The possibility of carbonyl fluoride as a new CVD chamber cleaning gas

Article abstract of DOI:10.1149/1.1669010

Carbonyl fluoride (COF₂) has been investigated as an alternative gas for plasma-enhanced chemical vapor deposition (PECVD) chamber cleaning in order to reduce greenhouse gases emitted from the cleaning process in semiconductor manufacturing. The cleaning performance of COF₂ and the environmental impact of its exhaust gases were evaluated using an experimental plasma tool. The results indicated that the cleaning performance of COF ₂ was equivalent to that of conventional C₂F₆. Furthermore, it was confirmed that the use of COF₂ would enable the reduction of global warming emissions by over 95% relative to the use of C ₂F₆, and thus COF₂ is considered to be a promising alternative cleaning gas.

Full text of DOI:10.1149/1.1669010

Journal of The Electrochemical Society, 151 ͑5͒ G297-G301 ͑2004͒
G297
0
013-4651/2004/151͑5͒/G297/5/$7.00 © The Electrochemical Society, Inc.
The Possibility of Carbonyl Fluoride as a New CVD Chamber
Cleaning Gas
Yuki Mitsui, b Yutaka Ohira,a,d Taaisuke Yonemura, Tsuyoshi Takaichi,a
a,c,z
a,e
Akira Sekiya, and Tatsuro Beppu
a
b
Research Institute of Innovative Technology for the Earth, and Research Center of Developing
Fluorinated Greenhouse Gas Alternatives, National Institute of Advanced Industrial Science and
Technology, Ibaraki 305-8565, Japan
c
Asahi Glass Company, Limited, Chemicals Company, General Division Technology, Chiba 290-8566, Japan
Daikin Industries, Limited, Chemical Division, Ibaraki 305-0841, Japan
Kanto Denka Kogyo Company, Limited, New Products Development Division, Tokyo 100-0005, Japan
d
e
Carbonyl fluoride (COF ) has been investigated as an alternative gas for plasma-enhanced chemical vapor deposition ͑PECVD͒
2
chamber cleaning in order to reduce greenhouse gases emitted from the cleaning process in semiconductor manufacturing. The
cleaning performance of COF and the environmental impact of its exhaust gases were evaluated using an experimental plasma
2
tool. The results indicated that the cleaning performance of COF2 was equivalent to that of conventional C F . Furthermore, it
2
6
was confirmed that the use of COF would enable the reduction of global warming emissions by over 95% relative to the use of
2
C F , and thus COF is considered to be a promising alternative cleaning gas.
2
6
2
©
2004 The Electrochemical Society. ͓DOI: 10.1149/1.1669010͔ All rights reserved.
Manuscript submitted March 17, 2003; revised manuscript received October 30, 2003. Available electronically March 17, 2004.
Perfluorocarbons ͑PFCs͒ such as CF and C F , etc., with strong
dustrial Hygienists ͑ACGIH͒ of COF is 2 ppm, which is larger than
4
2
6
2
infrared absorption coefficients and extremely long lifetimes in the
atmosphere are, therefore, greenhouse gases and their global warm-
ing potentials ͑GWPs͒ are quite high. They are mainly used for the
that of arsine ͑0.05 ppm͒ and of germane ͑0.2 ppm͒, both of which
are widely used in semiconductor manufacturing. Thus, compared
1
with these substances, the toxicity of COF is considered to be mod-
2
plasma enhanced chemical vapor deposition ͑PECVD͒ chamber
cleaning process in semiconductor manufacturing, and are contained
in the exhaust gases from the process as well. PFCs have roughly
ten thousands times more impact on global warming than carbon
dioxide because of their long lifetimes and high GWP values. Ac-
cordingly, the PFC emissions pose serious problems for the global
environment. With a view to reducing the quantity of greenhouse
gases emitted from the cleaning process, research and development
of alternative gases2 and abatement systems4 for exhaust gases
have been conducted.
