Content uploaded by Takayuki Watanabe
Author content
All content in this area was uploaded by Takayuki Watanabe
Content may be subject to copyright.
1 23
Plasma Chemistry and
Plasma Processing
ISSN 0272-4324
Volume 30
Number 6
Plasma Chem Plasma Process
(2010) 30:813-829
DOI 10.1007/s11090-010-9259-
y
Decomposition Mechanism of Fluorinated
Compounds in Water Plasmas Generated
Under Atmospheric Pressure
1 23
Your article is protected by copyright and
all rights are held exclusively by Springer
Science+Business Media, LLC. This e-offprint
is for personal use only and shall not be self-
archived in electronic repositories. If you
wish to self-archive your work, please use the
accepted author’s version for posting to your
own website or your institution’s repository.
You may further deposit the accepted author’s
version on a funder’s repository at a funder’s
request, provided it is not made publicly
available until 12 months after publication.
ORIGINAL PAPER
Decomposition Mechanism of Fluorinated Compounds
in Water Plasmas Generated Under Atmospheric
Pressure
Narengerile
•
Hironori Saito
•
Takayuki Watanabe
Received: 15 May 2010 / Accepted: 5 October 2010 / Published online: 24 October 2010
Ó Springer Science+Business Media, LLC 2010
Abstract Decomposition mechanism of HFC-134a, HFC-32, and CF
4
in water plasmas
at atmospheric pressure has been investigated. The decomposition efficiency of 99.9% can
be obtained up to 3.17 mol kWh
-1
of the ratio of hydrofluorocarbon (HFC) feed rate to the
arc power and 1.89 mol kWh
-1
of the ratio of perfluorocarbon (PFC) feed rate to the arc
power. The species such as H
2
, CO, CO
2
,CH
4
, and CF
4
were detected from the effluent gas
of both PFC and HFC decomposition. However, CH
2
F
2
and CHF
3
were observed only in
the case of HFC decomposition. The HFC and PFC decomposition generate CH
2
F, CHF
x
(x:1–2)
, and CF
y (y:1–3)
radicals, then those radicals were subsequently oxidized by oxygen,
leading to CO and CO
2
generation in the excess oxygen condition. However, when there is
insufficient oxygen available, those radicals were easily recombined with fluorine to form
by-product such as CH
2
F
2
, CHF
3
, and CF
4
.
Keywords Hydrofluorocarbons Perfluorocarbons Water plasmas Thermal plasmas
Introduction
Industrial development is accompanied with increase of environmental deterioration such
as global warming which is a big problem for human, animals, and plants in recent years.
Perfluorocarbon (PFC) used in the semiconductor industry and hydrofluorocarbon (HFC)
used in refrigeration, collectively called fluorinated compounds, have been identified as
greenhouse gas due to their large global warming potential (GWP) and long atmospheric
lifetimes. In order to save the environment, the Kyoto protocol was adopted at the third
session of the conference of the parties in Kyoto, Japan in 1997. The Kyoto treaty commits
industrialized nations to reducing emissions of greenhouse gases including HFC and PFC
by years 2008–2012 to 5% below 1990 levels.
Narengerile H. Saito T. Watanabe (&)
Department of Environmental Chemistry and Engineering, Tokyo Institute of Technology, Yokohama
226-8502, Japan
e-mail: watanabe@chemenv.titech.ac.jp
123
Plasma Chem Plasma Process (2010) 30:813–829
DOI 10.1007/s11090-010-9259-y
Author's personal copy
Conventional process for reducing emissions of fluorinated compounds, e.g. alternative
chemistry process, recovery and recycle, and process optimization, have proved to be
inadequate. To deal with the problem, processes for the destruction of PFC and HFC were
performed by means of abatement methods including combustion, catalytic destruction,
and plasma abatement. Combustion technologies that have historically been used for the
destruction of fluorinated compounds may fail to meet stringent environmental regulations
because of the generation of other hazardous substances in the effluent gas [1, 2]. The
catalytic process for controlling PFC emission involves passing the PFC-laden stream over
the catalyst [3, 4]. For the catalyst to be employed in a commercial PFC abatement process,
it must be both high reactive and durable, i.e. able to maintain a high decomposition
efficiency for an extended period of time. Plasma treatment has been widely studied and
applied to waste destruction, and has the advantages of high efficiency, flexibility, and
environmental compatibility.
Atmospheric nonthermal plasma process was attempted for the decomposition of highly
concentrated fluorinated compounds at low power levels. However, the nonthermal plasma
techniques showed some drawbacks such as low decomposition efficiency and generation of
harmful compounds if applied without catalyst. Jasinski et al. [5, 6] proposed the use of low-
power (\100 W) microwave discharges, operating in air plasma at atmospheric pressure, for
decomposition of highly concentrated fluorinated compounds. Their results showed that the
decomposition efficiencies were reached almost 100% for the fluorinated compounds such
as dichlorodifluoromethane (CFC-12), chlorodifluoromethane (HCFC-22), 1,1,1,2-tetraflu-
oroethane (HFC-134a), and trichlorofluoromethane (CFC-11) at a feed rate of 1.0 L min
-1
.
However, the electrical energy consumption was as low as 0.5 mol kWh
-1
. Furthermore,
the toxic compounds e.g. phosgene (COCl
2
) and carbonyl fluoride (COF
2
) were detected in
the effluent gas. Futamura and Annadurai [7] studied on the HFC decomposition by using
silent discharge plasma and surface discharge plasma in combination with MnO
2
and TiO
2
–
SiO
2
catalysts. Although these catalysts showed high activity for the HFC oxidation, their
irreversible deactivation was very severe due to fluorination of the catalysts.
