Electrodegradation of Chlorofluorocarbons in a Laboratory-Scale Flow Cell with a Hydrogen Diffusion Anode

Article abstract of DOI:10.1149/1.1639164

Trichlorofluoromethane (CFC 11) and 1,1,2-trichloro-1,2,2-trifluoroethane (CFC 113) have been cathodically reduced in a laboratory-scale flow cell at room temperature. The electrodes, 10 cm² in section, were a Cu cathode and a H₂-fed gas diffusion electrode. The cell reservoir contained 0.5 dm³ of methanol-water mixture with NH₄Cl, perchlorate of tetrabutylammonium (PTBA), or both. In the case of CFC 113, 50 ppm PdCl₂ was also added to the electrolyte. Electroreduction took place with the formation of dechlorinated derivatives by steps, methane and difluoroethene being the last derivatives for CFCs 11 and 113, respectively. The current efficiency in NH₄Cl was about 75%. It was improved for CFC 11 to 90-95% when PTBA was employed, both with NH₄Cl or alone. The use of PTBA for CFC 113 was unsatisfactory, because it interfered with Pd electrodeposition and the current efficiency decreased. However, copper corrosion during the electroreduction of both CFCs significantly decreased with the use of PTBA. The hydrogen diffusion anode appeared to be chemically stable during the process, although some concave deformation after prolonged hydrogen oxidation leading to an increase in the interelectrode gap was apparent. The improvement of its dimensional stability is discussed.

Full text of DOI:10.1149/1.1639164

B98
Journal of The Electrochemical Society, 151 ͑2͒ B98-B104 ͑2004͒
0013-4651/2004/151͑2͒/B98/7/$7.00 © The Electrochemical Society, Inc.
Electrodegradation of Chlorofluorocarbons in a
Laboratory-Scale Flow Cell with a Hydrogen Diffusion Anode
,z
Pere L. Cabot,a, Laura Segarra,a and Juan Casadob
*
a
´
´
LCTEM, Departament de Quımica Fısica, Universitat de Barcelona, 08028 Barcelona, Spain
b
´
DIC, Carburos Metalicos, S. A., 08005 Barcelona, Spain
Trichlorofluoromethane ͑CFC 11͒ and 1,1,2-trichloro-1,2,2-trifluoroethane ͑CFC 113͒ have been cathodically reduced in a
laboratory-scale flow cell at room temperature. The electrodes, 10 cm2 in section, were a Cu cathode and a H2-fed gas diffusion
electrode. The cell reservoir contained 0.5 dm3 of methanol-water mixture with NH4Cl, perchlorate of tetrabutylammonium
͑PTBA͒, or both. In the case of CFC 113, 50 ppm PdCl2 was also added to the electrolyte. Electroreduction took place with the
formation of dechlorinated derivatives by steps, methane and difluoroethene being the last derivatives for CFCs 11 and 113,
respectively. The current efficiency in NH4Cl was about 75%. It was improved for CFC 11 to 90-95% when PTBA was employed,
both with NH4Cl or alone. The use of PTBA for CFC 113 was unsatisfactory, because it interfered with Pd electrodeposition and
the current efficiency decreased. However, copper corrosion during the electroreduction of both CFCs significantly decreased with
the use of PTBA. The hydrogen diffusion anode appeared to be chemically stable during the process, although some concave
deformation after prolonged hydrogen oxidation leading to an increase in the interelectrode gap was apparent. The improvement
of its dimensional stability is discussed.
© 2004 The Electrochemical Society. ͓DOI: 10.1149/1.1639164͔ All rights reserved.
Manuscript submitted January 8, 2003; revised manuscript received September 5, 2003. Available electronically January 9, 2004.
The use of chlorofluorocarbons ͑CFCs͒ for industrial purposes
since around 1930 has lead to emissions and further accumulation of
these compounds in the atmosphere, destroying stratospheric ozone
and contributing to the greenhouse effect in the troposphere. Al-
though the production of CFCs was stopped since 1996 as a result of
international agreements, there are still large amounts in stock. The
estimated amount of banked CFCs are 2.25 Mton, about 45% being
trichlorofluoromethane ͑CFC 11͒ and about 45% dichlorodifluo-
romethane ͑CFC 12͒.1
trolyte, whereas H2 is fed from the opposite side. Hydrogen ab-
sorbed in Pd is then transported to the electrolyte side, where it is
electrochemically oxidized to proton. This electrode made the use of
separators unnecessary and it avoided the formation of polluting
by-products. Different dechlorinated compounds including com-
pletely dechlorinated ones were obtained with electroreduction effi-
ciencies up to 98% at high current densities. However, significant Pb
corrosion and current efficiencies of 25-50% were found in sodium
acetate and current densities of about 1 mA cmϪ2 11
thus justifying
,
In order to avoid eventual emissions, different destruction pro-
cesses such as incineration, pyrolysis, hydrolysis, catalytic oxida-
tion, catalytic hydrogenation, plasma destruction, UV destruction,
and destruction using high-energy radiation, have been proposed.
However, destruction methods require energy and waste removal,
and therefore, conversion of CFCs into useful products appears to be
more attractive, since it combines synthesis with degradation. This
can be done using chemical or electrochemical reactions. As an
example of the former, Pd black has been applied as a model cata-
lyst for the high-temperature hydrogenolysis of CFC 12 into CH2F2
͑HFC 32͒, which appears to be a new technology refrigerant.1
Electroreduction of CFCs appears to be adequate for such a com-
bination of electrodegradation and electrosynthesis, because chlo-
further effort to characterize corrosion phenomena and current effi-
ciency in different electrode-electrolyte systems.
