A carbon arc process for treatment of CF4 emissions

Article abstract of DOI:10.1021/es980533w

Light perfluorocarbons, such as carbon tetrafluoride, are produced or emitted from a variety of processes, including manufacture of aluminum and processing of semiconductor devices. At the same time, the long atmospheric lifetime and high global warming potential of such compounds makes them an environmental concern. A new process for the abatement of perfluorocarbon emissions using a carbon are plasma was investigated. In particular, the conversion of CF₄ to C₂F₄ and higher fluorinated species, including poly(tetrafluoroethylene) (PTFE) was demonstrated. General features of the reaction chemistry are discussed, including primary reactions to form radicals and ions and secondary reactions to form C₂F₄ and higher compounds. The conversion efficiencies and products obtained in the reported experiments indicate potential applicability of the process for point source emission control of high global warming potential perfluorocarbons.

Full text of DOI:10.1021/es980533w

Research Communications
A Carbon Arc Process for Treatment
of CF Emissions
4
D A N I E L T . C H E N , * , † M O S E S M . D A V I D , ‡
‡
G E O R G E V . D . T I E R S , A N D
‡
J O S E P H N . S C H R O E P F E R
3
M Occupational Health and Environmental
Safety Division, 3M Center, Building 260-3B-08,
St. Paul, Minnesota 55144-1000, and 3M Corporate Research
Laboratories, 3M Center, St. Paul, Minnesota 55144-1000
Light perfluorocarbons, such as carbon tetrafluoride, are
produced or emitted from a variety of processes, including
manufacture of aluminum and processing of semiconductor
devices. At the same time, the long atmospheric lifetime
and high global warming potential of such compounds makes
them an environmental concern. A new process for the
abatement of perfluorocarbon emissions using a carbon arc
plasma was investigated. In particular, the conversion
of CF4 to C2F4 and higher fluorinated species, including poly-
FIGURE 1. Experimental apparatus (not to scale).
This paper reports on a new alternative process for
mitigating fluorocarbon emissions, namely, the conversion
of light perfluorocarbons into fluorinated materials with
recoverable value. The preliminary work described in this
paper demonstrates the feasibility of a plasma arc process
(tetrafluoroethylene) (PTFE) was demonstrated. General
features of the reaction chemistry are discussed, including
primary reactions to form radicals and ions and secondary
reactions to form C2F4 and higher compounds. The
conversion efficiencies and products obtained in the
reported experiments indicate potential applicability of the
process for point source emission control of high global
warming potential perfluorocarbons.
for the conversion of CF
PTFE), and other fluorocarbons.
Plasma processes are a class of Advanced Oxidation
4 2 4
to C F , poly(tetrafluoroethylene)
(
Technologies (AOTs) gaining recognition for treatment of
gaseous and liquid waste. One type of thermal plasma is the
carbon arc; in this case, the arc is sustained between graphite
electrodes placed in a vacuum and driven by a high current
DC welding power supply. Extremely high current densities
6
7
2
(
10 -10 A/ cm ) are characteristic of the carbon arc process,
leading to evaporation and subsequent ionization of the
graphitic carbon; this occurs at the arc spots where the local
temperatures can exceed 10 000 K. The resulting gas phase
species are highly energetic with local particle velocities
Introduction
Perfluorinated compounds are a class of materials useful for
a variety of applications, ranging from semiconductor wafer
etching and precision parts cleaning to heat transfer. An
often exploited feature of these materials, their chemical
stability, also has an environmental consequenceslong
atmospheric lifetime. As such, perfluorocarbons represent
an environmental burden as potent greenhouse gases. The
most potentially damaging of these gases are the light
4
exceeding 10 m/ s (13); such highly energetic particles are
conducive for subsequent gas phase reaction chemistry.
While the temperatures within the arc spot are very high, the
bulk gas phase temperatures are relatively low; importantly,
the plasma gas temperature quenches over a much shorter
distance than that corresponding to a bulk gas heating
process, such as incineration.
4 2 6 4
perfluorocarbons, CF and C F ; for example, CF has been
Homogeneous gas phase reactions between energetic
molecules, electrons, ions, and radicals as well as hetero-
geneous reactions with carbon electrodes at elevated tem-
perature lead to the efficient conversion of even stable
estimated to have an atmospheric lifetime of between 2300
and 50 000 years (1-3). Although the total contribution of
perfluorocarbons to anthropogenic sources of global warming
may be small, strategies for their control should be developed.
4
molecules such as CF . Although there are discrete arc spots
4 2 6
The major emissions of CF and C F are from aluminum
on the electrodes, the reaction zone is considerably larger,
as these arc spots move over the surface of the electrodes.
The process can be operated at atmospheric pressure, though
the work described in this report was conducted below 10
Torr. Important practical advantages of the carbon arc
process include relatively low capital cost, high con-
version efficiencies, low bulk thermal temperature (com-
pared to incineration), and a wide flexibility in reaction
chemistries.
production, fluorochemical manufacturing, and semicon-
ductor device processing. Several strategies are currently
being investigated for mitigation of perfluorocarbon releases,
including source reduction (4-7), pressure swing adsorption
(
(
8), microwave plasma destruction (9), thermal destruction
10), cryogenic recovery, and membrane permeation (11).
Several of these techniques have recently been assessed and
reviewed by Mocella et al. (12).
Interestingly, the conversion of fluorocarbons at high
temperature or in plasma was investigated many years ago,
primarily directed toward the synthesis of PTFE. In the late
1950s to the mid 1960s, workers at du Pont (14, 15), Dow
*
To whom correspondence should be addressed. Phone: (651)
7
33-3122; fax: (651) 737-3069; e-mail: dtchen@mmm.com.
†
‡
3
3
M Occupational Health and Environmental Safety Division.
M Corporate Research Laboratories.
S0013-936X(98)00533-1 CCC: $15.00
Published on Web 08/29/1998

