Characteristics of Si Ox thin film deposited by atmospheric pressure plasma-enhanced chemical vapor deposition using PDMS O2 He

Article abstract of DOI:10.1149/1.3129464

Silicon oxide thin films were deposited using a modified, pin-to-plate, dielectric barrier discharge system with polydimethylsiloxane (PDMS), bubbled by He O2 gas mixtures at atmospheric pressure and a temperature of less than 50°C. Increasing PDMS flow r

Full text of DOI:10.1149/1.3129464

D248
Journal of The Electrochemical Society, 156 ͑7͒ D248-D252 ͑2009͒
0013-4651/2009/156͑7͒/D248/5/$25.00 © The Electrochemical Society
Characteristics of SiOx Thin Film Deposited by Atmospheric
Pressure Plasma-Enhanced Chemical Vapor Deposition
Using PDMSÕO2ÕHe
J. H. Lee, Y. S. Kim, J. S. Oh, S. J. Kyung, J. T. Lim, and G. Y. Yeomz
Department of Advanced Materials Science and Engineering, Sungkyunkwan University, Suwon 440-746,
Republic of Korea
Silicon oxide thin films were deposited using a modified, pin-to-plate, dielectric barrier discharge system with polydimethylsilox-
ane ͑PDMS͒, bubbled by He/O2 gas mixtures at atmospheric pressure and a temperature of less than 50°C. Increasing PDMS flow
rate in the gas mixture increased the deposition rate, but also increased the surface roughness due to the formation of particles in
the gas phase as a result of increased PDMS and silanol groups, leading to incomplete decomposition or oxidation of PDMS. The
increase in the ratio of oxygen flow rate to PDMS flow rate decreased the surface roughness with increasing deposition rate due
to the efficient oxidation of PDMS. However, when the oxygen flow rate was raised above 1 slm, due to the increased oxidation
of PDMS in the gas phase and the decreased PDMS dissociation by the decreased plasma density, the surface roughness was again
increased with decreasing deposition rate. At the gas mixture of 9 slm PDMS/He and 1 slm oxygen, a smooth, SiO2-like thin film
was obtained at a deposition rate of 12 nm/min.
© 2009 The Electrochemical Society. ͓DOI: 10.1149/1.3129464͔ All rights reserved.
Manuscript submitted July 14, 2008; revised manuscript received January 12, 2009. Published May 20, 2009.
Atmospheric pressure, plasma-enhanced chemical vapor deposi-
down voltage and a higher plasma density than the conventional
DBD at the same applied voltage due to the high electric field at the
multiple pins of the power electrode.12
tion ͑AP-PECVD͒ processes are recognized as promising and cost-
effective methods for wide-area coating on sheets of steel, glass,
polymeric web, etc. To reduce the production cost of large-area elec-
tronic devices such as flat-panel display devices and thin-film pho-
tovoltaics, it is desirable to develop alternative deposition processes
that operate at atmospheric pressure and utilize in-line processing.
SiO2 layers have a wide variety of electronic applications such as
gate dielectrics, insulators, and waveguides.1,2 In addition to elec-
tronic applications, SiO2 layers are routinely applied to polymers to
reduce the water vapor and oxygen permeation into versatile poly-
meric materials.3 AP-PECVD has been used for the low temperature
deposition of SiO2 thin films on thermally sensitive materials such
as flexible displays, organic light-emitting diode, and organic thin-
film transistors due to the difficulty of depositing SiO2 thin film on
these materials at high temperatures.