In this work, a new small-scale flow cell with a Pd-based
H2-diffusion anode and a copper cathode in parallel has been built
and tested. Methanol ͑70 vol %͒-water mixtures containing NH4Cl,
PTBA, or both have been employed. The selection of cathode and
electrolytes was based on the efficient electroreduction of CFC 113
to CTFE on Cu cathodes in NH4Cl-containing methanol-water
mixtures2 and on the catalytic effect of PTBA in the electroreduction
of CFC 113 in wetproofed carbon-black porous cathodes.18 The con-
version of CFCs 11 and 113, the current efficiencies, and the stabil-
ity of the electrodes at current densities of about 10 mA cmϪ2 have
been determined using different analytical techniques.
ride is easily removed from the molecule in this way. Diverse com-
Experimental
2-7
pounds such as chlorotrifluoroethylene ͑CTFE͒
and
tetrafluoroethylene4,8 can be produced from 1,1,2-trichloro-1,2,2-
trifluoroethane ͑CFC 113͒ and from 1,2-dichloro-1,1,2,2-
tetrafluoroethane ͑CFC 114͒, respectively, using a variety of elec-
trodes and electrolytes. More recently, difluoroethene ͑DFE͒ and
trifluoroethene ͑TFE͒ have also been obtained from CFC 113 on Pb
cathodes in methanol-water mixtures containing different electro-
lytes and PdCl2 dissolved at the parts per million level,9-11 which
catalyzed the formation of the most dechlorinated derivatives. It has
also been shown that HFC 32 and/or tetrafluoroethene are formed in
the electrochemical reduction of CFC 12 in aqueous12,13 or organic
media14-16 using lead,15,16 silver,12,14 or Pb-supported gas diffusion
electrodes ͑GDEs͒.14
The authors have employed a conventional laboratory closed cell
to dechlorinate CFC 11 ͑trichlorofluoromethane͒ and CFC 113 in
methanol-water mixtures containing NH4Cl.9,10,17 The electrodes
were a Pb cathode and a Pd-based hydrogen diffusion anode. In the
hydrogen diffusion anode, only one side is in contact with the elec-
The cathodes were 2 mm thick Cu disks having 4.6 cm diam.
The Cu composition, in weight %, was 0.013 Pb, 0.005 Sn, 0.002
Ni, and 0.031 P, Al, Sb, As, Bi, and Fe, being less than 0.001%.
Before the experiments it was cleaned with 50 vol % HNO3 .
The anode was built using a 150 ␮m thick Pd-30 Ag foil from
Goodfellow, activated with a layer of Pd black ͑2 mg cmϪ2͒ and an
external layer of Pt black ͑2 mg cmϪ2͒. The first layer was electro-
chemically generated, as indicated previously, but reducing the
deposition time to the half.9 Afterward, Pt black was electrodepos-
ited on it from a solution containing 3% H2PtCl6 • 6H2O and
0.02% Pb͑AcO͒ • 3H2O ͑both Merck for synthesis͒, at Ϫ15 mA
2
cmϪ2 and 50°C for 8 min, as indicated elsewhere.19 It was fed with
commercially pure H2 from Air Products and Chemicals at atmo-
spheric pressure.
The electrodes were coupled to a two-electrode undivided flow
cell made of stainless steel ͑SS͒ and polymethylmethacrylate
͑PMMA͒. Figure 1 shows a side view, with two SS bodies, for the
anode ͑SA͒ and the cathode ͑SC͒. The latter was solid whereas the
former was perforated to allow the H2 arrival at the anode. The
anode gas chamber was closed with a PMMA window ͑PW͒, thus
allowing observation of the foil, and had a H2 inlet and a purge.
* Electrochemical Society Active Member.
z E-mail: p.cabot@ub.edu
Downloaded on 2015-04-06 to IP 169.233.248.124 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 ͑2͒ B98-B104 ͑2004͒
B99
minϪ1 by means of a peristaltic pump ͑P͒, and then it was returned
to the reservoir. The electrolyte was mildly stirred during the elec-
trolyses by means of a magnetic stirrer ͑S͒ for homogenizing its
composition and temperature. The system also had a key for extract-
ing liquid sample ͑E͒. The extraction of gas samples was performed
by means of a syringe through the septum on the top of the reservoir.
The working electrolytes were prepared in methanol ͑70 vol %͒-
water mixtures. They contained 0.75 mol dmϪ3 NH4Cl, 0.1 mol
dmϪ3 PTBA, or 0.75 mol dmϪ3 NH4Cl ϩ 0.05 mol dmϪ3 PTBA.
The water was Millipore of Milli-Q quality. Methanol and NH4Cl
were Panreac PRS, and PTBA, Fluka, electrochemical grade.
Amounts of CFC 11 and CFC 113 of about 0.1 mol were added to
the electrolyte in different experiments. They corresponded to about
the half the saturation concentration of CFC 11,17 and near satura-
tion concentration of CFC 113,9 at 20°C. For CFC 113 electrodeg-
radation, PdCl2 ͑Merck for synthesis͒ was also used at a concentra-
tion of 50 ppm. It was added to the electrolyte as a small volume of
concentrated PdCl2 solution acidified with HCl ͑reagent grade
Merck͒ to pH 0. These solutions were not deaerated prior to the
experiments in order to avoid the loss of volatile compounds and to
study the behavior of the system in the presence of dissolved O2 ,
which can be useful in the case of a higher scale application.
The electrolyses were performed controlling the cell voltage in
the range 1-5 V to obtain quasi-steady currents of about 100 mA,
using an HP-6643A power supply, up to a significant CFC electro-
degradation. They were periodically stopped for the electrode ex-
amination. In these stops, the electrolyte of the cell was transferred
to the reservoir, the cell was removed, and the electrolyte bypassed
in order to avoid electrolyte and gas losses. Afterward, the elec-
trodes were examined, carefully rinsed in methanol and water, and
dried.
