Selective liquid-phase hydrodechlorination of chlorotrifluoroethylene over palladium-supported catalysts: Activity and deactivation

Article abstract of DOI:10.1007/s10562-010-0370-8

Liquid-phase hydrodechlorination of chlorotrifluoroethylene (CTFE) to trifluoroethylene (TrFE) with molecular hydrogen was studied over palladium-supported catalysts. BaSO₄, Al₂O₃ and activated carbon (AC) were used as supports, respectively. The results showed that the Pd/AC catalysts exhibit higher activity than Pd/BaSO₄ and Pd/Al₂O₃. The treatment of activated carbon with HNO ₃ led to a considerable increase of surface functional groups containing oxygen atoms, which resulted in a higher dispersion of palladium on the supports and enhancement of catalytic activity. The stability of the catalyst was investigated, one reason for the inhibition was the accumulation of NaCl on the surface of Pd/AC that blocks the pores of carbon support. The activity of Pd/AC could partially recovered by washing with water. The other irreversible deactivation of the catalyst are from the change of particle size and the pore structure, leaching of Pd, a decrease of BET surface area and Pd surface area of Pd/AC catalyst.

Full text of DOI:10.1007/s10562-010-0370-8

Catal Lett (2010) 138:68–75
DOI 10.1007/s10562-010-0370-8
Selective Liquid-phase Hydrodechlorination
of Chlorotrifluoroethylene over Palladium-Supported
Catalysts: Activity and Deactivation
•
Bao-Chuan Meng Zhao-Yang Sun
•
•
Jin-Peng Ma Gui-Ping Cao Wei-Kang Yuan
•
Received: 8 December 2009 / Accepted: 10 May 2010 / Published online: 22 May 2010
Ó Springer Science+Business Media, LLC 2010
Abstract Liquid-phase hydrodechlorination of chlorotri-
fluoroethylene (CTFE) to trifluoroethylene (TrFE) with
molecular hydrogen was studied over palladium-supported
catalysts. BaSO4, Al2O3 and activated carbon (AC) were
used as supports, respectively. The results showed that the
Pd/AC catalysts exhibit higher activity than Pd/BaSO4 and
Pd/Al2O3. The treatment of activated carbon with HNO3
led to a considerable increase of surface functional groups
containing oxygen atoms, which resulted in a higher dis-
persion of palladium on the supports and enhancement of
catalytic activity. The stability of the catalyst was investi-
gated, one reason for the inhibition was the accumulation
of NaCl on the surface of Pd/AC that blocks the pores of
carbon support. The activity of Pd/AC could partially
recovered by washing with water. The other irreversible
deactivation of the catalyst are from the change of particle
size and the pore structure, leaching of Pd, a decrease of
BET surface area and Pd surface area of Pd/AC catalyst.
by-products [1]. Catalytic hydrodechlorination of chlori-
nated organic compounds over noble metal catalysts has
been recognized as a facile and efficient procedure to
eliminate these compounds [2, 3]. Among noble metals in
Group VIII known for those procedures, palladium was
reported to be the most active metal that catalyze selective
replacing of chlorine with hydrogen. As Pd not only pro-
motes C–Cl bond cleavage efficiently, but also shows
poisoning resistance [4, 5].
Heterogeneous catalytic hydrodechlorination is con-
ventionally carried out in either gas [6–8] or liquid phase
[2, 3, 9]. The hydrogen source can be molecular hydrogen
[3, 6, 10] or other hydrogen donors such as alcohols,
2-propanol [11, 12], sodium hypophosphite, sodium boro-
hydride [13], as well as proaromatic compounds [14].
Trifluoroethylene (TrFE), produced by catalytic hyd-
rodechlorination of chlorotrifluoroethylene (CTFE), is an
important chemical intermediate for synthesis of many
useful compounds. For example, 1,1,1,2-tetrafluoroethane
(HFC-134a) is known as the best alternative of freons. It also
can copolymerize with olefins, for instance, the copoly-
merization product of TrFE with vinylidene fluoride (VDF)
has extraordinarily piezoelectric and pyroehctric properties,
that are widely used as transducers or detectors [15].