1998 Am erican Chem ical Society
VOL. 32, NO. 20, 1998 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3 2 3 7

FIGURE 2. Example GC analysis of gas phase reactor effluent.
FIGURE 3. Single ion traces for specific fluorocarbon fragments from mass spectrometry. (a) Ion m/z ) 169.00 (168.70-169.70) corresponding
+
+
+
3
to C
3
F
7
. (b) Ion m/z ) 131.00 (130.0-131.70) corresponding to C F . (c) Ion m/z ) 69.00 (68.70-69.70) corresponding to CF .
3 5
(
16), and MIT (17) demonstrated the feasibility of forming
into a wafer processing chamber (Plasma-Therm Systems
Inc.); the chamber dimensions were 66 cm in diameter and
18 cm in height. The reactant gases were fed through the
center of the graphite cone via 1/ 4-in. copper tubing.
Procedure. The anode and cathode were configured, and
the chamber cover was lowered. The chamber was first
PTFE in plasma processes; more recently, Swanepoel and
Lombaard have filed a patent on a similar process (18). The
principal motivation for this earlier work was to develop an
economical method of producing tetrafluoroethylene, the
monomer for PTFE, rather than any environmental consid-
erations. This work forms the basis for more detailed studies
of the feasibility of plasma processes for the mitigation of
fluorocarbon emissions.
evacuated to ∼5 mTorr; the flow of CF was then initiated
4
and allowed to stabilize. A variety of flow rates were
3
investigated ranging from 200 to 800 standard cm / min. To
initiate the arc, the chamber pressure was initially 1 Torr or
below. After striking the arc, a steady-state current flow was
rapidly achieved, with a typical current flow of 60 amp at a
potential of 20-30 V as measured at the power supply. Arc
stability was highly dependent on the gas-phase pressure
and the reactor configuration, as discussed below. After
achieving a steady-state plasma, the gas-phase pressure was
raised to ca. 5 Torr by reducing the vacuum pumping speed.
Steady-state operation for 10-15 min was readily achievable.
The product analyses of this report were obtained at an arc
Experimental Section
Equipm ent. The equipment used in this series of experi-
ments is shown in Figure 1. The power supply was a Miller
Maxstar 151. A graphite anode and cathode were milled to
form a matching cone and cup configuration. In the
configuration used, gap width was small and controlled;
consumption of the electrodes did not strongly influence
the gap space as there was a large effective area at the chosen
gap width. The cone measured 10 cm in height with a basal
diameter of 5 cm. The anode-cathode assembly was fixed
current of 62 amp at 23 V, with a CF
4
flow of 800 standard
3
cm / min at a pressure of 5.6 Torr.
3
2 3 8
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 32, NO. 20, 1998

