The influence of acid treatment of carbon nanofibers on the activity of palladium catalysts in the liquid-phase hydrodechlorination of dichlorobenzene

Article abstract of DOI:10.1016/j.apcata.2013.07.046

Palladium catalysts on the basis of acid treated carbon nanofibers (CNF) were studied. Using physico-chemical methods of investigation (low-temperature nitrogen adsorption-desorption, FTIR spectroscopy and Boehm titration) it has been shown that under the action of HCl and HNO₃ there take place a reduction in the specific surface area and the volume of pores and an increase of the concentration of surface functional groups. This has a substantial influence on the adsorptive properties of CNF, the dispersion of palladium and the catalytic activity in the liquid-phase hydrodechlorination of 1,2-dichlorobenzene. The reaction rate was observed to be the highest in the presence of Pd/CNF-HCl. Treatment of CNF by HCl gives the highest dispersion of palladium on the external surface of CNF. There is also an increase in the adsorption capacity for 1,2-dichlorobenzene and a decreased adsorption of the solvents on the surface of the acid-modified CNF.

Full text of DOI:10.1016/j.apcata.2013.07.046

Applied Catalysis A: General 467 (2013) 386–393
Contents lists available at ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
The influence of acid treatment of carbon nanofibers on the activity of
palladium catalysts in the liquid-phase hydrodechlorination of
dichlorobenzene
O.V. Netskinaa,∗, O.V. Komovaa, E.S. Taybana, G.V. Oderovaa, S.A. Mukhaa,
G.G. Kuvshinovb, V.I. Simaginaa,b
a Laboratory of hydrides investigation, Boreskov Institute of Catalysis SB RAS, Pr. Akademika Lavrentieva 5, Novosibirsk 630090, Russia
b Department of Chemical Engineering, Novosibirsk State Technical University, Pr. Karla Marksa 20, Novosibirsk 630092, Russia
a r t i c l e i n f o
a b s t r a c t
Article history:
Palladium catalysts on the basis of acid treated carbon nanofibers (CNF) were studied. Using physico-
chemical methods of investigation (low-temperature nitrogen adsorption–desorption, FTIR spectroscopy
and Boehm titration) it has been shown that under the action of HCl and HNO3 there take place a reduction
in the specific surface area and the volume of pores and an increase of the concentration of surface
functional groups. This has a substantial influence on the adsorptive properties of CNF, the dispersion
of palladium and the catalytic activity in the liquid-phase hydrodechlorination of 1,2-dichlorobenzene.
The reaction rate was observed to be the highest in the presence of Pd/CNF-HCl. Treatment of CNF by HCl
gives the highest dispersion of palladium on the external surface of CNF. There is also an increase in the
adsorption capacity for 1,2-dichlorobenzene and a decreased adsorption of the solvents on the surface
of the acid-modified CNF.
Received 19 February 2013
Received in revised form 19 July 2013
Accepted 22 July 2013
Available online xxx
Keywords:
Carbon nanofibers
Acid treatment
Surface function groups
Adsorption
© 2013 Elsevier B.V. All rights reserved.
Hydrodechlorination
Palladium catalyst
1. Introduction
Inorganic acid treatment is generally considered to be a means of
controlling the adsorptive properties of carbon materials by varying
Carbon materials provide optimal supports of catalysts for the
low-temperature hydrogenation of organic molecules, including
the liquid-phase hydrodechlorination of chlorinated organic com-
pounds [1–3]. These materials are characterized by a high adsorp-
tion capacity for organochlorine compounds [4] and are resistant to
the action of hydrochloric acid that forms in the course of the hydro-
dechlorination [5]. However, when choosing a carbon material for
a support application its porosity should be taken into considera-
tion since the presence of the chlorine atom in the benzene ring
makes impossible chlorobenzene adsorption in pores smaller than
0.36 nm [6,7]. Therefore, in such applications carbon materials with
a minimum volume of micropores should preferably be used.
Accessibility of the nanofibrous carbon surface to large organic
molecules and almost absolute absence of micropores make it
attractive catalyst support for hydrodechlorination of chloroaro-
matic compounds. [8,9]. By varying the synthesis conditions the
nanofibrous materials supports with different characteristics can
be obtained [10,11]. The support properties can also be modified
by a treatment in an oxidizing or a reducing medium. [12–15].
the concentration and nature of the oxygen-containing groups on
their surface [16–21]. In spite of the fact that an extensive literature
is available on the adsorption properties of oxidized carbons, the
studies of support acid treatment effects on Pd catalysts activity
are not numerous [22–28]. It has been suggested [25,29,30], that
acid treatment of activated carbon not only has an influence on the
adsorption properties of the final catalyst but also on the dispersion
of metal active component.