Porosity and surface area of the catalyst support have
remarkable effects on the catalytic activity and selectivity.
Francisco R R [16] found that in many cases the catalyst
properties were a function of porosity and surface area,
which was relatively straightforward, as the microporosity
is responsible for the high surface area of activated carbons
act as support to the deposited metal particles, thus
favoring large dispersion. Activated carbon with high sur-
face area and a well controlled porosity is essential to metal
dispersions, which usually results in a high catalytic
activity and selectivity. It was also found that the properties
Keywords Trifluoroethylene ꢀ Activated carbon ꢀ
Pd/AC catalyst ꢀ HNO3 pretreatment ꢀ Deactivation
1 Introduction
Chlorinated organic compounds exhibit high toxicity and
form a class of environmentally undesirable compounds
that are produced by a wide range of industrial processes as
B.-C. Meng ꢀ Z.-Y. Sun ꢀ J.-P. Ma ꢀ G.-P. Cao (&) ꢀ
W.-K. Yuan
East China University of Science and Technology, UNILAB
Research Center of Reaction Engineering, State-key Laboratory
of Chemical Engineering, Shanghai 200237, China
e-mail: gpcao@ecust.edu.cn
123

Selective Liquid-phase Hydrodechlorination of CTFE
69
of activated carbon were determined by not only its surface
area and pore structure, but also the amount of surface
functional groups [16, 17]. After pretreated with strong
oxidizing agents, such as HNO3 or H2O2, more oxygen
containing functional groups such as –OH and –COOH are
created on the surface of activated carbon [16], which
would enhance the interaction of the surface sites with
guest molecules. Surface treatment makes the carbon
surface more hydrophilic and accessible to the aqueous
solution of metal precursors during impregnation.
available in literature including the type of catalyst, the
processing conditions and catalyst deactivation.
In this paper, different catalyst supports such as BaSO4,
Al2O3 and activated carbon were used to load palladium.
Bohem titration for determining surface groups, N2 physical
adsorption and Scanning Electron Microscope (SEM) were
adopted to examine the changes of physical and chemical
properties of the surface of activated carbon after treated
with HNO3, respectively. CO chemisorption and Trans-
mission Electron Microscopy (TEM) were conducted to
clarify the effect of activated carbon pretreatment on the
dispersion of Palladium. In addition, we compared the
activity and selectivity of the Pd/AC catalysts whose sup-
ports that were treated with and without HNO3, respectively.
Furthermore, the stability of the Pd/AC catalyst was evalu-
ated, and the deactivation mechanism was discussed on the
basis of the characterization of fresh and used catalysts.
Unfortunately, major drawback of the hydrodechlori-
nation catalyst is its deactivation. Catalyst stability is rec-
ognized as very important aspect of catalyst studies for
industry and academia [18]. Thereby, since it is crucial to
optimize processing conditions and avoid premature cata-
lyst deactivation. Yet, researches on catalyst deactivation
receive less attention than the discovery of new catalysts
[18].
Urbano and Marinas [14] have reported that the deac-
tivation of the catalysts has been related to several factors,
i.e. the inhibitory effect of HX formed as by-product, for-
mation of carbonaceous deposits, sintering of the active
phase and degradation of the catalyst by the corrosive acid
formed in the reaction. However, formation of carbona-
ceous and sintering of the active metal are common spe-
cially in the gas phase hydrodechlorination processes
[8, 19] rather than liquid phase. At present, it is widely
accepted that the HCl formed from the hydrodechlorination
reaction is the main reason to inhibit the activity of pal-
ladium metal-supported catalysts. This inhibition effect of
the reaction product HCl, is manifested in a negative
reaction order (-1), as reported by Somorjai and
co-workers [20, 21] working with model catalysts.
The negative effect of HCl may result from two essen-
tial factors. Firstly, the competitive adsorption between
HCl and the substrate on the palladium surfaces as reported
by Coq [4], in his study on gas-phase hydrogenolysis of
chlorobenzene. Secondly, the corrosive medium produced
from the acid released may degrade either the support [9]
or the metal phase [22] of the catalyst. That is the reason
why a base, such as H2N–R–NH2, Zn powder, NaOH,
KOH, etc. is normally added to neutralize HCl and avoid
catalyst deactivation [2, 3, 10, 14].