The objective of the present work is to research and develop
alternative CVD chamber cleaning gases, which show performances
equivalent to conventional gases, and thus enable a great reduction
in greenhouse gas emitted from the cleaning process. At first, we
searched for a substitute gas among substances that had low GWP
values. However, the number of substances whose GWP values have
been measured is limited. Substances that are hydrolyzed with water
or moisture in the atmosphere are assumed to have short atmo-
spheric lifetimes and low GWP values. Thus the hydrolysis property
was one of important characteristics in the present search for alter-
native.
erate. In addition, COF is neither explosive nor combustible. There-
2
fore, COF can be handled with the existing safety control standards
2
for toxic gases implemented by semiconductor manufacturers. What
is more, since COF flows through the downstream of the chamber
2
as the component of the exhaust during C F cleaning, it is unlikely
2
6
that COF cleaning causes more trouble to the system than C F
2
2 6
cleaning.
,3
Experimental
The performance of the cleaning gas is usually evaluated at the
time of cleaning the deposits on the entire inside of the chamber
after the actual PECVD process. In the present study, the cleaning
performances and the environmental impact of its exhaust gases
were evaluated using the experimental plasma tool ͑manufactured
by Anelva Corporation͒ and the system as shown in Fig. 1a. The
chamber is equipped with circular parallel plate electrodes ͑the up-
per electrode with a diameter of 44 cm, and lower electrode with a
diameter of 42 cm͒. To the showerhead-type upper electrode, radio
frequency ͑rf͒ power of 13.56 MHz was applied. The mass flow
controller ͑MFC͒ set the flow rates of the tested gas and the additive
gas. The gases were introduced into the chamber through the upper
electrode showerhead. To dilute the exhaust from the chamber, ni-
COF is one of the by-products generated during plasma clean-
2
ing using perfluorocarbons. When a mixed gas of C F and O is
2
6
2
trogen was fed at the inlet of the dry pump at 15.5 L/min. SiO and
2
used for cleaning silicon compounds, the exhaust gas is known to
SiN are the major insulation films deposited by CVD in the semi-
x
contain nondecomposed C F , recombined CF , SiF produced
2
6
4
4
conductor large scale integration ͑LSI͒ process. In the present paper,
from silicon compounds, CO2 oxidized perfectly, which released
fluorine completely, CO oxidized imperfectly, which released fluo-
SiO film or/and a quartz plate were selected as samples because
2
SiO was revealed to requires longer time for an etching reaction
2
rine completely, COF which released fluorine incompletely, HF and
2
than SiN in the preliminary experimental evaluation.
x
F generated by the plasma but not utilized. Since COF , SiF , F ,
2
2
4
2
The cleaning performances was evaluated by two kinds of etch
rates, the etch rate on the lower electrode and the etch rate toward
the chamber wall without the pattern. For the measurement of the
and HF are toxic gases, they have been abated with water scrubbers.
The hydrolysis characteristic of COF is well known. The cleaning
2
performance of COF has been in doubt because it is not been com-
2
etch rate on the lower electrode, approximately 10,000 Å SiO thin
2
pletely dissociated by the plasma reaction and has a small number of
film deposited in advance on a 15 cm diam silicon wafer was etched
as a sample as shown in Fig. 1a. The film thickness was measured at
fluorine atoms in its molecule. However we selected COF for its
2
hydrolysis properties as the candidate of an alternative gas to study
in the present work.
9
fixed points on the wafer by a spectroscopic reflectometric film
thickness measuring instrument ͑model 3000 manufactured by Na-
nometorics, Japan͒ before and after the plasma discharged. Etch
rates were calculated by the differences in the film thickness and
etching times. As well as the etch rate, within-wafer uniformity was
obtained by the following formula. Poor uniformity means that the
The toxicity index, TLV ͑TWA͒ in American Conference of In-
z
E-mail: mitsui@cvd-rite.gr.jp
Downloaded on 2015-02-17 to IP 138.251.14.35 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

G298
Journal of The Electrochemical Society, 151 ͑5͒ G297-G301 ͑2004͒
Figure 2. Etch rate and uniformity vs. total gas flow.
where i indexes gases in the exhaust, Q /min is the mass flow in
i
kilogram/min during etching, and GWP100i is the 100 year inte-
grated global warming potential of the gas. GWP100i values in the
present paper were used the following values revised in 2001: CF4 ,
5
700; C F , 11900; C F , 8600; c-C F 10,000; CO , 1. Although
2 6 3 8 4 8 2
the GWP of COF has not been estimated, they react with water
2
very quickly and produce CO and HF. Their contribution to global
2
warming is as same as CO at most. Thus the GWP values in this
Figure 1. ͑a, top͒ Schematic diagram of experimental setup. ͑b, bottom͒
2
Quartz plate setting for etch rates toward chamber wall.
present paper were assumed to be the same with that of CO ͑ϭ1͒.