AP-PECVD of SiO2 at low temperature involves the dissociation
of volatile silicon precursors/oxygen with plasma and the formation
of SiO2 onto the substrate surface.4,5 Organo-silicon precursors are
advantageous precursors for PECVD of silicon compounds because
of the increased safety of manufacturing ͑in comparison with silane͒
arising from the stability, low flammability, and low toxicity of these
compounds.6,7 When PECVD is used for SiO2 formation, the in-
creased pressure can form unwanted solid particles which deposit
from the gas phase, thereby reducing the transparency of the depos-
ited films.8 Chemically, more reactive precursors are found to pro-
duce far more particles as a result of significant polymerization re-
actions and subsequent powder formation in the gas phase.9 For
example, organo-silicon precursors such as tetramethylcyclotetrasi-
loxane, octamethylcyclotetrasiloxane, polyhydrogenmethysiloxane,
and tetramethyldisiloxane are known to show a serious particle
problem during the deposition of SiO2 by AP-PECVD. However,
compared to other organo-silicon precursors, polydimethylsiloxane
͑PDMS͒ is known to exhibit superior material properties in addition
to fewer particle problems, possibly due to the lower chemical re-
activity of PDMS during the deposition with a dielectric barrier
discharge ͑DBD-type͒ AP-PECVD source.10,11
Experimental
Figure 1 shows the AP-PECVD in-line system used for the depo-
sition of SiO2. The discharge system was composed of a multipin
top electrode and a flat-ground electrode separated by about 8 mm.
Both electrodes were covered with a 3 mm thick ceramic plate and
had dimensions of 198 mm long ϫ 105 mm wide. Atmospheric
pressure plasmas were generated by applying 5 kV ac power
͑25 kHz͒ to the multipin electrode. At that condition, the power
consumed in the glow discharge was about 258.3 Ϯ 7 W. The sub-
strate temperature was maintained below 50°C using a chiller. The
process gas was fed into the system through the shower holes in the
ceramic material of the power electrodes. PDMS prepolymer and
curing agent ͑Sylgard 184, Dow Corning͒ were used as the Si pre-
cursors. The chemical structure of PDMS is composed of repeating
–OSi͑CH3͒2O– units.13 The PDMS prepolymer and curing agent
͑prepolymer: curing agent = 10:1͒ were fed into the system by bub-
bling He gas through the PDMS liquid reservoir kept at room tem-
perature.
To optimize the characteristics of the thin film, the PDMS and O2
flow rates were varied from 7 to 15 slm and from 0.6 to 1.4 slm,
In this study, SiO2 thin films were deposited using a modified
DBD called a “pin-to-plate-type DBD” to generate high density
plasma with a gas mixture of PDMS/O2. The effect of the gas mix-
ture on the physical and chemical properties of SiO2 was investi-
gated. The pin-to-plate DBD used in this study had a lower break-
Figure 1. Schematic diagram of the AP-PECVD in-line system ͑pin-to-plate
DBD type͒.
z E-mail: gyyeom@skku.edu
Downloaded on 2014-11-12 to IP 142.58.129.109 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 156 ͑7͒ D248-D252 ͑2009͒
D249
Figure 2. Deposition rate of silicon oxide films as a function of PDMS and
O2 flow rates using PDMS ͑bubbled by He͒/O2 at an applied voltage of 5 kV
͑25 kHz͒. When the PDMS flow rate was varied, the oxygen flow rate was
maintained at 1.0 slm, and when the oxygen flow rate was varied, the PDMS
flow rate was maintained at 9.0 slm.
respectively, while the other process parameters were fixed. During
the deposition, the substrates were fed into the plasma at the line
speed of 0.3 m/min.
The thickness of the deposited film was measured using a step
profilometer ͑Tencor, Alpha step 500͒, and the chemical bonding
states of the film were measured by X-ray photoelectron spectros-
copy ͑XPS; Thermo Electronics, Multilab ESCA2000͒ and Fourier
transform infrared spectrometry ͑FTIR; Bruker IFS-66/S͒. A Mg
source was used as the X-ray source for XPS. The surface charging
during XPS analysis was corrected by shifting the C 1s peak to
284.5 eV. The surface of the thin films was observed by field-
emission scanning electron microscopy ͑Hitachi, S-4700͒.
Figure 3. FTIR spectra of silicon oxide deposited at different ͑a͒ PDMS and
͑b͒ O2 gas flow rates by AP-PECVD. The deposition conditions are the same
as those in Fig. 2.
the plasma became unstable by showing a filamentary-type dis-
charge and the plasma density was decreased by the significant in-
crease of electron attachment to the oxygen atoms as a result of the
increased level of O2 gas in the gas mixture. The decreased electron
density in the plasma decreased the dissociation of the PDMS, and
thereby decreased the deposition rate when the O2 gas flow sur-
passed the critical value of 1 slm.