2 Experimental
2.1 Materials
Chlorotrifluoroethylene (purity = 99.5%) was obtained
from Shanghai 3F New Materials CO., Ltd., and used
without further purification. PdCl2 and all other chemical
reagents (minimum 96% purity) used in the experiments
were purchased from Sinopharm Chemical Reagent Co.,
Ltd. The activated carbon having a BET surface area of
1397 m2/g, and a mean particle size of about 50 lm was
purchased from SHHXTC. Ultrapure water (Advantec
RFD25ORA Water Distilling Apparatus) was used in all
the experiments. Hydrogen and nitrogen with the purities
of 99.99% were purchased from Shanghai Jifu Gas Co.,
Ltd.
2.2 Pretreatment of Activated Carbon
and Characterization
The typical procedure of pretreatment of activated carbon
was as follows: 10 g of activated carbon original was
introduced in 100 mL HNO3 (10 wt%) solution and stirred
at 90 °C under reflux for 4 h to ensure that the volume was
kept constant along the process. After the treatment, samples
were washed with ultrapure water until neutral pH and dried
in oven at 110 °C overnight. This sample was labeled as
AC1. The activated carbon without treated by HNO3 was
washed with ultrapure water until neutral, and noted as AC0.
Bohem titration with different strength of bases was
used to determine the contents of various functional groups
containing oxygen atoms on the surface of activated carbon
(AC0 and AC1), bases where NaOH was used for deter-
mining the amount of carboxyl groups, Na2CO3 for that of
Furthermore, the catalysts degradation due to the cor-
rosive environment in the reaction system also should be
taken into account, e.g. Forni et al. [23] detected 25% of
the BET area and 20% of palladium leaching from the
catalyst after running for 20 h in the process of hydrode-
chlorination of Polychlorinated biphenyls (PCBs) on Pd/C.
Similar results were corroborated by Concibido et al. [3]
and Cobo et al. [11].
However, for the hydrodechlorination of CTFE to TrFE
over palladium-supported catalysts, little information is
123

70
B. Meng et al.
lactones and carboxyl groups, NaOH for that of phenols,
lactones and carboxyls and NaOC2H5 for that of carbonyl,
phenols, lactones and carboxyls, respectively [24, 25].
Weighed amount of activated carbon sample (about 1.0 g)
was placed in 25 mL of 0.1 M solution of the base. After
sealing and equilibration with gentle agitation for 48 h at
30 °C, then the suspension was filtered. 10 mL of the fil-
trate was pipetted and 50 mL standard 0.1 M HCl solution
was added to it, back-titrated with standard 0.1 M NaOH.
The content of all those functional groups containing
oxygen atoms were calculated according to the assumption
of NaOH solution neutralizing acidic groups.
microscope with a LaB6 filament as the source of electrons,
operating at 200 kV.
Pd dispersion and average active particle size of the
catalyst was determined by measuring CO pulse-adsorp-
tion (Micromeritics Autochem 2910) in a flow of helium
after reduction in 10% H2–N2 at 450 °C for 2 h. The
catalyst surface was cleaned by heating in a flow of
helium at 120 °C for 1 h. The palladium dispersion
was calculated with respect to the total carbon monox-
ide uptake by assuming that carbon monoxide was
adsorbed on surface palladium atoms at a stoichiometric
ratio of 1:1.
The amounts of alkali consumption during these pro-
cedures can be given by Eq. 1.
The powder catalyst particle size distribution was
evaluated by laser diffraction using a Mastersizer 2000
analyzer (Malvern Instrument).
b ¼ ðV ꢁ cNaOH þ 25 ꢁ c0 ꢂ 50 ꢁ cHClÞ=m
ð1Þ
Atomic absorption spectrometry (AAS) was used to
determination the Pd content of the catalyst samples. The
procedure is as follows: aqua regia was added to a known
amount of the catalyst and the mixture was evaporated to
dryness. The final residue was then dissolved in 0.1 mol/L
HCl to a volume of 50 mL [3] and the solution was ana-
lyzed using TJA IRIS 1000.
where b is the amount of alkali consumption (mmol/g), V is
the volume of sodium hydroxide solution which used to
titrate excessive HCl (mL), c0 is the concentration of base
used (mol/L), and m is the amount of activated carbon
samples (g).