Figure 1. Side view of the undivided two-electrode flow cell, with two steel
bodies to press the cathode ͑SC͒ and the anode ͑SA͒, respectively, onto the
electrolyte compartment ͑EC͒. A PMMA window ͑PW͒ to observe the gas
chamber is also shown.
Each SS body had a terminal for electrical connection. The SS bod-
ies of the anode and the cathode, together with the electrodes de-
scribed previously were pressed onto the electrolyte compartment
͑EC͒ using Viton O-rings and SS screws. The electrolyte compart-
ment was also made with PMMA, with inlet and outlet allowing the
electrolyte circulation. The electrodes were placed vertically, face to
face, with an interelectrode gap of 4 mm. The electrode sections to
be exposed to the electrolyte were 10 cm2 both for the anode and the
cathode.
The cell was integrated in the system depicted in Fig. 2, where it
can be seen on the right of the figure ͑C͒. An electrolyte volume of
0.5 dm3 was placed in a closed glass reservoir ͑R͒, which in turn
was immersed in a water bath at 20°C ͑W͒. From the reservoir the
electrolyte was circulated through the cell at a flow rate of 25 mL
To study the possible metal corrosion and release in solution, the
electrolyte was analyzed by inductively coupled plasma ͑ICP͒ and
mass spectrometry ͑MS͒ detection, both during CFC electrodegra-
dation and in open-circuit conditions. The organic compounds of
small electrolyte volumes were mineralized with HNO3-HClO4 . Af-
ter evaporation and dissolution with HNO3 , the copper, palladium,
and platinum contents were determined using a Perkin-Elmer ELAN
6000 ICP-MS. The surface of the cathode was also analyzed by
X-ray diffraction ͑XRD͒ using a Bragg-Brentano ␪/2␪ Siemens
D-500 diffractometer.
As shown in Fig. 2, a gas phase over the electrolyte was always
present during the electrolyses, and then a product distribution of
CFC and its derivatives between the liquid and the gas according to
their solubility/volatility was expected. The identification of the de-
rivatives was performed by MS coupled to gas chromatography
͑GC͒ using a HP 5890 GC chromatograph with an HP-624 capillary
column of 60 m and an HP 5989A MS detector. The routine analyses
were performed using an HP 6890 chromatograph with the same
column and a thermal conductivity detector ͑TCD͒. The gas carrier
was He and the column flow was 1.0 mL minϪ1. The temperature
program for the oven was 20 min at 30°C, a temperature ramp of
15°C minϪ1 to 200°C, and then the temperature was held at 200°C
for 5 min. All the organic compounds were detected during the 20
min period of constant temperature of the oven. The analyses of the
gas phase were performed by direct injection of 400 ␮L of it to the
chromatograph by means of a syringe. The analyses of the liquid
were performed after extracting the corresponding organic com-
pounds in chloroform, as indicated previously.9 As some of the stan-
dards needed were not available, the amount of derivatives formed
during the electrolyses together with the remaining amounts of CFC,
both in the gas and in the liquid, were obtained correcting the chro-
matographic areas by the relative molar response ͑RMR͒ factors.
The results of the analyses are given in relative amounts of CFC and
its derivatives, excluding evolved hydrogen and solvent vapors, in
the electrolyte, R1 , and in the gas, Rg . Such relative amounts al-
lowed calculating the CFC conversion and the product distribution,
and they were also employed to determine the current efficiency.
The RMR factors were calculated from the molecular diameter ap-
Figure 2. Scheme of the laboratory-scale experimental setup for the CFCs
electrodegradation. The electrolyte is recirculated from a glass reservoir ͑R͒,
immersed in a water-bath ͑W͒, to the cell ͑C͒, by means of a peristaltic pump
͑P͒. A magnetic stirrer ͑S͒ and a key for sample extraction ͑E͒ are also
shown.
Downloaded on 2015-04-06 to IP 169.233.248.124 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

B100
Journal of The Electrochemical Society, 151 ͑2͒ B98-B104 ͑2004͒
proach, which gives accurate results for halogenated organic
compounds,20,21 being 112.3, 94.5, 78.4, 57.3, and 42.5, for CFC 11,
dichlorofluoromethane ͑DCFM͒, chlorofluoromethane ͑CFM͒, fluo-
romethane ͑FM͒, and methane, respectively, and 138.9, 123.8,
101.2, 84.9, and 74.7 for CFC 113, 1,2-dichloro-1,1,2-
trifluoroethane ͑DCTFE͒, CTFE, TFE, and DFE, also respectively.
The pH values of electrolyte samples were measured using a
Metrohm 632 pH meter.
Results and Discussion
Electroreduction of CFC 11.—The electrolyses in 0.75 M NH4Cl
was performed at a constant cell voltage of 1.0 V. In the first 7 h, the
cell current decreased from 100 to 70 mA and afterward it slowly
increased. The cell current was quasi-steady and ca. 120 mA after
about 13 h from the beginning of the experiment. The analyses of
the electrolyte and of the gas showed the preferential formation of
DCFM and CFM. FM was scarcely detected in the electrolyte,
whereas it was found to significant amounts in the gas. In addition,
small amounts of methane, difficult to quantify, were also deter-
mined in the latter. These compounds except methane were also
previously found when a Pb17 or Hg cathodes4 were employed.