Although an exhaustive discussion of the reaction chem-
istry is outside the scope of this paper, some general features
are worthy of discussion. If the carbon arc process is to be
used for light perfluorocarbon conversion, a potential target
reaction is the conversion of tetrafluoromethane to tet-
rafluoroethylene followed by polymerization to PTFE:
CF (g) + C(g) f C F (g)
4
2 4
nC F (g) f (C F ) (s)
2
4
2 4 n
The reactions occurring on the graphite surface and those
in the bulk phase will likely be very different. In the high
temperature arc region, difficult (e.g., breakdown of stable
bonds) reactions can occur; the resulting radicals and ions
recombine under cold bulk phase conditions, promoting
condensation and polymerization. It should also be noted
that the principal arc reactions are primarily taking place at
FIGURE 4. XRD pattern of solid residue showing presence of PTFE.
Analysis. Gas phase samples were analyzed by gas
chrom atography/ m ass spectrom etry (GC/ MS) using a
HP5890/ 5972 system. Separation was achieved using a 30-m
DB-5MS column (0.25 mm i.d., 1.0 µm film) column with a
temperature ramp between -40 (2 min hold) and 230 °C (10
the cathode spots. The reactions taking place with CF as
4
a feed and solid graphite may include reactions such as those
described below:
primary reactions: formation of radicals and ions
and CF units
°
C/ min). The separated components were subjected to
2
electron bombardment at 70 eV. Full-scan unit resolution
mass spectra from 29 to 500 m/ z were recorded. Additional
C(s) f C(g)
(vaporization of carbon)
gas chromatography experiments were used to confirm CF
4
4
C(g) + CF f 2CF2
(gas phase carbon reacting
with CF4)
conversion using an HP5880A chromatograph equipped with
a thermal conductivity detector. Solid-phase product was
sampled and analyzed by X-ray diffraction (Philips, copper
KR radiation) and FTIR (Bomem MB102).
C(s) + CF f 2CF2
(solid carbon reacting with CF4)
4
-
•
-
Results and Discussion
CF + e f CF + F
(electron impact
4
3
Gas phase pressure and gap spacing influenced arc stability,
with a larger dielectric constant (related to pressure) and
larger gap width being detrimental. During operation, the
generation of arc spots that moved over the surface of the
electrodes could be clearly seen. The preliminary nature of
this work precluded optimization; it is evident that more
efficient reactor configurations could be envisioned.
dissociation and ionization)
•
-
-
CF + e f CF + F
3
2
-
•
-
CF + e f CF + F
2
secondary reactions: formation of C F and
2
4
Gas phase samples were analyzed by GC/ MS. A typical
result from gas chromatography is shown in Figure 2, which
shows the presence of several compounds in addition to the
starting material. Quantitative analysis by gas chromatog-
higher compounds (examples)
2CF f C F
2
2
4
raphy showed an average conversion of 23% of CF
4
, with
•
•
CF + CF3 f C2F4
2 4
C F
being the primary product (21.6%). The space velocity
-
1
under standard conditions is 40 min ; at the reactor
conditions (pressure of 5.6 Torr), the effective space velocity
CF + C F f C F
2
2
4
3
6
8
-
1
is approximately 5430 min
The presence of high molecular weight materials is
consistent with CF being broken down in the presence of
carbon to produce CF
.
C F + CF f C F
3
6
2
4
4
In practice, a variety of reactions in addition to those above
take place in the carbon arc process, most of them dominated
by free radical chemistry. Such reactions include the stepwise
2
fragments. Figure 3 shows the single
+
ion traces for specific fragments corresponding to C
3
F
7
(m/ z
+
+
)
169, Figure 3a), C
m/ z ) 69, Figure 3c); various saturated and unsaturated
3
F
5
(m/ z ) 131, Figure 3b), and CF
3
4 2 2
dissociation of CF into CF . The literature suggests that CF
(
fragments are the fundamental building blocks for the
formation of oligomeric fluorocarbons under high temper-
ature conditions (19, 20).
species in the gas phase products are identified. Interestingly,
species up to perfluorodecane can be clearly identified.
Higher unsaturated compounds are also present, although
less readily identifiable. Although quantitative results are
difficult to extract, it is clear that species concentration
diminishes with higher molecular weight, consistent with
The carbon arc process offers wide flexibility in the
conversion of fluorochemicals to various products depending
on the co-reactants fed into the system. Many chemistries
that will be active have not yet been examined, such as the
addition of hydrocarbon co-reactants. The high reactivity
of gas phase hydrocarbons relative to graphite may allow
lower plasma energies to be used and higher conversions to
be achieved. One can certainly envision a slate of partially
fluorinated hydrocarbons (HFCs) resulting from a carbon
arc or other plasma process.
2
the idea that CF units are building chain length.
Analysis of the solid white material deposited on the walls
of the chamber shows the presence of PTFE, as confirmed
by FTIR and XRD (Figure 4). The PTFE produced from these
experiments was in powder form and loosely attached to the
reactor walls, indicating that some polymerization may be
taking place in the bulk gas phase in addition to polymer-
ization at the reactor walls.
The principal challenge in employing a carbon arc or other
thermal plasma process for control of perfluorocarbons is to
VOL. 32, NO. 20, 1998 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3 2 3 9