The chemical structure of the deposited SiO2 thin films was in-
vestigated using FTIR spectroscopy. Figure 3a and b shows the
FTIR spectra measured for various PDMS flow rates ͑O2 at 1 slm͒
and for various O2 flow rates ͑PDMS at 9 slm͒, respectively. The
SiO2 thin films were deposited on p-type ͗100͘ silicon wafers and
the thickness of the thin film was maintained at 200 Ϯ 25 nm. As
shown in Fig. 3a, as the PDMS flow rate increased from
Results and Discussion
Figure 2 shows the deposition rate of SiO2 measured as a func-
tion of PDMS ͑bubbled by He at room temperature͒ and O2 gas flow
rates. Glass was used as the substrate and was kept at a temperature
of less than 50°C. As shown in the figure, when the PDMS flow rate
was increased from 7 to 15 slm while the O2 gas flow rate was
maintained at 1 slm, the deposition rate was increased almost lin-
early from 8.1 Ϯ 0.2 to 23.6 Ϯ 0.4 nm/min. The increased deposi-
tion rate was related to the increased silicon source in the gas mix-
ture due to the increased PDMS flow rate and was, therefore, related
to the increased reaction of silicon with oxygen in the gas mixture.
When the oxygen flow rate was increased from 0.6 to 1 slm while
the PDMS flow rate was kept at 9 slm, the deposition rate was
increased from 11.05 Ϯ 0.5 to 12 Ϯ 0.5 nm/min, but then de-
creased to 10.9 Ϯ 0.3 nm/min when the O2 gas flow rate was fur-
ther increased to 1.4 slm. The initial increase in deposition rate as
the O2 gas flow rate was increased to 1 slm was attributed to the
increased oxidation of PDMS and was therefore believed to be re-
lated to the increased decomposition of PDMS by oxygen for the
formation of silicon oxide in the film.14 However, the decrease in
deposition rate with further increase in O2 gas flow rate above 1 slm
was attributed to the decrease in the dissociation of PDMS due to
the decreased plasma density. At an oxygen flow rate above 1 slm,
7 to 15 slm, the peak intensities related to the Si–O–Si bonding,
16
located at the wavenumber ranges of 1000–110015 and 1500 cm−1
,
and to the Si–OH bonding, located at the wavenumber ranges of
16
925–930 and 3350–3650 cm−1
PDMS flow rate increased the peak intensities related to the
,
increased. Also, the increased
17
Si–͑CH3͒x bonding located at 801 cm−1
.
However, when the O2
gas flow rate was increased from 0.6 to 1.4 slm, as shown in Fig.
3b, the observed bonding peak intensities related to the Si–O–Si
bonding, Si–͑CH3͒x bonding, and Si–OH bonding did not change
significantly. The increase of the Si–OH bonding peak with increas-
Downloaded on 2014-11-12 to IP 142.58.129.109 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

D250
Journal of The Electrochemical Society, 156 ͑7͒ D248-D252 ͑2009͒
Table I. Composition of the silicon oxide thin films deposited as a
function of the PDMS and O2 flow rates measured by XPS (the
hydrogen content was not included due to XPS detection limit).
Flow rate
Gas
͑slm͒
Silicon
Oxygen
Carbon
PDMS
7
9
11
13
15
29.8
30.2
28.9
29.8
29.9
68.9
68.3
68.9
67.6
67.1
1.4
1.5
2.3
2.6
3.0
O2
0.8
1
1.2
1.4
28.9
30.2
30.2
29.9
69.4
68.3
67.6
68.1
1.7
1.5
2.2
2.0
ing PDMS flow rate suggested the presence of a high silanol
group,18 ͓–OnSi͑OH͒4−n–͔n, in the film, thereby possibly indicating
the higher hydrophilic property and higher porosity of the deposited
film.13 The increase of the Si–͑CH3͒x bonding peak with increasing
PDMS flow rate indicated the possible formation of a soft rather
than hard film. The increase of both Si–OH and Si–͑CH3͒x bonding
peaks with increasing PDMS flow rate was attributed to the insuffi-
cient oxidation of the dissociated PDMS.
Using the XPS wide-scan data, the composition of the films de-
posited as a function of the PDMS and O2 gas flow rates shown in
Fig. 3 was estimated and the results are shown in Table I. As shown
in the table, the composition did not vary significantly with PDMS
and O2 flow rates, and the carbon content was generally low.