The purpose of this work was to investigate the activity
of Pd catalyst in the liquid-phase hydrodechlorination of 1,2-
dichlorobenzene (1,2-DCB) by molecular hydrogen depending
on the physicochemical properties of carbon nanofibers (CNF)
supports subjected to a treatment by nitric (CNF-HNO3) and
hydrochloric (CNF-HCl) acid. The choice of CNF as the object of
investigation was determined by the absence in them of a devel-
oped structure of micropores which is an essential requirement for
catalytic reactions involving large organic molecules [31].
2. Experimental
2.1. Material
∗
The CNF samples (Vꢀpore = 0.32 cm3/g; SBET = 151 m2/g) were
prepared by the pyrolysis of a (C1–C3) fraction over a nickel catalyst
Corresponding author. Tel.: +7 383 330 74 58; fax: +7 383 330 76 91.
E-mail address: netskina@catalysis.ru (O.V. Netskina).
0926-860X/$ – see front matter © 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.apcata.2013.07.046

O.V. Netskina et al. / Applied Catalysis A: General 467 (2013) 386–393
387
in a quartz flow microreactor at 550 ◦C [10]. A nickel catalyst was
prepared by coprecipitation of hydroxide mixtures from nickel and
aluminum salt solutions. Sediment was filtered, washed and dried
at 120 ◦C and then decomposed in nitrogen at 350 ◦C. The sample
was reduced with pure hydrogen at 550 ◦C for 3 h. According to the
chemical analysis the nickel content was 90 3 wt.%. Gas (C1–C3)
flow rate was set to meet the following conditions: 110 L of gas per
1 h per 1 g of Ni in the catalyst.
up of two immiscible fluids. In some experiments a organic phase
consisted from 7 cm3 of toluene, 4 cm3 of 2-propanol and 2.5 mL of
1,2-DCB.
The products of hydrodechlorination in the organic phase were
analyzed on a “Crystall-2000” chromatograph (JSC KUPOL, Russia)
with a flame-ionization detector. The instrument was operated
under an Ar carrier gas, using a hydrogen flame, in the temper-
ature range from 60 to 120 ◦C (heat rate is 10 ◦C/min), a column
length of 2 m and column diameter of 2.5 mm. The filling sorbent
was Chromaton N-AW.
The CNF fraction of 0.04–0.08 mm was used for further experi-
ments.
Hydrodechlorination performance is quantified in terms of
dechlorination degree (XCl):
2.2. Acid treatment of CNF
Treatment of CNF by acids was performed at room temperature
using solutions containing 10% (w/w) of the corresponding (HCl
and HNO3) acid. 5 g of CNF was placed into a 50 mL flask and 25 mL
of acid solution were added. The mixture was agitated with a mag-
netic stirrer for 20 h after which it was quickly warmed up to 80 ◦C
and kept at this temperature for an hour. The acid-treated CNFs
were repeatedly washed with distilled water. The acid treatment
procedure was repeated three times for each CNF sample and the
treated samples were dried under vacuum for two hours at 75 ◦C.
[1, 2-DCB]0 − [1, 2-DCB] − 0.5[CB]
XCl
=
(1)
[1, 2-DCB]0
where [1,2-DCB]0 is the initial concentration of 1,2-DCB; [1,2-DCB]
is the value of concentration of 1,2-DCB at reaction time t and [CB]
is the concentration of chlorobenzene (CB) at reaction time t.
2.6. Investigation methods
2.3. Study of the adsorption of 1,2-DCB, toluene and 2-propanol
from organic phase of reaction mixture
The specific surface area and pore volume were based on N2
adsorption isotherms, determined with a Micrometritics ASAP
2000 surface analyzer.
The content of impurities in CNF and of palladium in cata-
lysts was determined by X-ray fluorescence analysis on a VRA-30
instrument with a Cr anode of the X-ray tube. The relative error of
determination was 5%.
The transmission IR spectra were taken in air at ambient tem-
perature on a Bomem MB-102 Fourier spectrometer. For these
measurements 1.5 mg of a catalyst sample was mixed with 800 mg
of CsJ and pressed to a pellet.