Thermal plasmas can offer distinct advantages for waste treatment process; high
enthalpy to enhance the reaction rate, oxidation or reduction atmospheres in accordance
with required chemical reactions, and rapid quenching (10
5
–10
6
Ks
-1
) to produce
chemical non-equilibrium compositions. The PLASCON process developed by CSIRO in
Australia is one good example of thermal plasma waste treatment [8–10]. The arc was
generated by direct current (DC) discharge with the arc power from several kW up to
hundreds of kW. The plasma supporting gas was argon. This process can destroy a wide
range of fluorinated compounds including HFC, PFC, and chlorofluorocarbon (CFC), for
which decomposition efficiency was at levels above 99.99%. Ozawa et al. [11] studied on
decomposition of HFC-134a using water plasma. In order to obtain high decomposition
efficiency, they injected HFC-134a with a mixture of air and steam. When the steam and
air feed rate were set at 12 and 90 L min
-1
, respectively, the highest decomposition
efficiency of 99.99% was obtained at 30 A. Kim et al. [12] used a non-transferred DC arc
torch to decompose polychlorinated biphenyl (PCB) mixture waste using steam as the
plasma supporting gas. For the decomposition of PCB mixture waste containing 15%
trichlorobiphenyl (C
12
H
7
C
l3
), 12% tetrachlorobiphenyl (C
12
H
6
C
l4
), and 73% CCl
4
, the
decomposition efficiency of 99.999% had been obtained at feed rate of 19 g s
-1
(0.18 L min
-1
) at an arc power of 100 kW. Glocker et al. [13] decomposed the CFC by
using DC arc torch with steam as main plasma gas and argon as protecting gas in a power
range of 8-32 kW. However, the steam plasma process requires complicated system
including the heating-up of the steam feeding line for preventing from condensation.
814 Plasma Chem Plasma Process (2010) 30:813–829
123
Author's personal copy
Hrabovsky et al. [14, 15] developed hybrid plasma torch with combined stabilization of
electric arc by water vortex and gas flow. The cathode section was stabilized by a vortex
gas flow. Plasma flows through the cathode nozzle into the torch where the arc column was
surrounded by a water vortex. The water vortex was formed in three cylindrical segments
with a tangential water injection as it is in water-stabilized torches. The torch was operated
at an arc power from 80 to 300 kW with an exit centerline plasma velocity from 2.0 to
6.5 km s
-1
. The system was mainly studied for pyrolysis and gasification of waste for
production of syngas because it can produce oxygen-hydrogen plasma with extremely high
enthalpies.
We developed a DC water plasma torch which with a cathode of hafnium embedded
into a copper rod and a nozzle-type copper anode in our previous work [16]. The hafnium
used as cathode material can overcome the erosion problems and achieve a long operating
time in oxidation atmospheric. The torch can generate stable 100%-water plasmas using
DC discharge at the arc power of 1.0 kW without any steam generator or gas supply
system. Using the water plasma produced by the plasma torch, we succeed in decomposing
liquid waste of alcohol solutions [17] and gaseous waste such as HFC and PFC [18, 19]. In
References [18, 19], the decomposition efficiency of 99.9% can be obtained up to
1.54 mol kWh
-1
of the ratio of HFC-134a feed rate to the arc power where the maximum
feed rate was 0.59 L min
-1
at an arc power of 0.91 kW.
The objective of this study is to decompose fluorinated compounds by water plasmas
generated under atmospheric pressure and to investigate the decomposition mechanism of
fluorinated compounds in the water plasmas. The experiments have been performed with
different types of fluorinated compounds: HFC-134a and difluoromethane (HFC-32), and
tetrafluoromethane (CF
4
). In the first part, the optimum decomposition conditions were
determined by calculating the thermodynamic equilibrium in order to forecast and design
the decomposition process. In the second part, HFC-134a, HFC-32, and CF
4
were
decomposed by water plasmas generated by DC discharge at atmospheric pressure. The
decomposition efficiency obtained for each compound were compared and discussed. The
influence of operational conditions such as feed rate, additional oxygen, and arc power
were examined to help explain the decomposition mechanism. Finally, the decomposition
mechanism of fluorinated compounds in the water plasmas was discussed from the point of
view of chemical kinetics.
Thermodynamic Consideration
Thermodynamic equilibrium was calculated from the minimization of Gibbs free energy of
the system with assumption that chemical equilibrium is attained, which requires that
chemical reaction rates are sufficiently larger than rates of convection and diffusion. The
chemical composition obtained from the calculation is probably not accurate. However, the
thermodynamic equilibrium is useful to assess the important species in fluorinated com-
pounds decomposition. In this study, the thermodynamic equilibrium for HFC-134a with or
without water was calculated by FACT (Centre for Research in Computational Thermo-
chemistry, Canada). FACT is a computer program, linked to a database, for finding the
chemical equilibrium of a gas mixture by minimizing the Gibbs free energy. The calcu-
lations were performed with atmospheric pressure condition.
The equilibrium composition of the pyrolyzed HFC-134a as a function of temperature
between 500 and 6,000 K is shown in Fig. 1. As can be seen, decomposed species such as
F, H, HF, C(g), CF, and C
2
were observed in the decomposition system at a temperature of
Plasma Chem Plasma Process (2010) 30:813–829 815
123
Author's personal copy
6,000 K. CF, CF
2
, and CF
3
were generated in the largest quantities at 3,760, 2,900, and
2,690 K, respectively. The by-product of CF
4
and solid carbon appeared at room tem-
perature. It indicates that the HFC-134a was decomposed in high-temperature range, but
by-products such as CF
4
will be generated if the quenching step is not sufficiently fast.
Hence, optimization of the process is a prerequisite for HFC decomposition. The equi-
librium compositions of 1.0 mol HFC-134a with 2.0 mol H
2
O were calculated and shown
in Fig. 2. In this calculation, the ratio of HFC-134a to water was determined from the
stoichiometric coefficient of the following overall reaction (1).
CH
2
FCF
3
þ 2H
2
O ! 4HF þ H
2
þ 2CO ð1Þ
As shown in Fig. 2, CF, CF
2
,CF
3
, and further CF
4
can be eliminated by adding water. It
indicates that the unwanted by-products can be suppressed by adding water in the process
of HFC decomposition because CF, CF
2
, and CF
3
radicals are easily combined with F atom
or each other to generate PFC such as CF
4
or tetrafluoroethylene (TFE) in general.