Therefore, Cu has certain ability for defluorination. This capacity
has also been evidenced during the electroreduction of chlorotrifluo-
romethane ͑CFC 13͒ and CFC 12 on Cu-supported GDEs in aqueous
media.13,22 Pb-supported GDEs induced only dechlorination of CFC
12.13
The results of the analyses of the electrolyte and of the gas over
the latter during the electroreduction of CFC 11 in NH4Cl are shown
in Fig. 3a and b, respectively. Figure 3a shows the evolution of the
electrolyte composition in CFC and its derivatives together with the
conversion of CFC 11 into DCFM and CFM, in front of the charge
consumed. Fig. 3b shows the same in the gas phase, although with
the addition of FM, which was scarcely found in the electrolyte. As
indicated previously, methane could not be adequately quantified
and therefore, it has not been included in this figure. CFC 11 and its
derivatives were volatile and therefore appear to be distributed in the
gas and in the electrolyte according to their volatility. As FM is
more volatile than its related compounds with more Cl atoms, it is
mainly found in the gas phase. CFM is present in the gas in smaller
amounts than FM, but it dominates over the latter in the electrolyte.
Therefore, we may conclude that the relative amounts of CFC de-
rivatives resulting from the electrolysis decrease in the order
Figure 3. Relative molar composition of CFC 11 and its derivatives DCFM,
CFM, and FM ͑indicated͒ in ͑a͒ the liquid, R1 , and ͑b͒ the gas, Rg , in front
of the charge Q utilized in its electrolysis in methanol ͑70 vol%͒-water
mixture containing 0.75 mol dmϪ3 NH4Cl.
Ͼ
Ͼ
Ͼ
FM methane. This is in agreement with a
DCFM
CFM
stepped removal of Cl atoms from CFC during the electrolysis, fol-
lowing the sequence
CFCl3 ϩ 2eϪ ϩ Hϩ → CFCl2H ϩ ClϪ
CFCl2H ϩ Hϩ ϩ 2eϪ → CFClH2 ϩ ClϪ
CFClH2 ϩ Hϩ ϩ 2eϪ → CFH3 ϩ ClϪ
CFH3 ϩ Hϩ ϩ 2eϪ → CH4 ϩ FϪ
͓1͔
͓2͔
͓3͔
͓4͔
the electrolyte were 49.8% CFC 11, 48.4% DCFM, and 1.8% CFM.
The corresponding amounts of products are listed in Table I, where
the results of the main experiments of this work are summarized. As
shown in Reactions 1 and 2, DCFM needs 2F per mol, whereas
CFM needs 4F per mol. The charge required for this conversion was
9.0 kC, thus indicating a mean current efficiency of about 74%.
As shown in Fig. 3b, CFC 11 concentration in the gas phase is
much higher than in the liquid ͑its Rg being about 75% at the end of
the electrolysis͒. This is in contrast with the results of the analysis of
the liquid ͑its Rg being 48.4%͒. The dominance of the most volatile
compounds in the gas is expected if we assume phase equilibrium
and the absence of great deviations from the ideal behavior of the
mixtures. Therefore, we may suspect that no phase equilibrium is
attained in the system in the absence of suitable mixing of gas and
electrolyte. Note that a greater amount of volatile compounds were
being produced with time, thus requiring the use of adequate pres-
surized systems to avoid losses during operation and extraction of
liquid and gas samples. This has to be considered in further work.
The surface of the cathode was periodically revised after about
7-8 h in order to follow its evolution. It was seen that the cathode
surface gradually became brown, loosing its peculiar brightening
A stepped removal of Cl was also found when using Pb
cathodes,11,17 although no methane was released in such conditions.
The derivatives having fewer Cl atoms are found in a smaller con-
centration due to the also smaller concentrations of the correspond-
ing reactants. The dominance of the most dechlorinated compounds
should be obtained for much longer electrolysis time to allow a
significant conversion of DCFM. The electrolysis was stopped at
Q ϭ 12.2 kC, whereas a charge of about 69.5 kC (0.090 ϫ 8F,
where F is the Faraday constant͒ should be necessary for a complete
conversion of 0.090 mol CFC 11 into methane, assuming that only
Reactions 1-4 are taking place. They are not the only reactions be-
cause hydrogen gas can be evolved11 and oxygen dissolved can be
reduced. A rough estimation of the current efficiency for the conver-
sion of CFC 11 can be done assuming that most of the derivatives
formed are in the liquid phase, as shown elsewhere.11 When the
electrolysis was stopped, the R1 values of CFC and derivatives in
Downloaded on 2015-04-06 to IP 169.233.248.124 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 ͑2͒ B98-B104 ͑2004͒
B101
Table I. Summary of the results obtained for the main experiments of electrodegradation of CFCs 11 and 113 indicating the amount of CFC
consumed „nc…, amount of products „np…, consumed charge „Q…, current efficiency „␧…, initial potential „V0…, and mean corrosion rate of the
Cu cathode „
Electrolyte
…. The initial electrolyte volume was always 0.5 dm3.
vcor
͑␮g hϪ1 cmϪ2
͒
Q ͑kC͒
␧ ͑%͒
nc ͑mol͒
np ͑mol͒
V0 ͑V͒
vcor
12.2
74
1.0a
12
0.75 mol dmϪ3 NH4Cl
ϩ0.090 mol CFC 11
0.1 mol dmϪ3 PTBA
ϩ0.095 mol CFC 11
0.75 mol dmϪ3 NH4Cl
ϩ0.05 mol dmϪ3 PTBA
ϩ0.095 mol CFC 11
0.75 mol dmϪ3 NH4Cl
ϩ50 ppm PdCl2
4.52 ϫ 10Ϫ2
DCFM:4.36 ϫ 10Ϫ2
CFM:1.62 ϫ 10Ϫ3
DCFM:5.63 ϫ 10Ϫ2
CFM:1.81 ϫ 10Ϫ3
DCFM:1.77 ϫ 10Ϫ2
CFM:5.23 ϫ 10Ϫ4
5.81 ϫ 10Ϫ2
1.82 ϫ 10Ϫ2
12.2
3.96
95
91
4.0
1.0
0.8
3.5
12.8
4.10
80
52
1.0a
2.7
8.5
1.0
4.33 ϫ 10Ϫ2
8.48 ϫ 10Ϫ3
DCTFE:3.53 ϫ 10Ϫ2
CTFE:2.77 ϫ 10Ϫ3
TFE:6.55 ϫ 10Ϫ4
DFE:4.54 ϫ 10Ϫ3
DCTFE:2.60 ϫ 10Ϫ3
CTFE:3.36 ϫ 10Ϫ3
TFE:2.52 ϫ 10Ϫ3
ϩ0.084 mol CFC 113
0.1 mol dmϪ3 PTBA
ϩ50 ppm PdCl2
ϩ0.084 mol CFC 113
a Experiment at constant potential.