2
Tested gas dissociation efficiencies was defined with the follow-
ing formula
plasma is irratically generated, which is not desirable in the CVD
chamber cleaning process.
Dissociation efficiency͑%͒
ϭ 100 Ϫ tested gas concentration during etching/
͗
Within-wafer uniformity͑%͒
tested gas concentration before etching ϫ 100
͓3͔
ϭ ͑maximum etch rate Ϫ minimum etch rate͒/
͑
maximum etch rate ϩ minimum etch rate͒ ϫ 100
͘
͓1͔
The items common to the experimental conditions are the tested
gases, COF , candidate substitute gas; and C F , C F , c-C F
2
2
6
3
8
4 8
reference gas for evaluation. Oxygen is the additive gas. RF power
is 750 W ͑power density: 0.96 W/cm ͒. The lower electrode tem-
Analysis of the exhaust gases from the etching reaction between
2
the cleaning gas and quartz plate was performed by the Fourier
transform infrared ͑FTIR͒ spectroscopy ͑Infinity Gold by Mattson͒
with a 2 cm absorption cell with barium fluoride windows. As
shown in Fig. 1a, a 15 cm quartz plate was placed on the lower
electrode and exhaust gases from the chamber were fed into the
FTIR spectroscope through the dry pump outlet with dilute nitrogen.
Quantitative analysis of FTIR for the exhaust gas components was
conducted using the calibration curve measured with the reference
gas. The low values of the concentrations are attributable to dilution
with nitrogen at 15.5 L/min, and the concentration values were mea-
sured during continuous etching while absorption peaks ͑concentra-
tions of components͒ were constant.
perature is 300°C. The interelectrode distance is 50 mm. The etching
͑
discharge͒ time is 30 s for etch rate measurement using SiO film
2
on silicon wafer and about 200 s for the exhaust gas analysis using
a quartz plate by FTIR. Total gas flow rate (tested gas ϩ oxygen)
and pressure value were recorded for each experiment.
Results and Discussions
As mentioned in the introduction, hydrolysis characteristics of
COF were attractions to survey the performance as an alternative
2
for PECVD chamber cleaning gas of low GWP value. In order to
clarify the performance, a few experiments were performed.
The effects of the total gas flow rate ͑the tested gas and the
Although the etch rate toward chamber wall can be measured
directly by placing a SiO film wafer piece on the chamber inner
wall, another way was designed for the present study. The etch rate
additive gas, O ) on the etch rate are shown in Fig. 2. In this ex-
2
2
7
periment, C F and COF were fixed at 50 and 60% concentration
2
6
2
of SiO film on the electrode and the SiF concentration analyzed in
and at pressures of 250 and 350 Pa, respectively. Although the maxi-
2
4
8
the exhaust gas is proportional. SiF is also produced by the etching
mum etch rate of C F was higher than that of COF , the results of
4
2
6
2
reaction between the cleaning gas and the quartz plate placed up-
right near the chamber wall as shown in Fig. 1b. Therefore, the SiF4
flow rate can be considered to be proportional to the etch rate toward
the chamber wall. The flow rates were obtained by converting the
both gases show the same tendency that the etching performance for
C F and COF decrease slightly as the total gas flow rate is in-
creased. The etch rates were higher at a pressure of 350 Pa than at
2
6
2
2
50 Pa. However, in the case of only C F at 350 Pa, the uniformity
2
6
SiF concentrations into the mmol/min unit.