The content of acidic functional groups on CNF surface was
investigated by the acid–base titration method proposed by Boehm
using NaHCO3, Na2CO3, NaOH or sodium ethoxide solutions
[34,35]. One gram of the carbon and 70 mL of a 0.05 N solution of
bases were mixed in a conical flask and agitated at 100 rpm for 24 h
at 298 K under argon. Then a 18 mL aliquot of this solution was acid-
ified by the addition of 18 mL of HCl (0.05 N) and the sample was
back-titrated with NaOH (0.05 N). Adsorption from 0.1 N NaOEt in
ethanol was determined analogously. The points of neutralization
were detected by potentiometric titration using an ATP-02 auto-
matic titrator (Akvilon, Russia). Each sample was analyzed three
times and the average value was employed for further analysis.
Temperature-programmed desorption (TPD) is used to measure
the surface acidic groups of the CNF-HCl and CNF-HNO3. It was
carried out in a tabular quartz reactor of 3.00 mm in diameter and
100 mm in length, which coupled to quadrupolar mass spectrome-
ter (NETZSCH QMS 403 C Aëolos). The sample (60 mg) was heated in
the reactor from room temperature to 1000 ◦C at rate of 10 ◦C/min
under a flow of argon (40 cm3/min). CO and CO2 desorbed from the
sample were monitored by a mass spectrometer at an m/e of 28 for
CO and 44 for CO2.
The adsorption of 1,2-dichlorobenzene (1,2-DCB), toluene and
2-propanol from the organic phase of the reaction medium was
studied by the static method at room temperature using the
following concentrations: C1,2-DCB = 0.2 mol/L, Ctoluene = 5.87 mol/L
and C2-propanol = 4.65 mol/L. 1 g sample of the original or acid-
treated carbon nanofibers was placed into a 60 mL flask and then
50 mL of reaction mixture (1,2-DCB, toluene or 2-propanol) was
added. Changes in concentrations of the organic compounds dur-
ing the adsorption cycle were monitored using a GC–MS instrument
Agilent 7000 Triple Quadrupole GC/MS, column Agilent 19091S-
433:HP-5MS 5% Phenyl Methyl Silox. To obtain reliable data each
experiment was repeated for at least three times. The differences
between the experimental values did not exceed 3%.
2.4. Catalyst preparation
Supported palladium catalysts were prepared by incipient
wetness impregnation from hydrochloric solutions of palladium
chloride (PdCl2). The catalysts were dried at temperatures from
130 to 150 ◦C for two hours. The reduction of the catalyst was
performed at room temperature by adding an aqueous solution of
sodium borohydride (the Pd:NaBH4 molar ratio was 1:3) [32]. The
palladium loading of the catalyst was 1 0.05 wt.% according to the
chemical analysis.
2.5. Experimental procedure
The apparatus for the liquid-phase hydrodechlorination of
1,2-DCB consisted of two main parts: a rector unit and a volu-
metric block. The reaction was carried out under constant stirring
(1200 rpm), at 50 ◦C and constant pressure of hydrogen of 0.1 MPa.
At stirring rates of higher than 800 rpm the influence of the external
diffusion processes was shown to be negligible [32]. To obtain reli-
able data each experiment was repeated for at least three times.
The differences between the experimental values did not exceed
4%.
The high-resolution transmission electron microscopy (HR
TEM) studies of the catalysts were carried out using a JEM-2010
˚
(JEOL, Japan) instrument with a lattice resolution of 1.4 A and an
accelerating voltage of 200 kV. The local elemental analysis was
performed by the EDX method on an Energy-dispersive X-ray
Phoenix Spectrometer equipped with a Si (Li) detector with an
energy resolution of at least 130 eV or higher. The samples were
fixed on “holey” carbon films supported on copper grids and exam-
ined with an electron microscope. To construct the particle size
distribution diagrams the diameter of at least 300 particles has been
To 11 cm3 of a solution of 1,2-DCB in an organic phase contain-
ing of 7 cm3 of toluene and 4 cm3 of 2-propanol were added 4 cm3
of a 50% (w/w) solution of potassium hydroxide in water to neutral-
ize the forming HCl [32,33]. Thus, the reaction medium was made

388
O.V. Netskina et al. / Applied Catalysis A: General 467 (2013) 386–393
determined. The mean Pd particle sizes are quoted [36] as both a
¯
number average diameter (dn),
ꢀ
inidi
ꢀ
dn
=
.