Meanwhile, H
2
and CO were increased while C (s) was decreased.
The influence of H
2
OorO
2
supply on by-product formation during the HFC-134a
decomposition was studied by calculation of thermodynamic equilibrium. The total
amount of CF
y (y:1–4)
as a function of temperature is shown in Fig. 3. A large amount of
CF
y
was observed in the process of the HFC-134a thermal decomposition. With
increasing additional amount of O
2
, the total amount of CF
y
was greatly decreased over
2,400 K because CF
y
was oxidized by O
2
to generate CO in higher temperature range
([2,400 K). Comparatively, with an increasing amount of H
2
O, the total amount of CF
y
was decreased in lower temperature range (\3,000 K) because CF
y
was removed by H to
produce CH
4
and HF. It indicates that oxygen and hydrogen play an important role in
controlling the reformation of undesirable by-product when the decomposition of fluo-
rinated compounds.
10
-3
10
-2
10
-1
10
0
10
1
1000 2000 3000 4000 5000 6000
Mole [mol]
Temperature [K]
H
H
2
CF
2
C(s)
CF
4
HF
F
CF
C
3
C
2
C
2
H
CF
3
C(g)
Fig. 1 Equilibrium composition
of pyrolyzed 1.0 mol HFC-134a
816 Plasma Chem Plasma Process (2010) 30:813–829
123
Author's personal copy
Experimental
The decomposition system, which is the same as described in [19], is presented in Fig. 4.
The system consists of a DC power supply, a water plasma torch, a water absorption
material (Stainless steel fibers SUS 316L) for leading water into the arc region, a reaction
tube for decomposition of PFC and HFC, and a neutralization vessel for absorption of
fluorine. The torch is a DC non-transferred arc plasma generator of coaxial design with a
10
-3
10
-2
10
-1
10
0
10
1
1000 2000 3000 4000 5000 6000
Mole [mol]
Temperature [K]
H
O
H
2
O
H
2
CO
C
CO
2
C(s)
CH
4
HF
F
Fig. 2 Equilibrium composition
of 2.0 mol H
2
O and 1.0 mol
HFC-134a
10
-4
10
-3
10
-2
10
-1
10
0
10
1
10
2
2000 4000 6000 8000 10000
Total Amount of CF
y
[mol]
Temperature [K]
HFC-134a
HFC-134a + 1mol O
2
HFC-134a + 1mol H
2
O
HFC-134a + 2mol O
2
HFC-134a + 2mol H
2
O
Fig. 3 Total amount of CF
y (y:
1–4)
as a function of temperature
in different HFC-134a
decomposition system
Plasma Chem Plasma Process (2010) 30:813–829 817
123
Author's personal copy
cathode of hafnium embedded into a copper rod and a nozzle type copper anode. The
diameter of the hafnium was 1.0 mm. After arc ignition, the large heat from arc was
conducted by the copper anode to the water absorption material which is excellent in
electrical and optical performance and set between the anode case and evaporator. When
water was introduced into the torch with a controlled feed rate, plasma is created by
heating and ionization of steam that is produced by evaporation of water from the water
absorption material. Then the PFC and HFC were injected through the injection hole
towards the center of the water plasmas at the position of 6 mm from the nozzle exit. The
plasma and the injected gas are expected to undergo complex reactions in the reaction tube.
The hot gaseous mixture produced was quickly quenched in 1.0 M NaOH solution. The
quenching step is important to suppress by-product formation in the waste treatment
process.
The effluent gas was analyzed by a gas chromatograph (GC-TCD, Shimadzu GC-8A)
equipped with a thermal conductivity detector, and a quadruple mass spectrometer (QMS,
Ametek Dycor Proline). The recovery of fluorine was measured by neutralizing the
solution in the vessel. The decomposition efficiency and the recovery of fluorine were
calculated according to the following equations, respectively:
Decomposition efficiency ð%Þ
¼ 1
Concentration of the feeding fluorinated
compound in the effluent gas
,
Concentration of the feeding
fluorinated compound
!" #
100%: ð2Þ
Recovery of fluorine ð%Þ¼
Absorbed fluorine in
NaOH solution ðmol)
,
Total feeding
fluorine ðmol)
"#
100%: ð3Þ
The flow rate of plasma is depended on steam characteristics such as vapor pressure and
specific heat because the torch used in this study was a steam-stabilized torch. Hence, the
water feed rate for production of the plasma gas was controlled within the range of
40–65 mg s
-1
. The system was operated at atmospheric pressure with the arc power from
0.8 to 1.2 kW and the voltage from 110 to 150 V.
HFC Decomposition
The decomposition of HFC-32 and HFC-134a were carried out using the proposed
decomposition system. The effect of the HFC-32 feed rate on the recovery of fluorine and
the decomposition efficiency is shown in Fig. 5. The HFC-32 feed rate was changed from
0.33 to 1.0 L min
-1
. The experiments were operated at atmospheric pressure with the arc
current of 7 A. As can be seen, the feed rate had small effect on the decomposition
Fig. 4 Schematic diagram of
PFC and HFC decomposition
system
818 Plasma Chem Plasma Process (2010) 30:813–829
123
Author's personal copy
efficiency, but strongly affected the recovery of fluorine itself. The different between the
decomposition efficiency and the recovery of fluorine indicates the by-products formed in
the effluent gas. The composition of the effluent gas as a function of the feed rate is shown
in Fig. 6. The main components of the effluent gas were H
2
(49.3%), CO (27.7%), and CO
2
(10.1%) at feed rate of 1.0 L min
-1
. The insufficient oxygen leads to production of large
amount of CO and small amount of CO
2
if larger feed rates were applied.
The QMS spectra of the effluent gas produced by decomposing HFC-32 with different
feed rate are shown in Fig. 7. As shown in Fig. 7a, the intensities peak of 25(C
2
H
?
),
26(C
2
H
2
?
), 31(CF
?
), 32(CHF
?
), 33(CH
2
F
?
), 51(CHF
2
?
), and 52(CH
2
F
2
?