So, the content of carboxyl groups, lactone groups,
phenolic groups, carbonyl groups, acidic surface function
groups can be expressed as bNaHCO , bNa CO –bNaHCO ,
3
2.4 Hydrodechlorination Procedure
3
2
3
bNaOH–bNa CO , bNaOC H –bNaOH, and bNaOC H (mmol/g),
respectively.
2
3
2
5
2
5
The liquid-phase hydrodechlorination of CTFE with
molecular hydrogen was carried out in a 1-L high pres-
sure bath reactor. Gauges were equipped to record the
pressure and temperature inside the reactor throughout the
experiment.
2.3 Catalyst Preparation and Characterization
Pd/AC, Pd/BaSO4 and Pd/Al2O3 catalysts were prepared
by wet impregnation method. The precursor used was
PdCl2 and the supports were impregnated with aqueous
solution of the precursor in the appropriate concentration.
After the impregnation procedure, the mixture was filtered.
Then, the catalysts were dried at 110 °C overnight and
followed by calcination in nitrogen stream at 400 °C for
4 h. Finally, the catalysts thus obtained were reduced in
hydrogen stream at 300 °C for 4 h. Pd content of those
catalysts was 3 wt%.
A typical operational procedure used in each experiment
was as follows: a certain amount of CTFE, methanol,
water, sodium hydroxide and catalyst samples were added
in the reactor respectively. After the temperature leveled
off at 100 °C, which spent about 30 min, the reactor was
charged with hydrogen to the pressure of 1.25 MPa.
Simultaneous, the stirring device started and the reaction
timed. In the end of the experiment, the reactor was cooled
to room temperature and the products were collected and
analyzed by a gas chromatograph equipped with a flame
ionization detector (FID). All products were identified by
Micromass GCTTM GC-mass spectroscopy. The main
product was TrFE, in addition, a small amount of vinyli-
dene fluoride (VDF) was detected.
The nitrogen adsorption–desorption isotherms were
measured on a Micromeritics ASAP 2010 analyzer (USA).
The samples were degassed at 190 °C for 6 h to remove the
surface impurities. The surface area was determined by
adsorption of nitrogen at 77 K and calculated using the
Brunauer–Emmett–Teller (BET) method. The pore size
distributions were calculated using the Barrett–Joymer–
Halenda (BJH) method on the adsorption branch.
The reaction equation is:
CF2 = CFCl + H2 ! CF2 = CHF + HCl
Scaning electron microscopy (SEM) was employed to
get insight into the morphology of the samples. Samples
were sputtered with gold and the images were acquired in
JEOL JSM6360LV (Japan) operated at 20 kV.
In addition to the main reaction, there is a follow-up
reaction as follows to produce by-product,
CF2 = CHF + H2 ! CH2 = CF2 þ HF
The conversion of CTFE and selectivity of TrFE were
calculated using Eqs. 2 and 3.
Transmission electron microscopy (TEM) images were
collected using a JEOL JEM 2100 (Japan) electron
123

Selective Liquid-phase Hydrodechlorination of CTFE
71
‘‘catalyst’’; (iii) without any solid. In these three cases, the
conversion of CTFE was not detected.
FeedCTFE(mol)ꢂUnreactedCTFE(mol)
conversionCTFE
¼
FeedCTFE(mol)
ꢁ100ð%Þ
ð2Þ
3.3 The Effect of HNO3 Pretreatment on Activated
Carbon
TrFE produced (mol)
Feed CTFE (mol)ꢂUnreacted CTFE (mol)
ꢁ100ð%Þ ð3Þ
selectivityTrFE
¼
The influence of HNO3 treatment on textual properties and
apparent morphology of activated carbon (AC0 and AC1)
is shown in Table 1 and Fig. 2, respectively. It is demon-
strated in Table 1 and Fig. 2 that HNO3 pretreatment has
no significant impact on textural properties though BET
surface area of AC1 decreased slightly and apparent mor-
phology was changed little. It is generally reported that the
liquid phase oxidation does not change the texture of
activated carbons significantly [27, 28], although under
more drastic conditions a decrease in surface area and pore
volume has been observed while the average micropore
width increases, as a result of the collapse of pore walls
[29].