Figure 5. XPS narrow-scan data of the Si 2p peaks of silicon oxide thin
films deposited as a function of the O2 flow rate: ͑a͒ 0.6, ͑b͒ 1.0, and ͑c͒
1.4 slm. The deposition conditions are the same as those in Fig. 2.
However, as shown in the table, the carbon percentage increased
from 1.4 to 3.0% as the PDMS flow rate increased from 7 to 15 slm
due to the increase of Si–͑CH3͒x bonding in the film. The carbon
content as a function of the O2 flow rate was minimized at 1.0 slm
of O2, possibly due to the optimized oxidation and dissociation of
PDMS.
Using the XPS narrow-scan data, the bonding states of silicon in
the deposited films were investigated. Figure 4a and b shows the
XPS narrow-scan data of Si 2p measured for the samples in Fig. 3a
deposited at PDMS flow rates of 9 and 15 slm, respectively. The Si
2p peak locations were calibrated by positioning the C 1s peak at
284.5 eV. As shown in the figures, the peak location of Si 2p de-
creased from 103.28 to 103.08 eV when the PDMS flow rate in-
creased from 9 to 15 slm. The Si 2p peaks can be deconvoluted to a
few peaks related to silicon binding to R ͓R = methyl group
͑–CH3͒x: R–Si–O3: 102.8 eV, R2–Si–O2: 102.1 eV, R3–Si–O:
Figure 4. XPS narrow-scan data of the Si 2p peaks of silicon oxide thin
films deposited as a function of the PDMS flow rate: ͑a͒ 9.0 and ͑b͒ 15 slm.
The deposition conditions are the same as those in Fig. 2.
101.5 eV͔, Si–O4
͑SiO2, 103.4 eV͒, and Si–͑O,OH͒
4
͑104.3 eV͒.19,20 The deconvoluted results are also shown in the fig-
Downloaded on 2014-11-12 to IP 142.58.129.109 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 156 ͑7͒ D248-D252 ͑2009͒
D251
Figure 6. SEM micrographs of silicon oxide thin films deposited as a func-
tion of the PDMS flow rate: ͑a͒ 9.0 and ͑b͒ 15 slm. The deposition conditions
are the same as those in Fig. 2.
ures. As shown in the figures, the silicon binding peak containing
SiO2 decreased, while the peak containing R–Si–O3 increased when
the PDMS flow rate increased from 9 to 15 slm. Also, the silicon
binding peak containing Si–͑O,OH͒ increased with increasing
4
PDMS flow rate. The silicon binding peaks related to R3–Si–O and
R2–Si–O2 showed very low peak intensities at both PDMS flow
rates of 9 and 15 slm. The very low silicon binding peak intensities
related to R3–Si–O and R2–Si–O2 were believed to be attributed to
the relatively high dissociation of PDMS under our experimental
condition. However, the increase of the peak related to R–Si–O3
binding observed with increasing PDMS flow rate appeared to show
the reduced oxidation of the dissociated PDMS with increased
PDMS flow rate. The increase of the silicon binding peak containing
Figure 7. SEM micrographs of silicon oxide thin films deposited as a func-
tion of the O2 flow rate: ͑a͒ 0.6, ͑b͒ 1.0, and ͑c͒ 1.4 slm. The deposition
conditions are the same as those in Fig. 2.
Si–͑O,OH͒ observed with increased PDMS flow rate was also re-
4
lated to the incomplete PDMS oxidation.20
the dissociation of PDMS decreased and the Si–͑CH3͒x bonding
again increased due to the decreased plasma density resulting from
the formation of unstable plasma.