H2 → 2Hϩ ϩ 2eϪ
͓6͔
and color, presenting sometimes some small blue spots. This indi-
cated some Cu corrosion during the electrolysis, because the elec-
trode was rinsed and dried when the electrolysis was stopped. The
examination of the cathode surface by XRD showed only small
amounts of Cu2O as Cu oxidized species, and therefore, some
amount of copper could be dissolved into the solution. This became
evident by the slight blue color of the electrolyte after sufficient time
of electrolysis. ICP analysis of the electrolyte showed a copper
and only one Hϩ ion is consumed per 2F in each one of Reactions
1-4.
The plots of R1 and Rg for CFC 11 and its derivatives in PTBA
had the same shape as in NH4Cl. However, a greater CFC conver-
sion in PTBA was apparent, because the values of R1 for CFC 11,
DCFM, and CFM at the end of the electrolysis were 38.9, 59.3, and
1.9%, respectively. This means a required charge of 11.6 kC, which
corresponds to a mean current efficiency of about 95% ͑see Table I͒.
In this case, the corrosion rate of Cu was ca. 0.8 ␮g hϪ1 cmϪ2, about
one order of magnitude smaller than in NH4Cl ͑no blue color in the
electrolyte was appreciated͒. The visual inspection of the Cu cath-
ode clearly indicated that it was cleaner than in NH4Cl, and the
XRD analyses of its surface showed only Cu metal. The use of
PTBA has the penalty of a higher cell voltage, but the current effi-
ciency of CFC11 electrodegradation is greater and Cu corrosion
much smaller than in NH4Cl. The explanation of the observed dif-
ferences between NH4Cl and PTBA in the corrosion rate and the
amount equivalent to a Cu corrosion rate of about 12 ␮g hϪ1 cmϪ2
.
This shows that a cathodic current density of ca. 10 mA cmϪ2 ap-
pears to be insufficient to avoid the cathode corrosion. Dissolved
oxygen plays a role in the copper corrosion in aqueous media. How-
ever, copper oxidation by CFC 11 and/or its derivatives, as in the
case of Fe with CCl4 ,23 could not be discarded at first. For this
reason, a parallel experiment of Cu corrosion for about 9 h was
performed using the same system as that employed for the CFC 11
electrodegradation, except that no current flowed between the elec-
trodes. In this case, the amount of copper in the electrolyte was
equivalent to a Cu corrosion rate of about 86 ␮g hϪ1 cmϪ2, i.e.,
much higher than that found during electrodegradation. In addition,
the analysis of the electrolyte showed the formation of DCFM, as in
the electrodegradation process, indicating that CFC 11 could oxidize
copper through the following redox reaction
CFCl3 ϩ Cu ϩ Hϩ → CFCl2H ϩ Cu2ϩ ϩ ClϪ
͓5͔
We may therefore conclude that oxygen and CFC 11 are able to
oxidize Cu and that Cu corrosion could probably be avoided at
sufficiently negative potentials, with higher cathodic current densi-
ties. However, to study the effect of the electrolyte in Cu corrosion,
CFC 11 ͑0.095 mol͒ was electrolyzed in 0.1 mol dmϪ3 PTBA at the
same current densities as in NH4Cl. In this case, the cell voltage to
obtain a quasi-stationary current of about 100 mA was higher than
for the latter electrolyte. The initial cell voltage was 4.0 V and it was
decreased with time to 1.9 V, at the end of the electrolysis ͑see curve
a in Fig. 4͒. These higher cell voltages were most probably due to
the higher resistance of the electrolyte. Although the decrease in
potential to keep constant the cell current could be attributed to
different reasons, we feel that the main one was the increase in
conductivity due to the decrease of pH. Figure 5 shows such a pH
decrease during the electrolyses in this experiment ͑curve a͒. A pH
decrease also took place when NH4Cl was employed as supporting
electrolyte ͑curve b͒, thus justifying the increase in current density
with time at constant potential in such conditions. This net acidifi-
cation clearly resulted from the anodic process, because two Hϩ ions
were produced per 2F
Figure 4. Cell voltage, V, vs. consumed charge, Q, for different experiments
at constant cell current of 100 mA using ͑a͒ 0.1 mol dmϪ3 PTBA with CFC
11 and ͑b͒ 0.1 mol dmϪ3 PTBA with 50 ppm acidic PdCl2 and CFC 113.
Downloaded on 2015-04-06 to IP 169.233.248.124 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

B102
Journal of The Electrochemical Society, 151 ͑2͒ B98-B104 ͑2004͒
Figure 5. Measured pH vs. consumed charge, Q, for different experiments
using ͑a͒ 0.1 mol dmϪ3 PTBA with CFC 11, ͑b͒ 0.75 mol dmϪ3 NH4Cl with
CFC 11, ͑c͒ 0.1 mol dmϪ3 PTBA with 50 ppm acidic PdCl2 and CFC 113,
and ͑d͒ 0.75 mol dmϪ3 NH4Cl with 50 ppm acidic PdCl2 and CFC 113.
current efficiencies can be found in the hydrophobic character of the
tetrabutylammonium cations. It is known that they can be strongly
adsorbed at the electrode,24 thus propitiating more suitable condi-
tions for the CFC 11 reduction.