deteriorated when the total flow rate was 700 standard cubic centi-
4
In order to evaluate the environmental impacts of the cleaning
gases, as the index of global warming gas emissions during the
etching of the quartz plate, kilograms of carbon equivalent per
minute ͑kgCE/min͒ was used, which has been defined with the fol-
lowing formula
meters per minute ͑sccm͒ or more and the plasma generation be-
came unstable. Meanwhile, in the case of COF at 350 Pa, the uni-
2
formity was maintained even at 900 sccm. This indicates that COF2
achieves more stable plasma generation than C F , which is advan-
2
6
tageous for the actual cleaning condition. Based on these experimen-
tal results, basic experimental conditions were fixed at a total gas
flow rate of 300 sccm and at pressure of 250 Pa.
kgCE ϭ ⌺͑12/44͒͑Q /min͒͑GWP100i͒
͓2͔
i
Downloaded on 2015-02-17 to IP 138.251.14.35 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, 151 ͑5͒ G297-G301 ͑2004͒
G299
Figure 5. Exhaust gas analysis of COF vs. gas concentration.
2
pared, it was found that the exhaust gases by using COF did not
2
include C F , and that the CF concentration is extremely low in the
2
6
4
case of COF rather than that of C F . It seems that the amount of
2
2 6
Figure 3. Etch rate vs. gas concentration ͑%͒.
CF produced as by-product is slight in the case of COF for the
4
2
plasma reaction because of the large binding energy of CϭO in
COF molecules. Although the dissociation ratio of COF is equiva-
2
2
The effects of the introductory concentrations on etch rates are
shown in Fig. 3. The gases were COF , C F , and C F . In this
lent to that of C F , COF in the exhaust reacts with water easily
2
6
2
2
2
6
3 8
and can be removed by an existing scrubber. This suggests that the
amount of greenhouse gas emitted will be reduced greatly, when
evaluation, the total gas flow rate and chamber pressure were fixed
at 300 sccm and 250 Pa. The maximum etch rate on the electrode of
COF2 was slightly lower than C F and roughly the same level
COF is used for cleaning. It is also points out that compared with
2
2
6
exhaust gases emitted from the actual CVD cleaning process in the
4
,5
compared with C F and c-C F . The etch rate formed curves
3
8
4 8
manufacturing line, the C F and SiF concentrations are low. The
2
6
4
with an upper convex peak in relation to the concentration. The
maximum etch rates occurred at approximately 50% concentration
for C F , at approximately 80% concentration for COF , at ap-
low C2F6 concentration and the high dissociation efficiency were
attributable to the small amount of total gas flow rate. The low SiF4
concentration was attributable to a small etching area because only
one quartz plate was placed on the lower electrode. The concentra-
tion of CO was also low because it was estimated that the oxidation
2
6
2
proximately 30% concentration for C F , and at approximately 25%
3
8
concentration for c-C F . The concentrations at which the peaks
4
8
appeared depended on the number of carbon atoms; when the num-
ber of carbon atoms increases from 1 to 4, more oxygen is required,
was well done in this plasma tool and/or more COF was recom-
2
bined with surplus fluorine attributable to a small etching area. ͑CO
reacts with F2 very quickly.͒ As shown in Fig. 4 and Fig. 5, the
increase in introduction concentration of the tested gas led to the
6
lowering the concentration at which the peak appears. The effects
of the introductory concentration on the exhaust gas components are
shown in Fig. 4 for C F and Fig. 5 for COF . The total gas flow
2
6
2
increase in the CF concentration in the exhaust. Carbons containing
4
rate was fixed at 300 sccm, and the chamber pressure was fixed at
50 Pa. The exhaust gas from the plasma reaction with the quartz
cleaning gases are finally oxidized producing CO and fluorine radi-
2
2
cals. However, side reaction of CF consumes fluorine radicals in
4
plate on the lower electrode was analyzed by FTIR. The percentage
part is small because the exhaust gas is diluted with 15.5 L/min
nitrogen. The components’ percentage does not add up to 100%
because nitrogen, oxygen, and fluorine are not analyzed by FTIR
without their infrared absorption. When Fig. 4 and Fig. 5 were com-
vain to cleaning purpose. Thus, the decrease in the introductory
oxygen concentration was seen to cause the decrease of CO and the
2
increase of CF4 and finally the decrease in fluorine radicals and
lowering in the etch rate.
Comparison of kgCE/min in the exhaust gases is shown in Fig. 6.