Figure 5a-c shows the XPS narrow-scan data of Si 2p measured
for the samples in Fig. 3b deposited with O2 gas flow rates of 0.6, 1,
and 1.4 slm, respectively. Each Si 2p peak was deconvoluted simi-
larly to the related binding peaks shown in Fig. 4. As shown in the
figures, the silicon binding peak intensities related to R3–Si–O,
The surface morphology of the thin film deposited as a function
of PDMS and O2 gas flow rates was investigated by scanning elec-
tron microscopy ͑SEM͒ and the results are shown in Fig. 6 and 7,
respectively. Figure 6a and b shows the results for PDMS flow rates
of 9 and 15 slm, respectively, while the O2 gas flow rate was kept at
1 slm. Figure 7a-c shows the results for O2 gas flow rates of 0.6, 1,
and 1.4 slm, respectively, while the PDMS flow rate was kept at
9 slm. The thickness of the deposited thin film was maintained
at 200 nm. As shown in Fig. 6, the surface roughness increased
as the PDMS flow rate increased from 9 to 15 slm, particularly as it
increased from 13 to 15 slm, possibly due to particle formation
in the gas phase resulting from homogeneous condensation reactions
or due to porosity in the thin film resulting from the increased
R2–Si–O2, and Si–͑O,OH͒ were low and did not change signifi-
4
cantly with increasing O2 flow rate, which was similar to the result
presented in Fig. 3b. However, for the silicon binding peaks related
to SiO2 and R–Si–O3, at an O2 gas flow rate of 1 slm, a maximum
peak was obtained for the peak related to SiO2, while a minimum
peak was obtained for the silicon binding peak related to R–Si–O3.
These two peaks were believed to be related to the highest efficiency
of gas dissociation and oxidation of PDMS at an O2 flow rate of
1 slm. That is, at an O2 flow rate of less than 1 slm, Si–͑CH3͒x
bonding remained in the film due to the insufficient oxidation of
dissociated PDMS. However, at an O2 flow rate of more than 1 slm,
Si–͑OH͒ bonding in the film with the increased PDMS flow rate
x
Downloaded on 2014-11-12 to IP 142.58.129.109 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

D252
Journal of The Electrochemical Society, 156 ͑7͒ D248-D252 ͑2009͒
͑data not shown͒. Also, when the thin film was scratched with twee-
zers, the film deposited at a higher PDMS flow rate was more easily
scratched by the increased Si–͑CH3͒x bonding, indicating the forma-
tion of a soft thin film which is unsuitable as a water-permeation
diffusion barrier.21
SiO2-like thin film with a low carbon content of 1.5% and a smooth
surface was obtained at a deposition rate of 12 nm/min.
Acknowledgments
This research was supported by a grant ͑no. F0004041-2008-31͒
from the Information Display R&D Center, one of the 21st Century
Frontier R&D Programs funded by the Ministry of Knowledge
Economy of the Korean government.
For the thin film deposited as a function of O2 gas flow rate, the
increase of O2 gas flow rate from 0.6 to 1 slm decreased the surface
roughness so that a very smooth film surface with nearly no particles
on the surface was obtained at the latter O2 gas flow rate, as shown
in Fig. 7b. The roughness of the 0.6 slm O2 film was attributed to
the insufficient oxidation of PDMS due to the insufficient oxygen in
the gas mixture. However, the surface roughness again increased as
the O2 flow rate was increased above 1 slm, which was attributed to
the gas-phase oxidation of PDMS leading to particle formation.
Also, in this experiment, even though some particles were observed
on the deposited SiO2 surface for some of the conditions, when
comparing with the SiO2 deposited using hexamethyldisilazane ob-
tained by a previous experiment,21 a smoother thin film could be
obtained possibly due to the characteristics of PDMS having less
chemically reactive precursors.
Sungkyunkwan University assisted in meeting the publication costs of
this article.
References
1. J. S. Jung, J. Y. Kwon, Y. S. Park, D. Y. Kim, H. S. Cho, K. B. Park, W. Xianyu,
H. Yin, and T. Noguchi, J. Korean Phys. Soc., 45, S861 ͑2004͒.
2. J. Y. Lee and S. H. Lee, J. Korean Phys. Soc., 45, 558 ͑2004͒.
3. A. Gruniger and P. R. von Rohr, Surf. Coat. Technol., 174, 1043 ͑2003͒.
4. M. A. Liberman and A. J. Lichtenberg, Principles of Plasma Discharges and Ma-
terials Processing, Wiley Interscience, New York ͑1994͒.
5. J. H. Lee, C. H. Jeong, J. T. Lim, V. A. Zavaleyev, S. J. Kyung, and G. Y. Yeom,
Jpn. J. Appl. Phys., Part 1, 45, 8430 ͑2006͒.