As in the case of NH4Cl, a parallel experiment of 6 h without
current flow was performed. The electrolyte containing PTBA and
CFC was circulated through the cell at open-circuit conditions. The
period extraction of electrolyte samples showed increasing amounts
of DCFM with time, thus indicating that Reaction 5 also takes place
with PTBA. The amounts of DCFM were smaller, but significant,
being about 3.5% of the DCFM formed when the quasi-stationary
current of 100 mA was applied.
The PTBA concentration employed was near saturation and this
precluded using higher concentrations. As expected, higher cell volt-
ages had to be applied to obtain currents of about 100 mA when its
concentration decreased. Thus, an initial cell voltage of 5.5 V was
required to obtain a cell current of about 100 mA in 0.05 mol dmϪ3
PTBA. Therefore, this salt was employed in combination with
NH4Cl and 0.095 mol CFC 11. The corresponding results are also
shown in Table I. The cell voltage to maintain a quasi-stationary
current of about 100 mA was in the range 1.0-1.2 V, in agreement
with the experiment using NH4Cl. The corrosion rate of Cu ͑3.5 ␮g
hϪ1 cmϪ2͒ and the current efficiency ͑91%͒, intermediate between
NH4Cl and PTBA, makes NH4Cl ϩ PTBA preferable in front of
NH4Cl alone.
The gas pressure over the electrolyte increased during the elec-
trolyses. This was assigned to the formation of the most volatile
compounds, methane, and even hydrogen. Such a pressure increase
did not indicate continuing the electrolyses because the system was
not specially prepared for this. Therefore, a new system, prepared to
sustain high pressures, has to be built up to study greater CFC con-
versions in CFC-concentrated electrolytes.
Figure 6. Relative molar composition of CFC 113 and its derivatives
DCTFE, CTFE, TFE, and DFE ͑indicated͒, in ͑a͒ the liquid, R1 , and ͑b͒ the
gas, Rg , in front of the charge, Q, employed during its electroreduction in
methanol ͑70 vol %͒-water mixture with 0.75 mol dmϪ3 NH4Cl and 50 ppm
acidic PdCl2 .
the previous experiments with CFC 11, the pH of the electrolyte
decreased with time due to the net acidification produced by the
anodic reaction. The increase in the cell current with time may then
be attributed to a slight increase in the electrolyte conductivity. The
XRD analyses of the surface of the cathode at the end of the elec-
trolysis showed the formation of electrodeposited Pd, a PdCu alloy,
together with very small amounts of Cu2O.
As the great part of CFC and its derivatives are in the liquid, the
R1 values of Fig. 6a show that the amount of products formed de-
Ͼ
Ͼ
Ͼ
CTFE TFE. This se-
crease in the order DCTFE
DFE
quence is also found for the Rg values of the same derivatives except
for DCTFE ͑Fig. 6b͒, because the latter is the less volatile product,
and it is expected to be mostly in the electrolyte. The formation of
DCTFE and CTFE can be explained by two different two-electron
reductions of CFC 1139
Electroreduction of CFC 113.—Figures 6a and b show the dis-
tribution of CFC 113 and its derivatives, in the electrolyte and in the
gas, respectively, during the electrolysis of 0.084 mol CFC in 0.75
mol dmϪ3 NH4Cl plus 50 ppm of acidic PdCl2 at a constant cell
voltage of 1.0 V. The initial cell current was about 100 mA and it
presented a net increase with time to about 130 mA. The initial pH
of the electrolyte, curve d in Fig. 5, was smaller than that of NH4Cl
with CFC 11, curve b, because of the addition of acidic PdCl2 . As in
CFCl2-CF2Cl ϩ 2eϪ ϩ Hϩ → CFCIH-CF2Cl ϩ ClϪ
CFCl2-CF2Cl ϩ 2eϪ → CFCl ϭ CF2 ϩ 2ClϪ
͓7͔
͓8͔
Downloaded on 2015-04-06 to IP 169.233.248.124 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 ͑2͒ B98-B104 ͑2004͒
B103
loosely adhered and the amount of CFC derivatives in the electrolyte
was smaller than in NH4Cl alone. DFE was scarcely detected in the
liquid and the current efficiency was comparable to that obtained in
PTBA. This salt then appears to interfere with Pd black electrodepo-
sition and to reduce the current efficiency of its electrodechlorina-
tion when compared with NH4Cl alone, thus being undesirable for
CFC 113 electrodegration with dissolved PdCl2 .
where Reaction 8 is the main reaction taking place without PdCl2 in
solution.2 As discussed elsewhere,9,10 electrodeposited Pd and sub-
valent Pd complexes may catalyze the consecutive hydrogenation of
ethylene derivatives according to the following reaction stoichiom-
etries
CFCl ϭ CF2 ϩ 2eϪ ϩ Hϩ → CFH ϭ CF2 ϩ ClϪ
CFH ϭ CF2 ϩ 2eϪ ϩ Hϩ → C2F2H2 ϩ FϪ
͓9͔
͓10͔
Anode behavior.—As indicated previously, one of the objects of
this paper was to study the behavior of the Pd-30Ag foil of the
hydrogen diffusion anode after prolonged hydrogen oxidation. A
couple of anodes were employed in the electrolyses reported here.
The ICP analyses of the electrolyte did not show significant Pd or Pt
release from the foils. In addition, no methanol derivatives were
found in the GC analyses. This shows that the essential reaction in
the hydrogen diffusion anode is hydrogen oxidation, the oxidation of
Pd, Pt, or methanol being insignificant. However, the Pd-30Ag foil
suffered some deformation. The foil, initially flat, became concave,
its concavity being quite regular and slowly increasing with time.
The interelectrode gap increased in the central part of the electrodes
as a result and the electrodes could not become shorted. A 25 ␮m
thick Pd foil with a section of about 0.5 cm2 was employed in
previous experiments of CFC electrodegradation9,11,17 and the foil
deformation was much less apparent because of the electrode size.