Specifically, we set the concentration of C F at 50%, C F at 30%,
2
6
3 8
c-C F at 25%, and the concentration of COF at 60 and 80%. On
4
8
2
this evaluation, the total gas flow rate and chamber pressure were
fixed to be 300 sccm and at 250 Pa. The plasma reaction with SiO2
using COF reduced the kgCE/min of greenhouse gases emitted by
2
over 95% relative to using of C F . COF generated hardly any CF
2
6
2
4
as by-products during the plasma cleaning reaction compared with
those generated by using C F . Greenhouse gases emissions were
2
6
reduced by using less C F and c-C F than by using COF because
3
8
4
8
2
C F and c-C F could not avoid recombining with CF and high
3
8
4
8
4
GWP gases themselves were not decomposed perfectly. Moreover,
as shown in Fig. 3, the COF maximum etch rate appeared when the
2
concentration was 80%. When the concentration was lowered to
60%, the etch rate fell by only approximately ca. 10% from the
maximum whereas the CF quantity in the exhaust gas dropped by
4
approximately 50% as shown in Fig. 6. Thus, it was concluded that
the optimum concentration of COF is around 60% in this study.
2
Figure 4. Exhaust gas analysis of C F vs. gas concentration.
The C F maximum etch rate appeared when the concentration was
2
6
2 6
Downloaded on 2015-02-17 to IP 138.251.14.35 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

G300
Journal of The Electrochemical Society, 151 ͑5͒ G297-G301 ͑2004͒
Figure 8. The kgCE/min vs. the total gas flow rate.
Figure 6. Comparison of kgCE/min in the exhaust gases.
exhaust was analyzed by FTIR. In addition, the etch rates ͑quartz
plate͒ on the lower electrode are also shown in Fig. 10 under the
same experimental conditions. They showed the pressure dependen-
cies of the SiF4 quantity produced. C2F6 with a concentration of
5
0%, and many cleaning recipes adopt this concentration. Thus, we
set the concentration of C F at 50% as the reference.
2
6
Comparison of the effects of the total gas flow rate on the disso-
ciation efficiency is shown in Fig. 7. The increase in the total gas
flow rate from 300 to 900 sccm led to similar decreases in dissocia-
tion efficiency of both C F and COF from 60 to 40%. The de-
crease in etch rates due to the increase in the total gas flow rate as
shown in Fig. 2 is also due to the decrease in dissociation efficiency.
The decrease in dissociation efficiency seems attributable to a
shorter time during which gas molecules receive discharge energy,
due to the increase in the total gas flow rate.
Comparison of the effects of the total gas flow rate on the kgCE/
min is shown in Fig. 8. Even when the total gas flow rate of COF2
increased, CF4 was generated, and the kgCE/min remained slight
because the GWP value of nondecomposed COF was estimated to
be the same as that of CO . In contrast, the increase in the total gas
5
0% and COF with a concentration of 60% were tested.
2
The total gas flow rate was 300, 500, 700, 900 ͑sccm͒ and the
chamber pressure was 150, 250, 350 ͑Pa͒.
2
6
2
It was found that the etch rates toward the chamber wall of COF2
were 20 to 25% higher than that of C F , although the etch rates on
2
6
the electrode of COF was approximately 10% lower than that of
2
C F at the same pressure. As shown in Fig. 9 and 10, the COF
2
6
2
cleaning performance at 350 Pa was equivalent to that of C F at
2
6
250 Pa in consideration the etch rates both on the electrode and
toward the chamber wall. For amount of SiF quantity, the etch rates
4
toward the chamber wall were smaller than the etch rates on the
lower electrode because the quartz plate accounts for only approxi-
mately one-tenth of the entire chamber wall circumference and only
the 50 mm edge portion of the 15 cm quartz plate upright was
exposed between the electrodes.
2
2
flow rate of C F resulted in the decrease in the dissociation effi-
ciency, which led to an increase in nondecomposed C F and in the
2
6
2
6
In the case of C F and COF , the effect of the total flow rate on
2
6
2
kgCE/min. The decrease in dissociation efficiency due to the in-
crease in the total gas flow rate further consolidated the superiority
of COF to C F in terms of environmental impact.
Comparison of the etch rates toward the chamber wall are shown
in Fig. 9. The quartz plate was placed upright near the chamber wall.