) were observed
0
20
40
60
80
100
0
20
40
60
80
100
0.0 0.2 0.4 0.6 0.8 1.0 1.2
Recovery of Fluorine
Decomposition Efficiency
Arc Current: 7 A
Recovery of Fluorine [%]
Decomposition Efficiency [%]
CH
2
F
2
Feed Rate [L min
-1
]
Fig. 5 Effect of feed rate on
fluorine recovery and HFC-32
decomposition efficiency
10
-4
10
-3
10
-2
10
-1
10
0
0.0 0.2 0.4 0.6 0.8 1.0 1.2
Mole Fraction [-]
CH
2
F
2
Feed Rate [L min
-1
]
H
2
CH
2
F
2
CH
4
CO
CO
2
Arc Current: 7 A
Fig. 6 Mole fraction of
compositions of effluent gas as a
function of HFC-32 feed rate
Plasma Chem Plasma Process (2010) 30:813–829 819
123
Author's personal copy
at feed rate of 1.0 L min
-1
. Comparing the spectra to the data in NIST library [20], the
intensities peak of 51(CHF
2
?
) and 52(CH
2
F
2
?
) correspond to CHF
3
and undecomposed
CH
2
F
2
, respectively. However, those complex intensities peak disappeared during
decomposition of HFC-32 at a low feed rate of 0.33 L min
-1
, indicating a complete
decomposition, shown in Fig. 7b.
The effect of the HFC-134a feed rate on the decomposition efficiency and the recovery
of the fluorine is shown in Fig. 8. The arc current was set at 7 A. As can be seen, when the
feed rate had been increased from that of 0.215 to 1.0 L min
-1
, the HFC-134a decom-
position efficiency remained at 97%, while the recovery from the fluorine greatly
10
-13
10
-12
10
-11
10
-10
Intensity [a.u.]
m/z [-]
C
+
H
2
O
+
F
+
CO
2
+
CO
+
H
2
+
CH
2
F
2
: 0.333 L min
-1
Arc Current: 7 A
CH
3
+
(b)
01020304050607080
01020304050607080
10
-13
10
-12
10
-11
10
-10
Intensity [a.u.]
m/z [-]
C
+
H
2
O
+
F
+
CH
CH
2
F
+
CHF
2
+
CH
2
F
2
+
H
2
+
CHF
CF
+
CO
2
+
CO
+
O
+
or CH
4
+
CH
2
F
2
: 1.0 L min
-1
Arc current: 7 A
C
2
H
2
+
C
2
H
+
(a)
Fig. 7 Mass spectra of gaseous product after decomposition of HFC-32 with different feed rate
(a) 0.33 L min
-1
,(b) 1.0 L min
-1
0
20
40
60
80
100
0
20
40
60
80
100
0.0 0.2 0.4 0.6 0.8 1.0 1.2
Recovery of Fluorine
Decomposition Efficiency
Decomposition Efficiency [%]
Recovery of Fluorine [%]
CH
2
FCF
3
Feed Rate [L min
-1
]
Arc Current: 7 A
Fig. 8 Effect of feed rate on
fluorine recovery and HFC-134a
decomposition efficiency
820 Plasma Chem Plasma Process (2010) 30:813–829
123
Author's personal copy
decreases. The composition of the effluent gas as a function of the feed rate is shown in
Fig. 9. As the HFC-134a feed rate increases, H
2
and CO
2
decrease while CO, CH
2
F
2
,CH
4
,
and undecomposed HFC-134a increase, because the plasma temperature as well as the
residence time in high temperature range was decreased by increasing feed rate. In addi-
tion, HFC feed rate increases carbon-containing species such as CF, CF
2
,CF
3
, CHF, and
CH
2
F in the water plasmas.
The composition of the effluent gas during decomposition of HFC-134a at arc current of
5 or 7 A was analyzed by QMS and shown in Fig. 10. In the both case, the HFC-134a feed
rate was controlled at 1.0 L min
-1
. As shown in Fig. 10a, the intensities peak of 2(H
2
2?
),
12(C
?
), 13(CH
?
), 14(CH
2
?
), 15(CH
3
?
), 16(O
?
), 18(H
2
O
?
), 19(F
?
), 26(C
2
H
2
?
),
28(CO
?
), 31(CF
?
), 32(CHF
?
), 33(CH
2
F
?
), 44(CO
2
?
), 50(CF
2
?
), 51(CHF
2
?
), 62(C
2
F
2
?
),
63(C
2
HF
2
?
), 69(CF
3
?
), 83(C
2
HF
4
?
) were observed at the arc current of 5 A. Comparing
the spectra to the data in NIST library, the intensities peak of 15(CH
3
?
), 26(C
2
H
2
?
),
33(CH
2
F
?
), 51(CHF
2
?
), and 69(CF
3
?
) correspond to CH
4
,C
2
H
2
,CH
2
F
2
, CHF
3
, and CF
4
,
respectively. The intensities peak of 63(C
2
HF
2
?
) and 83(C
2
HF
3
?
) were considered to be
derive from the undecomposed HFC-134a. When the arc current was increased from 5 to
7 A, the intensities peak of 31(CF
?
), 50(CF
2
?
), 63(C
2
HF
2
?
), 69(CF
3
?
), and 83(C
2
HF
3
?
)
disappeared, meanwhile the intensity peak of 33(CH
2
F
?
) was also decreased as shown in
Fig. 10b. Herein, the weak intensity peak of 33(CH
2
F
?
) correspond to CH
2
F
2
. The
unwanted by-products had been generated in the effluent gas even with a high arc current
of 7 A. In comparison to that of HFC-32, decomposition efficiency of the HFC-134a was
slightly larger than that of the HFC-32 decomposition because the HFC-134a included the
weak C–C single bond. The by-products of CH
2
F
2
and CHF
3
had been detected from the
effluent gas for both of the HFC-32 and HFC-134a decomposition, but CF
4
was only
detected in the case of the HFC-134a. Mole fractions of H
2
and CH
4
in the effluent gas
produced by HFC-32 decomposition were larger than that of HFC-134a.