Use of a greater size electrode in this work then magnifies this effect
so that it is much more apparent.
The deformation of the Pd-30Ag foil can be explained by the
inner pressures developed in the foil due to the hydrogen dissolu-
tion. It is known that volumes of hydrogen much higher than a given
volume of Pd can be dissolved in it.26 In this process, two palladium
hydride phases are formed at room temperature, the relative stresses
appearing in the region between such phases leading to the defor-
mation of the thin Pd foils. As indicated elsewhere,26 only one hy-
dride phase is formed at temperatures over 77°C in Pd-30Ag foils,
and therefore, electrodegradation of CFCs at these temperatures has
to be explored in order to study the stability of such foils under these
conditions. Use of a mechanical reinforcement at the hydrogen side
of the foil can be suggested. A temperature increase is expected to
lead to a greater loss of volatiles from the electrolyte, thus increas-
ing the pressure in the gas phase over the electrolyte. Therefore, the
use of a new system, especially prepared to support high pressures,
appears necessary. This has to be tested in further work.
We also suggested the formation of TFE by the two-electron
reduction of DCTFE9 in a similar reaction, producing tetrafluoroet-
hylene from CFC 1144,8
CFCIH-CF2Cl ϩ 2eϪ → CFH ϭ CF2 ϩ 2ClϪ
͓11͔
The relative amount of DFE was higher than that of TFE and
CTFE. This indicates a relatively rapid transformation of CTFE into
TFE ͑Reaction 9͒ and a more rapid transformation of TFE into DFE
͑Reaction 10͒. This is also supported by the initial increase and the
further decrease in the relative amount of CTFE both in the electro-
lyte ͑Fig. 6a͒ and in the gas ͑Fig. 6b͒, although it is more apparent in
the latter. Note, however, that the use of higher current densities
favored the formation of the most dechlorinated derivatives on the
Pb cathode.9,10
The electroreduction efficiency was calculated from the liquid
composition, as in the case of CFC 11 ͑see Table I͒. The conversions
into DCTFE, DFE, CTFE, and TFE, were 42, 5.4, 3.3, and 0.78%,
respectively (R1 values corresponding to Q ϭ 12.8 kC in Fig. 6a͒.
According to Reactions 7-11, the charge employed in such a con-
version was 10.2 kC, thus corresponding to a current efficiency of
about 80%. The corrosion rate of Cu was 8.5 ␮g hϪ1 cmϪ2, some-
what smaller than that found for CFC 11, probably due to the partial
protection of the Cu surface by electrodeposited Pd black. No Cu
corrosion in open circuit was studied for CFC 113, because no Pd
electrodeposition was expected in this condition and therefore the
cathode would not be the same. The Cu corrosion may vary during
Pd electrodeposition, i.e., at different Pd black coverage. In fact, this
opens a new study of CFC degradation in open circuit, where the
CFC can be reduced on Pd whereas Cu is being oxidized in an
overall process similar to Reaction 5. However, it is out of the scope
of the present paper.
The different derivatives formed during the electrolysis of 0.084
mol CFC 113 in 0.1 mol dmϪ3 PTBA with acidic PdCl2 presented
similar relative amounts as in NH4Cl, in the range 3-4%, except
DFE, which was much smaller ͑see Table I͒. This could be due to
the different form of Pd black electrodeposition in NH4Cl and in
PTBA, because PdCl24Ϫ , which is generally involved in the Pd
electrodeposition,25 was not present with PTBA. Adherence prob-
lems of Pd black were evident, because small black particles ap-
peared in the electrolyte and the XRD analysis of the cathode sur-
face at the end of the electrolysis scarcely showed the presence of
Pd, in contrast with XRD analysis of the cathode when NH4Cl was
employed. The visual inspection also indicated that Pd electrodepo-
sition was deficient, because the color of Cu only changed to be
dark, not black as in NH4Cl. However, some completely dechlori-
nated derivatives were still produced. With the use of PTBA, greater
cell voltages than in NH4Cl were also found to obtain quasi-
stationary cell currents of about 100 mA ͑see curve b in Fig. 4͒. The
cell voltages were smaller, however, than for the electrolysis of CFC
11 ͑curve a͒. This can be explained again by the electrolyte pH,
which was higher in the electrolyte with CFC 11 ͑curve a in Fig. 5͒
than in the electrolyte with CFC 113 ͑curve c in Fig. 5͒, because
acidic PdCl2 was added to the latter. It is interesting to observe that
the corrosion rate in PTBA was only 1.0 ␮g hϪ1 cmϪ2, thus con-
firming its positive effect against the cathode corrosion. However,
the current efficiency was only about 52%.
Conclusions
A laboratory-scale flow cell was built to study the behavior of the
electrodes and the formation of the derivatives during the electro-
degradation of CFCs 11 and 113. The electrolyte was nondeaerated
0.75 mol dmϪ3 NH4Cl, 0.1 mol dmϪ3 PTBA, or 0.75 mol dmϪ3
NH4Cl ϩ 0.05 mol dmϪ3 PTBA, adding 50 ppm PdCl2 only for
CFC 113. The electrolyte volume of the cell was 0.5 dm3 and it was
circulated at 25 mL minϪ1 between a Cu cathode and a H2-fed
Pd-30Ag hydrogen diffusion anode activated with Pd ϩ Pt blacks,
the interelectrode gap being 4 mm. The electrolyses were performed
at current densities of ca. 10 mA cmϪ2. A gas chamber was enabled
over the electrolyte, which permitted the distribution of the CFC
derivatives in both phases according to their volatility.