(2)
(3)
ini
¯
and surface area-weighted diameter (ds),
ꢀ
inidi3
ꢀ
ds =
,
inidi2
where ni is the number of particles of diameter di.
3. Results
As has been described in Section 1, treatment of carbon materi-
als by acids has a considerable influence on their physicochemical
properties. The resulting changes in the nature of surface functional
groups, the structure of surface layers, porosity and specific sur-
face area and the removal of impurities may have an influence on
the catalytic properties of Pd supported on an oxidized carbon. A
study of acid-treated carbon nanofibers has been performed and
the obtained data were compared with those for an untreated CNF.
Fig. 1. Nitrogen adsorption–desorption isotherm at 77 K (A) and pore size distribu-
tion (B) for the initial CNF, CNF-HCl and CNF-HNO3.
3.1. Characterization of acid-treated carbon nanofibers
3.1.1. Chemical composition
According to X-ray fluorescence data (Table 1) treatment of CNF
by hydrochloric and nitric acids decreases the amount of impuri-
ties in the samples. In the case of CNF-HNO3 the content of nickel
was reduced by a factor of 5 as compared with the untreated CNF.
In the case of CNF-HCl the amount of nickel was reduced only by
a factor of 2.5. Carbon materials are capable of a strong adsorption
of nitrate and chloride ions. X-ray fluorescence analysis confirmed
an enhanced content of chloride ions in CNF-HCl. The nitrogen
impurities could not be detected because of the limitations of this
technique.
Fig. 2. FTIR-spectra CNF, CNF-HCl and CNF-HNO3.
3.1.2. Textural characteristics
of the specific surface area in the CNF samples kept in 10% (w/w)
solutions of acids. Thus, for CNF, CNF-HCl and CNF-HNO3 the val-
ues of the specific surface area are equal to 151, 143 and 145 m2/g,
respectively. An analogous decrease in the specific surface area of
activated carbon upon treatment by nitric and hydrochloric acids
has been observed in [38,39].
In Fig. 1 are shown the low-temperature nitrogen adsorption-
desorption isotherms and pore size distribution for CNF, CNF-HCl
and CNF-HNO3. It can be seen that they are practically identical
to each other and are indicative of a mesoporous structure of the
samples under study. The irreversible capillary condensation in
mesopores is supported by the presence of a hysteresis loop at
P/P0 > 0.4 (the desorption curve runs above the adsorption curve).
An analysis of the shapes of the observed hysteresis loops using
the IUPAC classification allows their assignment the H1 type, a
type characteristic of uniformly packed globules having close vol-
umes. [37]. For all of the CNF samples the determined volume of
micropores does not exceed 1–2% of the total pore volume (Table 1).
Fig. 1B shows the pore size distribution in the regions of meso-
3.1.3. Surface group determination
To determine the composition of the surface functional groups
in the CNF, CNF-HCl and CNF-HNO3 samples were studied by FTIR
spectroscopy, Boehm titration and TPD.
The IR spectra of the CNF, CNF-HCl and CNF-HNO3 samples are
shown in Fig. 2. It is seen that there are no substantial differences
between the spectra of CNF and CNF-HCl but in the spectrum of
CNF-HNO3 there appear new lines at 1720, 1620 (sh.), 1580 and
1360 cm−1. There is an increase in the amount of the adsorbed
water as indicated by increased relative intensities of the absorp-
tion bands at 3450 and 1620 (sh.) cm−1 due to the vibrations of
water molecules [40]. The absorption band at 1720 cm−1 indicates
the presence on the sample surface of carbonyl (C O) and carboxyl
˚
˚
(from 20 to 500 A) and macro- (>500 A) pores. It can be seen that
treatment by acids did not lead to any essential changes in the
porous structure of CNF. All the samples show the typical bimodal
˚
˚
pore size distribution-peaks at 145 A and 390 A. The total pore vol-
ume is equal to 0.32, 0.28 and 0.27 cm3/g for CNF, CNF-HCl and
CNF-HNO3, respectively. This is consistent with the reduced values
Table 1
Physicochemical properties of CNF before and after treatment by hydrochloric and nitric acids.