10
-4
10
-3
10
-2
10
-1
10
0
0.0 0.2 0.4 0.6 0.8 1.0 1.2
Mole Faction [-]
CO
2
CO
H
2
CH
2
F
2
CH
4
Arc Current: 7 A
CH
2
FCF
3
CH
2
FCF
3
Feed Rate [L min
-1
]
Fig. 9 Mole fraction of
compositions of effluent gas as a
function of HFC-134a feed rate
Plasma Chem Plasma Process (2010) 30:813–829 821
123
Author's personal copy
The effect of additional O
2
on the HFC-134a decomposition efficiency and the recovery
of fluorine is shown in Fig. 11. The HFC-134a with oxidation gas of O
2
was injected
through injection hole into the water plasmas. The HFC-134a feed rate remained at
0.515 L min
-1
while the O
2
feed rate was increased. The arc current was set at 7 A. As the
additional O
2
feed rate increases, the decomposition efficiency remained nearly constant
while the recovery of fluorine increases. The composition of the effluent gas as a function
of O
2
feed rate is presented in Fig. 12. The increasing additional O
2
feed rate lead to
increasing CO
2
and decreasing CO because CH
2
F, CHF, and CF
y
were completely oxi-
dized by the excess oxygen to generate CO
2
. When additional O
2
feed rate was
0.5 L min
-1
, the compositions of the effluent gas were mainly CO
2
(47.8%), H
2
(32.4%),
10
-13
10
-12
10
-11
10
-10
m/z [-]
CO
+
CO
2
+
C
+
Arc Current: 5 A
CH
2
FCF
3
: 1.0 L min
-1
H
2
O
+
F
+
O
+
CH
3
+
CH
2
F
+
C
2
HF
4
+
CF
3
+
C
2
HF
2
+
C
2
F
2
+
CHF
2
+
CF
2
+
C
2
H
2
+
Intensity [a.u.]
H
2
+
CH
2
+
CF
+
CH
+
CHF
+
(a)
0 20 40 60 80 100 120
0 20 40 60 80 100 120
10
-13
10
-12
10
-11
10
-10
Intensity [a.u.]
m/z [-]
CO
+
CO
2
+
H
2
+
Arc Current: 7 A
CH
2
FCF
3
: 1.0 L min
-1
H
2
O
+
F
+
O
+
CH
3
+
CH
2
F
+
C
2
H
2
+
C
+
(b)
Fig. 10 Mass spectra of gaseous product after decomposition of 1.0 L min
-1
HFC-134a (a) at 5 A; (b) at
7A
0
20
40
60
80
100
0
20
40
60
80
100
0.0 0.1 0.2 0.3 0.4 0.5 0.6
Recovery of Fluorine
Decomposition Efficiency
Decomposition Efficiency [%]
Recovery of Fluorine [%]
Feed Rate of O
2
[L min
-1
]
Arc Current: 7 A
CH
2
FCF
3
: 0.515 L min
-1
Fig. 11 Effect of O
2
addition on
fluorine recovery and HFC-134a
decomposition efficiency
822 Plasma Chem Plasma Process (2010) 30:813–829
123
Author's personal copy
and CO (19.7%). The undecomposed HFC-134a and the by-product of CH
2
F
2
were sup-
pressed to level of 1%. Furthermore, CH
3
F and CF
4
were not detected from the effluent
gas.
PFC Decomposition
Decomposition of CF
4
by the water plasmas was tested because CF
4
is one of the most stable
gas among the PFC and its decomposition is extremely difficult. The effect of feed rate on the
decomposition efficiency and the recovery of fluorine are shown in Fig. 13. At the arc
10
-4
10
-3
10
-2
10
-1
10
0
0.0 0.1 0.2 0.3 0.4 0.5 0.6
Mole Faction [-]
Feed Rate of O
2
[L min
-1
]
CO
2
CO
H
2
O
2
CH
2
F
2
CH
4
Arc Current: 7 A
CH
2
FCF
3
: 0.515 L min
-1
CH
2
FCF
3
Fig. 12 Mole fraction of
compositions of effluent gas as a
function of additional O
2
feed
rate in HFC-134a decomposition
system
0
20
40
60
80
100
0
20
40
60
80
100
0.0 0.2 0.4 0.6 0.8 1.0 1.2
Decomposition Efficiency
Recovery of Fluorine
Decomposition Efficiency [%]
Recovery of Fluorine [%]
CF
4
Feed Rate [L min
-1
]
Arc Current: 7 A
Fig. 13 Effect of feed rate on
fluorine recovery and CF
4
decomposition rate
Plasma Chem Plasma Process (2010) 30:813–829 823
123
Author's personal copy
current of 7 A, both the CF
4
decomposition efficiency and the recovery of fluorine were
lower than 70% even at small feeding rate of 0.215 L min
-1
. As the feed rate increases, the
CF
4
decomposition efficiency decreases because high CF
4
feed rate means that less energy
will be available for decomposing each CF
4
molecular. The fact that there was no difference
found between the CF
4
decomposition efficiency and the recovery of fluorine indicates that
there is negligible formation of by-products containing fluorine. The composition of the
effluent gas as a function of the feed rate is shown in Fig. 14. The main compositions were
CO
2
(68.9%), CF
4
(24.9%), H
2
(4.5%), CO (1.1%), and O
2
(0.6%) at feed rate of
0.215 L min
-1
. No by-product was detected by GC-TCD except undecomposed CF
4
.
Compared with the HFC-32 decomposition, the mole concentration of CO-generating was
much smaller than in the case of CF
4
decomposition at the same feed rate supplied, because
the CF
4
decomposition efficiency as well as the carbon generation was very smaller. CO
2
decreases significantly with increasing of CF
4
feed rate. This decrease was due to increase of
the C-containing species owing to the large feed rate in the high temperature region.