6. K. Schmidt-Szalowski, Z. Rzanek-Boroch, J. Sentek, Z. Rymuza, Z.
Kusznierewicz, and M. Misiak, Plasmas Polym., 5, 173 ͑2000͒.
7. J. H. Lee, C. H. Jeong, H. B. Kim, J. T. Lim, S. J. Kyung, and G. Y. Yeom, Thin
Solid Films, 515, 917 ͑2006͒.
8. Y. Sawada, S. Ogawa, and M. Kogoma, J. Phys. D, 28, 1661 ͑1995͒.
9. S. R. Coulson, I. S. Woodward, J. P. S. Badyal, S. A. Brewer, and C. Willis,
Langmuir, 16, 6287 ͑2000͒.
Conclusion
A silicon oxide thin film was deposited using a pin-to-plate,
DBD-type, AP-PECVD with a PDMS ͑prepolymer: curing agent
= 10:1͒ polymer source/O2 gas at a substrate temperature lower than
50°C, with varying PMDS and O2 gas flow rates.
10. D. R. Lide, Handbook of Chemistry and Physics, 76th ed., CRC, Boca Raton, FL
͑1995͒.
11. L. O’Neill, L. A. O’Hare, S. R. Leadley, and A. J. Goodwin, Chem. Vap. Deposi-
tion, 11, 477 ͑2005͒.
12. Y. H. Lee, S. J. Kyung, C. H. Jeong, and G. Y. Yeom, Jpn. J. Appl. Phys., Part 2,
The increase of PDMS ͑bubbled through He flow at room tem-
perature͒ at a fixed O2 gas flow rate increased the deposition rate
due to the increased silicon source availability that was induced by
the increased PDMS flow rate. However, the surface roughness also
increased with increasing PDMS flow rate at a fixed O2 gas flow
rate, which was attributed to the increased porosity in the film as
seen by the increased Si–͑OH͒ bonding in the film, in addition to the
possible particle formation in the gas phase by a homogeneous re-
action. When the O2 gas flow rate was varied at a fixed PDMS flow
rate, the deposition rate peaked at a certain O2 flow rate, at which
point the surface roughness was minimized, possibly due to the
optimum dissociation and oxidation of PDMS having the lowest
Si–͑OH͒ bonding in the film. At an ac power of 5 kV ͑25 kHz͒ and
at a gas flow rate of PDMS ͑9 slm bubbled by He͒/O2 1 slm, a
44, L78 ͑2005͒.
13. D. C. Duffy, J. C. McDonald, O. J. A. Schueller, and G. M. Whitesides, Anal.
Chem., 70, 4974 ͑1998͒.
14. J. Pulpytel, F. Arefi-Khonsari, and W. Morscheidt, J. Phys. D, 38, 1390 ͑2005͒.
15. G. R. Nowling, M. Yajima, S. E. Babayan, M. Moravej, X. Yang, W. Hoffman, and
R. F. Hicks, Plasma Sources Sci. Technol., 14, 477 ͑2005͒.
16. F. Massines, N. Gherardi, A. Fornelli, and S. Martin, Surf. Coat. Technol., 200,
1855 ͑2005͒.
17. D. Trunec, Z. Navratil, P. Stahel, L. Zajickova, V. Bursikova, and J. Cech, J. Phys.
D, 37, 2112 ͑2004͒.
18. L.-Y. Chen and F. C.-N. Hong, Appl. Phys. Lett., 82, 3526 ͑2003͒.
19. M. R. Alexander, R. D. Short, F. R. Jones, W. Michaeli, and C. J. Blomfield, Appl.
Surf. Sci., 137, 179 ͑1999͒.
20. J. Viard, E. Beche, D. Perarnau, R. Berjoan, and J. Durand, J. Eur. Ceram. Soc.,
17, 2025 ͑1997͒.
21. J. H. Lee, Thuy T. T. Pham, Y. S. Kim, J. T. Lim, S. J. Kyung, and G. Y. Yeom, J.
Electrochem. Soc., 155, D163 ͑2008͒.
Downloaded on 2014-11-12 to IP 142.58.129.109 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).