CFC 11 was electrodechlorinated in the different electrolytes,
gradually losing its Cl atoms up to CH4 . The current efficiency in
NH4Cl was ca. 74%. It increased in PTBA ϩ NH4Cl and PTBA to
ca. 91 and 95%, respectively, although the cell voltage had to be
considerably higher to obtain the same current density as in NH4Cl.
The use of PTBA significantly reduced the cathode corrosion.
The electroreduction of CFC 113 in NH4Cl led to different
dechlorinated derivatives, their amounts decreasing in the order
Ͼ
Ͼ
Ͼ
CTFE TFE, and the current efficiency being
DCTFE
DFE
about 80%. The cathode corrosion was smaller than for CFC 11
because of the Pd black electrodeposition. PTBA also decreased the
cathode corrosion, but Pd black electrodeposition was poor and the
current efficiency decreased to about 50%.
Electrolysis of CFC 113 in NH4Cl ϩ PTBA led to the same Pd
electrodeposition problems as in the case of PTBA. The deposit was
Downloaded on 2015-04-06 to IP 169.233.248.124 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

B104
Journal of The Electrochemical Society, 151 ͑2͒ B98-B104 ͑2004͒
8. K. M. Smirnov, A. P. Tomilov, L. G. Feoktistov, and M. M. Gol’din, Zh. Prikl.
Khim. (S.-Peterburg), 51, 701 ͑1978͒.
9. P. L. Cabot, M. Centelles, L. Segarra, and J. Casado, J. Electrochem. Soc., 144,
3749 ͑1997͒.
10. P. L. Cabot, M. Centelles, L. Segarra, and J. Casado, J. Electroanal. Chem., 435,
255 ͑1997͒.
11. P. L. Cabot, L. Segarra, and J. Casado, Electrochem. Solid-State Lett., 6, B15
͑2003͒.
12. E. Delli, S. Kouloumtzoglou, G. Kyriacou, and C. Lambrou, Chem. Commun.
(Cambridge), 1998, 1693.
13. N. Sonoyama and T. Sakata, Environ. Sci. Technol., 32, 375 ͑1998͒.
14. A. Schizodimou, G. Kyriacou, and Ch. Lambrou, J. Electroanal. Chem., 471, 26
͑1999͒.
The increasing amount of volatile derivatives with time increased
the pressure in the cell, thus suggesting building up a system pre-
pared to support inner pressures significantly over the atmospheric
one.
In spite of its chemical stability, the Pd-30Ag foil of the hydro-
gen diffusion anode presented some concave deformation, which
increased the interelectrode gap during prolonged H2 oxidation. The
dimensional stability of the anode needs improvement and its long-
time operation has to be studied in depth.
Acknowledgments
15. N. Georgolios, G. Kyriacou, and G. Ritzoulis, J. Appl. Electrochem., 31, 207
͑2001͒.
16. N. Sonoyama, K. Ezaki, H. Fujii, and T. Sakata, Electrochim. Acta, 47, 3847
͑2002͒.
17. P. L. Cabot, M. Centelles, L. Segarra, and J. Casado, J. Electrochem. Soc., 147,
3734 ͑2000͒.
18. V. L. Kornienko, G. A. Kolyagin, G. V. Kornienko, and Yu. V. Saltykov, Sov.
Electrochem., 28, 412 ͑1992͒.
19. P. L. Cabot, E. Guezala, J. C. Calpe, M. T. Garcia, and J. Casado, J. Electrochem.
Soc., 147, 43 ͑2000͒.
20. M. J. Height, E. M. Kennedy, and B. Z. Duglogorski, J. Chromatogr., A, 841, 187
͑1999͒.
21. E. F. Barry, R. S. Fischer, and D. M. Rosie, Anal. Chem., 44, 1559 ͑1972͒.
22. N. Sonoyama and T. Sakata, Environ. Sci. Technol., 32, 4005 ͑1998͒.
23. M. M. Scherer, J. C. Westall, M. Ziomek-Moroz, and P. G. Tratnyek, Environ. Sci.
Technol., 31, 2385 ͑1997͒.
24. H. Lund and O. Hammerich, in Organic Electrochemistry, 4th ed., p. 275, Marcel
Dekker, New York ͑1991͒.
The authors gratefully acknowledge the financial support of this
work by Carburos Metalicos S.A. The authors also thank the Servei
´
´
`
Cientıfico-Tecnics de la Universitat de Barcelona for the GC-MS
and ICP-MS analyses.
LCTEM assisted in meeting the publication costs of this article.
References
1. E. J. A. X. van de Sandt, A. Wiersma, M. Makkee, H. van Bekkum, and J. A.
Moulijn, Appl. Catal., A, 155, 59 ͑1997͒.
´
2. A. Savall, R. Abdelhedi, S. Dalbera, and M. L. Bouguerra, Electrochim. Acta, 35,
1727 ͑1990͒.
´
3. A. Savall, R. Abdelhedi, S. Dalbera, and M. L. Bouguerra, J. Appl. Electrochem.,
20, 1045 ͑1990͒.
4. Montecatini Edison S.p.a., Italian Pat. 852,487 ͑1969͒.
5. K. M. Smirnov, A. P. Tomilov, L. G. Feoktistov, and M. M. Gol’din, Zh. Prikl.
Khim. (S.-Peterburg), 51, 703 ͑1978͒.
25. J. F. Llopis and F. Colom, in The Encyclopedia of the Electrochemistry of the
Elements, pp. 253-275, Marcel Dekker, New York ͑1976͒.
26. J. N. Armor, CHEMTECH, 1992, 557.
6. S. Wawzonek and S. Willging, J. Electrochem. Soc., 124, 860 ͑1977͒.
7. V. L. Kornienko, G. A. Kolyagin, and G. V. Kornienko, Russ. J. Electrochem., 33,
896 ͑1997͒.
Downloaded on 2015-04-06 to IP 169.233.248.124 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).