For the gases tested at the fixed concentrations, the chamber pres-
both etch rates was slight, but the effect of the chamber pressure was
great. Specifically, as the chamber pressure decreased, the etch rate
toward the chamber wall increased whereas the etch rate on the
electrode decreased. Although it is difficult to identify substances to
be cleaned in the case that they are distributed unevenly, it is effec-
tive to set a pressure for the cleaning recipe optimized for deposition
conditions and for uneven distribution.
2
2 6
sure and the total gas flow rate were varied as below, and SiF in the
4
Figure 7. The dissociation efficiency vs. total gas flow rate.
Figure 9. Comparison of the etch rates toward chamber wall.
Downloaded on 2015-02-17 to IP 138.251.14.35 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, 151 ͑5͒ G297-G301 ͑2004͒
G301
product, CF , was remarkably low and the GWP of nondecomposed
4
COF was estimated to be the same as that of CO for its hydrolysis
2
2
property.
Thus, in view of the characteristics of COF , it was concluded
2
that COF is a strong candidate to replace C F and that the use of
2
2 6
COF would enable a great reduction in greenhouse substances gen-
2
erated during cleaning. Provided adequate safety measures are
implemented, the use of COF as a cleaning gas would greatly re-
2
duce the greenhouse gas emission from the semiconductor industry.
Acknowledgments
This study was conducted as part of the research projects com-
missioned by New Energy and Industrial Technology Development
Organization ͑NEDO͒.
National Institute of Advanced Industrial Science and Technology as-
sisted in meeting the publication costs of this article.
Figure 10. Etch rate on the lower electrode vs. the chamber pressure.
References
1
2
. The Science of Climate Change IPCC Climate Change, ͑1995͒.
. W. Entley, W. Hennessy, and J. Langan, Electrochem. Solid-State Lett., 3, 99
Conclusion
͑
2000͒.
3
4
5
. L. Pruette, S. Karecki, R. Reif, L. Tousignant, W. Reagan, S. Kesari, and L.
Zazzera, J. Electrochem. Soc., 147, 1149 ͑2000͒.
. C. Allgood, M. Mocella, H. Chea, and H. Sawin, J. Electrochem. Soc., 150, G112
͑2003͒.
. Y. Mitsui, Y. Kosano, T. Takaichi, T. Beppu, and A. Sekiya, in Proceedings of 8th
Annual International Semiconductor Environmental Safety and Health (ISESH)
Conference, Kenting, Taiwan ͑2001͒.
The CVD chamber cleaning performances of COF2 and C2F6
were evaluated by measuring the etch rates on the lower electrode
for SiO film and the etch rates toward the chamber wall for the
2
quartz plate placed upright. The results showed that the performance
of COF was equivalent to that of C F in consideration of the etch
2
2 6
rates both on the electrode and toward chamber wall. It was also
confirmed that the generation of plasma is more stable and the
plasma tended to expand more toward the chamber wall in the case
6. Y. Kosano, Y. Mitsui, T. Takaichi, A. Sekiya, and T. Beppu, in Proceedings of 7th
Annual International Semiconductor Environmental Safety and Health (ISESH)
Conference, Dresden, Germany ͑2000͒.
7
. S. Suzuki, E. Wani, T. Sunada, K. Shibata, T. Morimura, Y. Mitsui, Y. Kosano, T.
Takaichi, T. Beppu, and A. Sekiya, in The Global Climate Change: A Coordinated
Response by Electrochemistry and Solid-State Science and Technology, A. Wieck-
owski, Editor, PV 2000-20, p. 231, The Electrochemical Society Proceedings Se-
ries, Pennington, NJ ͑2001͒.
of COF than in the case of C F .
2
2 6
Their environmental impacts during the cleaning were evaluated
by the kgCE/min of global warming gas based on exhaust gases
analysis. It was found that the kgCE in the case of COF was greatly
2
8. T. Yonemura, K. Fukae, Y. Ohira, Y. Mitsui, T. Takaichi, A. Sekiya, and T. Beppu,
reduced compared to that in the case of C2F6 because the by-
J. Electrochem. Soc., 150, G707 ͑2003͒.
Downloaded on 2015-02-17 to IP 138.251.14.35 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).