3 Results and Discussions
3.1 The Effect of Different Catalyst Support
The activities of the synthesized catalysts were investigated
in the hydrodechlorination of CTFE to TrFE in a batch
autoclave and the experimental results were displayed in
Fig. 1. It is seen from Fig. 1 that Pd/AC1 is the most active
catalyst, followed by Pd/AC0, Pd/Al2O3 and Pd/BaSO4, i.e.
activated carbon is superior to Al2O3 and BaSO4 when
used as catalyst support, and the N2 physisorption data of
the catalysts are shown in Table 1. Table 1 shows that this
phenomenon can attributed to high surface area and well
developed porosity of activated carbon which are essential
for achieving large metal dispersions and high catalytic
activity. It also can be seen from Fig. 1 that Pd/AC1
appears higher catalytic activity compared to Pd/AC0,
whose reason will be interpreted in Sect. 3.3.
The contents of functional groups containing oxygen
atoms on AC0 and AC1 were analyzed by Bohem titration
method, and the results are shown in Table 2. As shown in
Table 2, although the concentration of lactone groups
decreases slightly, the activated carbon treated with HNO3
obviously possesses higher concentration of functional
groups containing oxygen atoms, in which carboxyl groups
increase significantly. This can be interpreted from the
mechanism for carbon oxidation proposed by Figueiredo
et al. [27] as follows.
3.2 Blank Experiments with CTFE
Cf þ O2 ! carbonyls, ethersðþCO)
Cf þ O2 ! lactones, anhydrides
Cf þ carbonyls, ethers þ O2 ! CO2
þ CO or carbonyls, ethers
ðIÞ
ðIIÞ
In order to find out whether any reaction product can be
formed in the absence of palladium-supported catalyst,
several experiments with hydrogen and CTFE were made
under the same conditions as the catalytic reaction, namely:
(i) using activated carbon original (AC0) as ‘‘catalyst’’; (ii)
using activated carbon after treatment with HNO3 (AC1) as
ðIIIÞ
ðIVÞ
Cf þ carbonyls, ethers þ O2 ! lactones, anhydries
Table 1 The N2 physisorption data of different catalysts, activated
carbon treated with and without HNO3, respectively, and the catalyst
after 10 runs
Catalyst
BET surface
area, m2/g
Pore volume,
cm3/g
Pore size,
nm
Fresh Pd/BaSO4
Fresh Pd/Al2O3
AC0
7.13
136.74
0.010
0.230
0.390
0.350
0.330
0.300
0.077
20.810
4.950
2.180
2.170
2.170
2.160
2.350
1397.34
1411.25
1175.29
1188.20
373.32
Fresh Pd/AC0
AC1
Fresh Pd/AC1
Pd/AC1 after 10 runs
AC0 the activated carbon without the treatment of HNO3, AC1 the
activated carbon with the treatment of HNO3
Fig. 1 The results on different catalysts
123

72
B. Meng et al.
Fig. 2 SEM images of AC0
(a) and AC1 (b)
Table 2 The concentration s of functional groups containing oxygen atoms in AC0 and AC1
Activated
carbon
Carboxyl groups,
mmol/g
Lactone groups,
mmol/g
Phenolic hydroxyl groups, Carbonyl groups,
mmol/g
Acidic surface function groups,
mmol/g
mmol/g
AC0
AC1
1.0064
1.7700
0.0803
0.0605
0.2749
0.4104
0.6118
0.8730
1.9734
3.1140
Fig. 3 TEM images of Pd/AC0
(a) and Pd/AC1 (b)
To clarify the effect of activated carbon pretreatment on
the dispersion of palladium, TEM and CO chemisorption
were conducted with Pd/AC catalysts respectively and
TEM images of Pd/AC0 and Pd/AC1 were present in
Fig. 3. As is seen in Fig. 3, when HNO3 treated activated
carbon (AC1) is used as support, the catalyst gave a quite
smaller size of Pd particles and a quite higher dispersion of
Pd particles. We selected randomly about 50 Pd particles
from Fig. 3a and b, respectively, measured the diameters of
the particles, and calculated the average diameters. The
average Pd particles diameter is 11.984 nm on Pd/AC0,
and 4.923 nm on Pd/AC1.