The concentration of CO generated by decomposing CF
4
was much lower than that of
the HFC decomposition under the same decomposition conditions. The concentrations of
CO
2
and CO in the effluent gas provide important information of the decomposition
mechanism, because these concentrations are related to the reaction temperature and the
oxidation condition. CO is stable at higher temperatures than CO
2
. The higher CO con-
centration obtained for HFC indicting that the HFC decomposition at higher temperature
and the CO
2
reduction by the H present in HFC.
The decomposition rate as a function of the ratio of feed rate to the arc power is shown
in Fig. 15. In the case of without O
2
, the arc power of 0.98 kWh was remained constant
while the CF
4
feed rate increases. The decomposition efficiency was as low as 67%. Hence,
the experiment of CF
4
decomposition with O
2
was performed in order to improve the
decomposition efficiency. The CF
4
and the additional O
2
feed rate was remained at
0.215 L min
-1
while the arc power was increased. A CF
4
decomposition efficiency of 99%
can be obtained at an electrical energy consumption of 1.89 mol kWh
-1
.
10
-3
10
-2
10
-1
10
0
0.0 0.2 0.4 0.6 0.8 1.0 1.2
Mole Faction [-]
CF
4
Feed Rate [L min
-1
]
CO
2
CO
H
2
Arc Current: 7 A
CF
4
O
2
Fig. 14 Mole fraction of
compositions of effluent gas as a
function of CF
4
feed rate
824 Plasma Chem Plasma Process (2010) 30:813–829
123
Author's personal copy
This result was compared with the work of Sun and Park [21]. They used DC non-
transferred arc torch to decompose CF
4
by argon plasma with additional gas such as H
2
and
O
2
at atmospheric pressure. The almost 100% decomposition efficiency was obtained at the
feed rate of 1.0 L min
-1
CF
4
with 15 L min
-1
Ar, 2 L min
-1
O
2
, and 4 L min
-1
H
2
. Their
results thus obtained that the CF
4
electrical energy consumption was 0.30 mol kWh
-1
.
The PLASCON system has been widely used in industrial waste decomposition process.
For instance, decomposition efficiency of 99.99% was obtained for CCl
2
F
2
at feed rate of
40 L min
-1
at an arc power of 15 kW [10]. In this case, the electrical energy consumption
was 6.5 mol kWh
-1
when the injection mixture gas ratio of CCl
2
F
2
to H
2
O was 1:1. By
comparison, the decomposition efficiency for the PLASCON system is much higher than
that of for the water plasma system because the electrical arc power of PLASCON system
was 15 times larger than that of the water plasma torch. The large arc power provides a
high plasma temperature source which leading to thorough dissociation of the decompo-
sition gas and the additional gas. In addition, the PLASCON system as specialize in
commerce, its design was enhanced energy conservation. Waste enters the torch at a
specially designed injection manifold and instantly mixes with the plasma. Also, in order to
prevent the formation of any undesired organic molecules, the hot gases exiting the flight
tube undergo rapid quenching by direct sprays of cool alkaline liquor. Compared with the
PLASCON system, the water plasma torch is extremely simple and small although its
decomposition efficiency is not ideal. For the application of water plasma torch in industry,
additional efforts should be needed to deal with improving decomposition efficiency by
means of modifying of the apparatus’s design or increasing arc power.
Discussion
Thermal plasma decomposition of fluorinated compounds strongly depends on their
chemical structures at a given plasma temperature. To summarize the experimental results,
0
20
40
60
80
100
0.0 0.5 1.0 1.5
CF
4
Without O
2
CF
4
With O
2
CF
4
Decomposition Efficiency [%]
Feed Rate / Arc Power [mol kWh
-1
]
Fig. 15 CF
4
decomposition
efficiency plotted as a function of
the ratio of CF
4
feed rate to arc
power
Plasma Chem Plasma Process (2010) 30:813–829 825
123
Author's personal copy
higher decomposition efficiency obtained for HFC-134a were attributed to the low bond
energy of C–C bond (348 kJ mol
-1
) in contrast to those of C–H and C–F bonds. The
decomposition mechanism of HFC-134a, HFC-32, and CF
4
were discussed on the basis of
the experimental results and the chemical kinetics.
The distribution of plasma temperatures of the water plasma jet were described in our
previous work [18]. The average temperature of the water plasmas in the region just
downstream of the nozzle exit was measured by emission spectroscopy for free water
plasma jet. The average temperatures were determined from the Boltzmann plot from
hydrogen atoms. As the result, the average temperature was 4,000 K at a position of 6 mm
from the nozzle exit at an arc current of 7 A. Hence, the average decomposition tem-
perature for HFC and PFC were assumed 4,000 K for kinetics consideration.
The decomposition rates were calculated for HFC-134a, HFC-32, and CF
4
and listed in
Table 1. For calculating the reaction rate, we assumed that only the reactions R1, R2, R3,
and R4 run to completion within 6.0 mm of the nozzle exit, respectively. The concen-
trations of the HFC-134a, HFC-32, and CF
4
had been calculated by estimation of that: (1)
HFC-134a, HFC-32, and CF
4
are the ideal gaseous state (2) the feeding water dissociated
100%. The rate constant parameters were taken from the literatures [22–24]. Owing to the
lack of reliable data at high temperature, most of these constants were used for tempera-
tures over the recommended range. Temperature dependent rate parameters were described
by the extended Arrhenius equation as the following:
k ¼ AT
n
expðE=RTÞð4Þ
The calculation results obtained that the reactions R1, R2, and R3 were important steps
for primarily decomposition of HFC-134a, HFC-32, and CF
4
, respectively, because of their
relatively high decomposition rate. Herein, it should be noted that the dissociation of HFC-
134a, HFC-32, or CF
4
molecular by electron attachment was considered negligible because
the decomposition gas was injected into the downstream region of the water plasmas.
As can be seen from Table 1, the decomposition rates of HFC-134a and HFC-32 were
1.57 9 10
5
and 2.56 9 10
3
mol cm
-3
s
-1
, respectively. Hence, the high decompostion
rates lead to complete decomposition of HFC-134a and HFC-32 in high temperature
region. In contrast, the CF
4
decomposition rate was just 5.76 9 10
-2
mol cm
-3
s
-1
. The
CF
4
decomposition needs long decomposition time because of the exceedingly slow
decomposition rate, result in low decomposition efficiency and generation of large amount
of CF
y (y:1–3)
radicals in low temperature range.
