that does not absorb
the water as presented in Fig. 1. It is expected that the surface region fully changed to dense silicon-oxide from SiOF
by ions or radicals in the N2 O plasma for 4 min. However,
there is another possible explanation that the slow etch rate
of the surface is due to the densification of the surface region
only by the ion bombardment without the oxidation. To
clarify that oxidation occurs at the surface region, the elemental composition depth profile was obtained from XPS
analysis and cross sectional view of SiOF film was observed
using scanning electron microscopy 共SEM兲.
Figure 4 shows the XPS depth profile results and cross
sectional SEM images of SiOF films before 关Figs. 4共a兲 and
4共b兲兴 and after 关Figs. 4共c兲 and 4共d兲兴 the plasma treatment for
4 min. XPS depth profile result was obtained after the sputtering process for 30 s to remove the contaminant on the film
surface. The result shows that the plasma treatment removes
the fluorine atoms on the film surface and changes the surface region to silicon oxide. Considering the change of
I SiuF /(I SiuO⫹I SiuF) 共Table I兲 and XPS result, it is thought
that the oxidation occurs in the surface region. XPS results
also show no incorporation of nitrogen in the films, which
implies that no reaction occurred between SiOF film and
nitrogen atom. The cross sectional SEM image shows a layer
with different brightness for the plasma treated film. It seems
that the bright layer is the oxidized region by the plasma
treatment.
The effect of N2 O plasma treatment on the water absorption of SiOF films and the origin of stabilization were studied. The N2 O plasma treatment effectively reduced the water
absorption. It was found that the N2 O plasma treatment
causes the oxidation of SiuF bond on the film surface and
the oxidation reaction makes the surface dense silicon oxide.
Thus, the oxidized surface inhibits the water absorption of
films.
This work is partially supported by the Korea Science
and Engineering Foundation 共No. 95-0300-15-01-3兲 and the
Brain Korea 21 project in 2001.
L. Peters, Semicond. Int. 23, 108 共2000兲.
R. K. Laxman, Semicond. Int. 18, 71 共1995兲.
3
V. Pankov, J. C. Alonso, and A. Ortiz, J. Vac. Sci. Technol. A 17, 3166
共1999兲.
4
H. Yang and G. Lucovsky, J. Non-Cryst. Solids 254, 128 共1999兲.
5
S. P. Kim, S. K. Choi, Y. Park, and I. Chung, Appl. Phys. Lett. 79, 185
共2001兲.
6
K. M. Chang, S. W. Wang, C. J. Wu, T. H. Yeh, C. H. Li, and J. Y. Yang,
Appl. Phys. Lett. 69, 1238 共1996兲.
7
M. Yoshimaru, S. Koizumi, and K. Shimokawa, J. Vac. Sci. Technol. A 15,
2915 共1997兲.
8
S. Lee and J.-W. Park, J. Vac. Sci. Technol. A 17, 458 共1999兲.
9
P. C. Chen, K. Y. J. Hsu, H. L. Hwang, and J. Y. Lin, J. Appl. Phys. 76,
5508 共1994兲.
10
Y. C. Chen, M. Z. Tang, I. C. Tung, M. Chen, M. S. Feng, H. C. Cheng,
and C. Y. Chang, Jpn. J. Appl. Phys., Part 1 38, 4226 共1999兲.
11
Y. S. Park, J. K. Lee, I. S. Jung, S. B. Heo, and J. Y. Lee, Appl. Phys. Lett.
75, 3811 共1999兲.
12
M. S. Haque, H. A. Naseem, and W. D. Brown, J. Appl. Phys. 81, 3129
共1997兲.
13
J. Batey and E. Tierney, J. Appl. Phys. 60, 3136 共1986兲.
1
2
Downloaded 19 Jul 2009 to 143.248.110.216. Redistribution subject to AIP license or copyright; see http://apl.aip.org/apl/copyright.jsp