lactones, anhydries ! CO2ðþ COÞ þ 2Cf
carbonyls, ethers ! CO þ Cf
ðVÞ
ðVIÞ
where, Cf is a free active site on the activated carbon
surface. Reactions (V) and (VI) related to the decomposi-
tion of the surface groups can be neglected. It was reported
that the gas phase oxidation of activated carbon mainly
increases the concentration of hydroxyl and carbonyl sur-
face groups, while oxidation in the liquid phase increases
especially the concentration of carboxylic groups [27].
As shown above, activated carbon treated with HNO3
produced a decrease in surface area and microporosity due
to the destruction of the pore walls by the oxidation
treatment. The concentration of functional groups con-
taining oxygen atoms increased significantly especially for
carboxylic groups, which made the surface more accessible
to the aqueous solutions of the metal precursors during
impregnation. It is reported [16] that in impregnation step,
the interaction of palladium precursor with activated car-
bon is favored by the presence of oxygen surface groups,
which make the carbon surface more hydrophilic, and thus,
a better dispersion of the palladium precursor is achieved.
Table 3 presents the dispersion and the average particle
size of Palladium catalysts on the activated carbon deter-
mined by CO chemisorption. The diameter of Pd particles
is 14.1856 nm on Pd/AC0, and 5.3518 nm on Pd/AC1,
which are agree well with that from TEM images. Obvi-
ously, it can be seen from Table 3 that Pd dispersion
Pd/AC1 is significantly improved compared to Pd/AC0,
and the particle size of palladium is smaller. The results of
Fig. 3 and Table 3 show that Pd dispersion of a carbon-
supported Pd catalyst is closely related to the concentration
123

Selective Liquid-phase Hydrodechlorination of CTFE
73
Table 3 CO chemisorption
Catalyst
Pd dispersion, %
Pd particle
diameter, nm
Pd surface area,
m2/g metal
results of Pd/AC0, Pd/AC1
fresh and after used 10 runs
Fresh Pd/AC0
7.8986
20.9363
18.2974
14.1856
5.3518
6.1237
35.1883
93.2710
81.5148
Fresh Pd/AC1
Pd/AC1 after 10 runs
Fig. 4 Activity and selectivity of Pd/AC1 catalyst used in repeated
hydrodechlorination runs (open square—conversion of CTFE; filled
square—selectivity of TrFE)
Fig. 5 Reactivation of Pd/AC1 catalyst with water after 10 runs
3.5 Recovering the Used Catalyst by Aqueous
Washing
of functional groups containing oxygen atoms on activated
carbon which make the carbon surface more hydrophilic
and a high Pd dispersion was achieved.
Washing the used catalyst desorbs the substance that
adsorbed on the surface of the catalyst. This method is
facile to identify the causes of the catalyst deactivation;
moreover, the reactivated catalyst could be regenerated.
Thus, the catalyst used after 10 runs was recovered and
washed by stirring in water for 24 h, before a new
hydrodechlorination run using a fresh solvent was carried
out.
3.4 The Lifespan of Pd/AC1 Catalyst
The Pd/AC1 catalyst was recovered after each run and
directly reused in the next run using a fresh solvent.
Moreover, NaOH was added to eliminate HCl liberated in
the hydrodechlorination reaction. Ten hydrodechlorination
runs with the same catalyst were carried out in the mixture
of methanol and water. The result of lifespan test of
Pd/AC1 catalyst is shown in Fig. 4. As shown in Fig. 4, the
activity of Pd/AC1 catalyst decreased slightly in the first
three runs, and then maintained stability up to the 10th run.
For the first run, the fresh catalyst was used, and the con-
version of CTFE is 95.96%. However, after the 10 runs of
the catalyst, the conversion of CTFE decrease to 77.04%.