The by-product formation mechanism were studied from the comparative study of the
reaction rate constant due to unknown the decomposed species concentration which cannot
Table 1 Decomposition rates for first steps of HFC-134a, HFC-32, and CF
4
decomposition
No. Reactions Forward reaction Decomposition rate Re
AnE
R1 CH
2
FCF
3
? CH
2
F ? CF
3
7.9 9 10
16
0 386 1.57 9 10
5
[22]
R2 CH
2
F
2
? H ? CH
2
F ? HF 1.6 9 10
-11
2 148 2.56 9 10
3
[23]
R3 CF
4
? H ? CF
3
? HF 4.1 9 10
-11
1.6 173 5.76 9 10
-2
[23]
R4 CF
4
? CF
3
? F 3.39 9 10
-1
-4.64 122 6.06 9 10
-17
[24]
Rate coefficient of the forward reaction is k = AT
n
exp(-E/RT), where A is in s
-1
and cm
3
molecule
-1
s
-1
for 1 and 2 reactant species, respectively, the activation energy E is in kJ mol
-1
, R is the ideal gas constant,
and T is the temperature. Unit for decomposition rate is mol cm
-3
s
-1
826 Plasma Chem Plasma Process (2010) 30:813–829
123
Author's personal copy
be measured experimentally. The reactions which have high possibilities in the thermal
plasmas are described in Table 2 involving 20 species and 17 reaction steps [10, 24–37].
The rate reaction constants were calculated at a temperature of 2,000 K which is estimated
from thermodynamic calculation.
The experimental results of the CF
4
decomposition were showed that CHF
3
,CH
2
F
2
, and
CH
3
F had not been generated during the CF
4
decomposition, indicating CF
3
,CF
2
, and CF
radicals were not combined with H radical. For CF
4
,CF
3
,CF
2
, and CF, the reaction rate
constants of their reduction and oxidation (R2–R7) were extremely larger than that of their
recombination with H, O or each other. In addition, the produced hot gaseous were quickly
quenched during the experiment. Hence, all the reactions of their recombination with H, O
or each other were not listed Table 2. From the perspective of activation energy, reaction
(5) is initiated through an electrophonic addition step of H atom to trifluoromethyl radical.
Since the addition step is greatly exothermic, chemically activated energized trifluorom-
ethane (CHF
3
*) is formed. Subsequently, CHF
3
* is decomposed through HF elimination
(6). HF elimination (7) is superior to the F dissociation (8) for the activation energy for
reaction (7) is much lower than that of reaction (5).
CF
3
þ H ! CHF
3
ð5Þ
CHF
3
! CF
2
þ HF ð6Þ
CF
3
þ H ! CF
2
þ HF ð7Þ
CF
3
þ H ! CHF
2
þ F ð8Þ
For CF
2
removal, the reaction of (9) is less important because of its relatively small
reaction rate constant in comparison with reactions of (10) and (11).
Table 2 Kinetics parameters for fluorinated compounds decomposition by water plasmas
No. Reactions Forward reaction Re.
AnE(kJ)
R1 CH
2
F ? F ? CHF ? HF 8.3 9 10
-11
00[25]
R2 CF
3
? O ? COF
2
? F 2.6 9 10
-11
00[26]
R3 CF ? O ? CO ? F 3.9 9 10
-11
00[27]
R4 CF
2
? O ? CO ? 2F 3.9 9 10
-11
00[10]
R5 CF
3
? H ? CF
2
? HF 8.9 9 10
-11
00[26]
R6 CF
2
? H ? CF ? HF 6.6 9 10
-11
00[28]
R7 CF ? H ? CH ? F 6.1 9 10
-11
079[29]
R8 CF
2
? F ? M ? CF
3
? M 1.8 9 10
-24
-910[24]
R9 CF
3
? F ? M ? CF
4
? M 8.9 9 10
-28
-512[24]
R10 CF
2
? H
2
? CH
2
F
2
4.7 9 10
-12
0 142 [30]
R11 CF
3
? H
2
? CHF
3
? H 2.6 9 10
-11
071[31]
R12 CF
2
? O
2
? COF
2
? O 6.6 9 10
-13
072[32]
R13 COF
2
? H ? COF ? HF 7.5 9 10
-11
096[33]
R14 COF
2
? O ? CO
2
? F
2
2.2 9 10
-11
00[34]
R15 COF ? O ? CO
2
? F 1.0 9 10
-10
00[35]
R16 COF ? H ? CO ? HF 5.7 9 10
-11
0.5 0 [36]
R17 COF ? F ? CO ? F
2
2.9 9 10
-10
0.5 137 [37]
Plasma Chem Plasma Process (2010) 30:813–829 827
123
Author's personal copy
CF
2
þ O
2
! COF
2
þ O k
1
¼ 6:6 10
13
molecule
1
cm
3
s
1
ð9Þ
CF
2
þ H
2
! CF þ HF k
2
¼ 1:6 10
12
molecule
1
cm
3
s
1
ð10Þ
CF
2
þ O
2
! CO þ2F k
3
¼ 3:8 10
11
molecule
1
cm
3
s
1
ð11Þ
Therefore, the chemically stable hydroflourocarbons and carbonyl fluoride products
were not detected from the CF
4
decomposition process. The low CF
4
decomposition
efficiency was considered due to the recombination of CF
3
with F (12).
CF
3
þ F ! CF
4
k
4
¼ 2:0 10
12
molecule
1
cm
3
s
1
ð12Þ
A relatively-low concentration of CH
2
F
2
, CHF
3
, and CF
4
was generated during
decomposition of HFC-134a, as shown in Fig. 10. It is proposed that they were formed by
the recombination of CH
2
F and CHF radicals with fluorine (13)–(15).