increase the selectivity to the desired reactions. It is likely
that continued investigation into the quenching of activated
species will yield advances toward this goal.
(9) Hartz, C. L.; Bevan, J. W.; Jackson, M. W.; Wofford, B. A. Environ.
Sci. Technol. 1998, 32, 682-687.
(10) Mawle, P. Eur. Semicond. 1995, 5, 17-19.
(
11) Chernyakov, I.; Hsiung, T.; Schwarz, A.; Yang, J. Air Products,
U.S. Patent 5,730,779, March 24, 1998.
Acknowledgments
(
(
12) Mocella, M. T. Mater. Res. Soc. Symp. Proc. 1997, 447, 29-34.
The authors thank 3M Corporate Research and 3M Specialty
Materials Division for supporting this work. Important
analytical support was provided by Brian Lynch, Gerry
Lillquist, and Bill Peterson. In addition, the insights of Allen
Siedle, Fred Behr, Gerald Bauer, David Kracht, and Lou Errede
are gratefully acknowledged.
13) Johnson, P. C. The Cathodic Arc Plasma Deposited Thin Films.
In Thin Film Processes II; Vossen, J. L., Kern, W., Eds.; Academic
Press: Boston, MA, 1991; pp 209-280.
(
14) Denison, J. T.; Edlin, F. E.; Whipple, G. H., E. I. du Pont
deNemours and Company, U.S. Patent 2,852,574, September
1
6, 1958.
(
(
(
15) Farlow, M. W., E. I. duPont de Nemours and Company, U.S.
Patent 3,081,245 March, 12, 1963.
Literature Cited
(
1) Ravishankara, A. R.; Solomon, S.; Turnipseed, A. A.; Warren, R.
F. Science 1993, 259, 194-199.
16) Von Tress, W. R. The Dow Chemical Company, U.S. Patent
3
,133,871, May 19, 1964.
17) Baddour, R.; Bronfin, B. Ind. Chem. Process Des. Dev. 1965, 4,
62.
(
(
2) Cicerone, R. J. Science 1979, 206, 59-61.
3) Zander, R.; Solomon, S.; Mahieu, E.; Goldman, A.; Rinsland, C.
P.; Gunson, M. R.; Abrams, M. C.; Chang, A. Y.; Salawitch, R. J.
Belg. Geophys. Res. Lett, 1996, 23, 2353-2356.
1
(18) Swanepoel, J.; Lombaard, R. Atomic Energy Corporation of South
Africa Limited, European Patent EP648530A1, October 14, 1994.
(
4) Tabereaux, A. T.; Richards, N. E.; Satchel, C. E. Light Met.
(
Warrendale, PA) 1995, 325-333.
(
(
19) Errede, L. J. Org. Chem. 1962, 27, 3425-3430.
20) Weigert, F. J. J. Fluorine Chem. 1993, 65, 67-71.
(
(
5) Williams, J. D. Mater. Res. Soc. Symp. Proc. 1997, 447, 43-48.
6) Streif, T.; DePinto, G.; Dunnigan, S.; Atherton, A. Semicond. Int.
1
997, 20, 129-134.
7) Langan, J.; Maroulis, P.; Ridgeway, R. Solid State Technol. 1996,
9, 115-122.
8) Tom, G. M.; McManus, J.; Knolle, W.; Stoll, I. Mater. Res. Soc.
Symp. Proc. 1994, 344, 267-272.
(
(
Received for review May 22, 1998. Revised manuscript re-
ceived July 30, 1998. Accepted August 3, 1998.
3
ES980533W
3
2 4 0
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 32, NO. 20, 1998