There is about 19.72% of activity of the catalyst was lost.
Meanwhile, the selectivity of TrFE after the first cycle
remained unchanged ([97%). The dissociation energies of
C–F and C–Cl bonds are 109 and 81 kcal/mol, respectively
[29]. Due to the more strength of the C–F bond, fluorides are
more resistant to the hydrogenation. This difference makes
it possible for the selective hydrodehalogenation reactions
in polyhalides containing different halogen atoms [10]. For
hydrodehalogenation of CTFE, the main product is TrFE
favorable while the quantity of VDF can be ignored.
Figure 5 shows that washing the catalyst with water
could recover the catalytic activity partially, which signi-
fied that it is effective in washing out the substances
absorbed on the catalyst.
It has been reported that in the catalytic hydrode-
halogenation reaction, the addition of bases, e.g. NaOH
or NH3, is effective in suppressing the deactivation of
catalyst, because the halogen ions responsible for the
poisoning are removed from the catalyst surface [3, 14].
Salts produced in the reactions such as NaCl are likely to
deposit on the catalyst surface decreasing the catalytic
activity; washing the catalyst with water could restore the
activity of the catalyst [3, 10, 30]. Ukisu et al. [31] in
their studies on hydrogenation of aromatic chlorides over
Rh-Pt/C catalyst have found that the activity of the cat-
alyst gradually decreased attribute to the accumulation of
NaCl on the catalyst surface and water could wash away
these salts and the activity of the catalyst could recovered
completely.
123

74
B. Meng et al.
Fig. 6 SEM images of Pd/AC1
catalysts fresh (a) and after used
10 runs (b), respectively
3.6 Other Factors for Deactivation
In Sect. 3.5, we found that NaCl accumulation on catalyst
surface is one of the reasons causing deactivation, and after
washed with water, the catalytic activity could recover
partially, so there must be other reasons causing
deactivation.
The SEM images of Pd/AC1 catalysts fresh and after
used for 10 runs are shown in Fig. 6. It is clearly observed
from Fig. 6 that the particle size and pore structure of the
catalyst are changed, i.e. catalyst reuse procedure
decreased the porosity and BET surface area of the acti-
vated carbon, the results of Nitrogen adsorption–desorption
analysis shows that the BET surface area, pore volume and
pore size are decreased from 1188.20 m2/g, 0.30 cm3/g,
2.16 nm to 373.32 m2/g, 0.077 cm3/g, 2.35 nm respec-
tively after used 10 runs (see Table 1). Support pore
occlusion and partial destruction of the support structure
have been recognized as possible reasons for surface area
decrease [32].
Fig. 7 The particle size distribution of Pd/AC1 catalyst fresh and
after used 10 runs
the catalyst, which also contribute to the inreversible
deactivation of catalyst.
4 Conclusions
The particle size distribution of Pd/AC1 catalysts (fresh
and after 10 runs) was also characterized, and the results
are shown in Table 1 and Fig. 7. The results from Table 1
and Fig. 7 strengthened the conclusion shown in Fig. 6 that
the particle size distribution of Pd/AC1 after 10 runs was
more non-uniform, i.e. the number of small particle
decreased slightly while the number of big particle
increased significantly, which also leads to a decrease of
BET surface area and catalytic activity.
The results obtained in the catalytic hydrodechlorination of
CTFE over various palladium-supported catalysts allow us
to draw several interesting conclusions as regards effects
of supports of the catalyst and the reasons of catalyst
deactivation.
(1) Because of high surface area and well developed
porosity of activated carbon, Pd/AC catalysts show
higher activity than Pd/Al2O3 and Pd/BaSO4.
According to the results of CO chmisorption shown in
Table 3, Pd/AC1 after 10 runs showed a slightly decrease
in Pd disperision and Pd surface area, and at the same time,
particle size of Pd increased a bit.
(2) Activated carbon after HNO3 treatment produced a
decrease in surface area and microporosity due to the
destruction of the pore walls by the oxidation
treatment. While the concentration of functional
groups containing oxygen atoms was increased sig-
nificantly especially carboxylic groups, which make
the carbon surface more hydrophilic, and thus, a
better dispersion of the palladium precursor is
achieved.