CH
2
F þ F ! CH
2
F
2
ð13Þ
CHF þ F ! CHF
2
ð14Þ
CHF
2
þ F ! CHF
3
ð15Þ
Comparing the results of QMS measurements show in Fig. 10a, b, it can be seen that
CHF
3
and CF
4
were generated only at low arc current. The amount of CH
2
F
2
generated
was much larger than that of CF
4
for the decomposition of HFC-134a by the water
plasmas. This means that the CH
2
F radical was much more stable and much more dom-
inant than CF
3
radical in the water plasmas. It can be explained by electron affinity. Carbon
atom is in the electron delocalization state in CF
3
radical compared with the carbon atom
of in CH
2
F radical, when three F atom which is the most electronegative atom pulls on the
bonding electrons. As a result, CH
x
F easily combines with F to generate by-product.
However, H will combine with F atom to improve the decomposition of CH
x
F and CF
y
.
Conclusions
A feasibility of decomposition process for fluorinated compounds such as HFC-134a,
HFC-32, and CF
4
were demonstrated by water plasmas generated by DC discharge. The
decomposition efficiency and the reduction unwanted by-product were improved by adding
oxygen as oxidation gas. The compositions in the effluent gas were mainly H
2
,CO
2
, CO,
and CH
4
in which the mole fraction of syngas of H
2
and CO was between that of 25–70 for
HFC decomposition and between that of 10–20 for PFC decomposition. By-products of
CH
2
F
2
, CHF
3
, and CF
4
were detected in trace amounts in HFC decomposition process;
however there is no by-product formed in PFC decomposition. The by-product formation
strongly depends on the availability of F atom to combine with the other decomposed
species. The decomposition mechanism as well as the by-product formation mechanism
was also studied.
References
1. Takizawa K, Takahashi A, Tokuhashi K, Kondo S, Sekiya A (2005) Combust Flame 141:298–307
2. Takizawa K, Takahashi A, Tokuhashi K, Kondo S, Sekiya A (2006) J Fluorine Chem 127:1547–1553
828 Plasma Chem Plasma Process (2010) 30:813–829
123
Author's personal copy
3. Takita Y, Tanabe T, Ito M, Ogura M, Muraya T, Yasuda S, Nishiguchi H, Ishihara T (2002) Ind Eng
Chem Res 41:2585–2590
4. Hannus I (1999) Appl Catal A 189:263–276
5. Jasinski M, Dors M, Mizeraczyk J, Lubanski M, Zakrzewski Z (1999) J Tech Phys 40:197–200
6. Jasinski M, Szczucki P, Dors M, Mizeraczyk J, Lubanski M, Zakrzewski Z (2000) Czech J Phys
50:285–288
7. Futamura S, Gurusamy A (2005) J Electrostat 63:949–954
8. Murphy AB, McAllister T (1998) Appl Phys Lett 73:459–461
9. Heberlein J, Murphy AB (2008) J Phys D Appl Phys 41:053001
10. Murphy AB, Farmer AJD, Horrigan EC, McAllister T (2002) Plasma Chem Plasma Process 22:371–385
11. Ohno M, Ozawa Y, Ono T (2007) Int J Plasma Environ Sci Technol 1:159–165
12. Kim SW, Park HS, Kim HJ (2003) Vacuum 70:59–66
13. Glocker B, Nentwig G, Messerschmid E (2000) Vacuum 59:35–46
14. Hrabovsky M, Konrad M, Kopecky V, Sember V (1997) IEEE Trans Plasma Sci 25:833–839
15. Hrabovsky M, Kopecky V, Sember V, Kavka T, Chumak O, Konrad M (2006) IEEE Trans Plasma Sci
34:1566–1575
16. Watanabe T (2005) ASEAN J Chem Eng 5:30–34
17. Nishioka H, Saito H, Watanabe T (2009) Thin Solid Films 518:924–928
18. Watanabe T, Tsuru T (2008) Thin Solid Films 516:4391–4396
19. Narengerile, Saito H, Watanabe T (2009) Thin Solid Films 518:929–935
20. http://webbook.nist.gov
21. Sun JW, Park DW (2003) Korean J Chem Eng 20:476–481
22. Millward GE, Tschuikow-Roux E (1972) J Phys Chem 76:292–298
23. Berry RJ, Ehlers CJ, Burgess DR Jr, Zachariah MR, Marshall P (1997) Chem Phys Lett 269:107–116
24. Modica AP, Sillers SJ (1968) J Chem Phys 48:3283–3289
25. Beiderhase Th, Hack W, Hoyermann K (1995) Z Phys Chem 188:227–241
26. Takahashi K, Sekiuji Y, Yamamori Y, Inomata T, Yokoyama K (1998) J Phys Chem A 102:8339–8348
27. Van Hoeymissen J, De Boelpaep I, Uten W, Peeters J (1994) J Phys Chem 98:3725–3731
28. Yamamori Y, Takahashi K, Inomata T (1999) J Phys Chem A 103:8803–8811
29. Garrett BC, Truhlar DG (1979) J Am Chem Soc 101:5207–5217
30. Yu H, Kennedy EM, Uddin A, Sullivan SP, Dlugogorski BZ (2005) Environ Sci Technol 39:3020–3028
31. Hranisavljevic J, Michael JV (1998) J Phys Chem A 102:7668–7673
32. Bauer SH, Hou KC, Resler EL Jr (1969) Phys Fluids 12:125–132
33. Richter H, Vandooren J, Van Tiggenlen PJ (1994) J Chim Phys 91:1748–1762
34. Atkinson R, Baulch DL, Cox RA, Hampson RF Jr, Kerr JA, Rossi MJ, Troe J (1997) J Phys Chem Ref
Data 26:521–1011
35. Herron JT (1988) J Phys Chem Ref Data 17:967–1026
36. Gangloff HJ, Milks D, Maloney KL, Adams TN, Matula RA (1975) J Chem Phys 63:4915–4926
37. Modica AP (1970) J Phys Chem 74:1194–1204
Plasma Chem Plasma Process (2010) 30:813–829 829
123
Author's personal copy