In order to determine the difference between the fresh
and the used catalyst, the palladium content of the catalyst
after 10 runs was also detected via AAS method. The Pa
content of the catalyst after 10 runs is 2.394%. Compared
with that of the fresh catalyst (3% Pd), there were about
20.2% of Pd lost from the catalyst after used 10 runs, which
agrees well with the results shown in Fig. 4. The main
reason of deactivation of the catalyst was Pd losing from
(3) The stability of Pd/AC1 catalyst was investigated.
The catalyst exhibits a slightly deactivation in
123

Selective Liquid-phase Hydrodechlorination of CTFE
75
8. Meshesha BT, Chiment RJ, Medina OF et al (2009) Appl Catal B
87:70
9. Aramendia MA, Borau V, Garcia IM et al (1999) J Catal 20:392
10. Xia C, Xu J, Wu W et al (2004) Catal Commun 5:383
11. Cobo M, Quintero A, Correa CMD (2008) Catal Today 133–
135:509
repeated runs while the used catalyst after washed
with water, the activity could recovered partially,
signifying that the reversible deactivation was due to
the accumulation of NaCl on the surface of Pd/AC
and blocks the pores of carbon support and washing
with water could remove it.
12. Ukisu Y (2008) Appl Catal A 349:229
13. Pri-Bar I, James BR (2007) J Mol Catal A: Chem 264:135
14. Urbano FJ, Marinas JM (2001) J Mol Catal A: Chem 173:329
15. Pinto NJ, Bao ZX (1998) Chin J High Press Phys 12:223
16. Rodriguez-Reinoso F (1998) Carbon 36:159
17. Pigamo A, Besson M, Blanc B et al (2002) Carbon 40:1267
18. Birtill JJ (2003) Catal Today 81:531
(4) The irreversible deactivation of the Pd/AC1 catalyst
attributed to the change of particle size and the pore
structure, losing of palladium, decrease of BET
surface area and Pd surface area.
19. Wiersma A, van de Sandt E, den Hollander MA et al (1998)
J Catal 177:29
20. Rupprechter G, Somorjai GA (1997) Catal Lett 48:17
21. Ribeiro FH, Gerken CA, Somorjai GA et al (1997) Catal Lett
45:149
Acknowledgements This work is financially supported by Shanghai
3F New Materials CO., Ltd. and Shanghai Huayi (Group) Company.
22. Hoke JB, Gramiccioni GA, Balko EN (1992) Appl Catal B 1:285
23. Forni P, Prati L, Rossi M (1997) Appl Catal B 14:49
24. Boehm HP (1994) Carbon 32:759
References
1. Murena F, Famiglietti V, Gioia F (1993) Environ Prog 12:231
2. Aramendia MA, Borau V, Garcia IM et al (2002) J Mol Catal A:
Chem 184:237
3. Concibido NC, Okuda T, Nishijima W et al (2007) Appl Catal B
71:64
4. Coq B, Ferrat G, Figueras F (1986) J Catal 101:434
5. Kulkarni PP, Deshmukh SS, Kovalchuk VI et al (1999) Catal Lett
61:161
6. Srinivas ST, Jhansi Lakshmi L, Lingaiah N et al (1996) Appl
Catal A 135:201
7. Barrabes N, Cornado D, Foettinger K et al (2009) J Catal 263:239
25. Boehm HP (2002) Carbon 40:145
26. Alonso F, Beletskaya IP, Yus M (2002) Chem Rev 102:4009
27. Figueiredo JL, Pereira MFR, Freitas MMA et al (1999) Carbon
37:1379
28. Noh JS, Schwarz JA (1990) Carbon 28:675
29. Moreno-castilla C, Carrasco-marin F, Maldonado-hodar FJ et al
(1998) Carbon 36:145
30. Xia C, Liu Y, Xu J et al (2009) Catal Commun 10:456
31. Ukisu Y, Kameoka S, Miyadera T (2000) Appl Catal B 27:97
32. Yuan G, Keane MA (2003) Catal Today 88:27
123