Sample
Contents of impurities (wt.%)
Pore volume (cm3/g)
Micropores
Ni
Fe
0.014
Cl
Al
S
Mesopores
Macropores
CNF
CNF-HCl
CNF-HNO3
0.94
0.39
0.18
–
0.11
–
0.08
–
–
0.01
–
0.003
0.002
0.006
0.242
0.218
0.211
0.074
0.060
0.053
<0.01
<0.01
<0.01

O.V. Netskina et al. / Applied Catalysis A: General 467 (2013) 386–393
389
Fig. 3. CO2 (A) and CO (B) releases during the TPD of the CNF-HCl and CNF-HNO3 samples.
Table 2
(Table 3). This is in agreement with the data obtained by the Boehm
method.
Acidic surface oxides of CNF, CNF-HCl and CNF-HNO3 (according to the Boehm’s
titration method).
Sample
Titration using (␮mol/g)
3.2. Study of the adsorption of 1,2-DCB, toluene and 2-propanol
from organic phase of reaction mixture
NaOEt
112
Na2CO3
NaHCO3
NaOH
CNF
CNF-HCl
CNF-HNO3
9
22
42
5
7
8
12
26
3
5
9
69
82
233 12
6
8
179 11
294
Changes of the physicochemical properties of carbon materials
will certainly have an influence on the adsorption of the reac-
tion mixture components in the liquid-phase catalytic reactions.
Adsorption of 1,2-DCB, toluene and 2-propanol from the organic
phase of the reaction medium was used to compare the adsorption
properties of studied CNFs depending on the applied acid treat-
ment. The changes in the composition of the reaction medium were
determined by gas chromatography–mass spectrometry. Accord-
ing to the obtained results (Table 4), treatment by hydrochloric
acid increases the adsorption of 1,2-DCB on CNF-HCl and decreases
the adsorption of 2-propanol, especially in the early stages of the
process (the adsorption equilibrium of 1,2-DCB on studied CNFs
is attained within the first 5 min). The adsorption of 1,2-DCB and
toluene on CNF-HNO3 was decreased.
8
165
126
(
COOH) functional groups [41]. The absorption band at 1580 cm−1
is assignable to the stretching vibrations of the aromatic ring [42].
The absorption band at 1360 cm−1 appears to be due to NO3− vibra-
tions [43,44] indicating that there takes place adsorption of nitrate
ions. This is in good agreement with the results of a FTIR study of
oxidized carbon nanofibers [45].
The concentration of acidic sites on the carbon surface after
its acid-treatments was studied by Boehm’s method [34,35]. It is
assumed that sodium bicarbonate (pKa = 6.7) neutralizes carboxylic
acids, sodium carbonate (pKa = 10.25) neutralizes carboxylic acids
and lactones, sodium hydroxide (pKa = 15.74) neutralizes car-
boxylic acids, lactones, and phenols, whereas sodium ethoxide
(pKa = 20.58) reacts with all oxygen species, even extremely weak
acids [12,46]. According to the Boehm titration data (Table 2)
the content of surface functional groups in the acid-treated CNF
was increased. The carbon surface was enriched by weakly acidic
oxygen-containing groups (carbonyl, phenolic, lactone or lactol
groups) under action of hydrochloric acid. In case of CNF treatment
by nitric acid the concentration of strongly acidic groups such as
carboxylic acids and phenolic groups was increased several times
but the amount of weakly acidic oxygen-containing groups changes
little.
Thus, it has been established that changes in the physicochemi-
cal properties of acid-treated CNFs may have a substantial influence
on the adsorption process of 1,2-DCB in the cause of the catalytic
conversion, especially in the case of CNF-HCl-based catalysts.
3.3. Study of Pd catalysts based on untreated CNF and
acid-treated CNF
Palladium catalysts on the basis of CNF, CNF-HCl, CNF-HNO3
have been prepared. After their reduction by sodium borohydride
at room temperature the particle size of the active component was
studied by HR TEM. It can be seen from Fig. 4 that the particle size
of Pd does not exceed 10 nm. A detailed examination of the particle
size distribution of Pd has shown the metal dispersion to decrease in
the order Pd/CNF-HCl < Pd/CNF < Pd/CNF-HNO3. In the Pd/CNF cat-
alyst Pd particles were present not only on the CNF surface but
also inside the channels (Fig. 5). This makes them inaccessible to
the large molecules of the reagents. Thus, pretreatment of CNF by
The difference in the surface chemistry between CNF-HCl and
CNF-HNO3 is most clearly demonstrated by TPD. The CO and CO2
profiles for the oxidized CNFs are shown in Fig. 3. An analysis
of this data shows that CNF treated by HCl mainly contain car-
bonyl and lactone groups and on the surface of the CNF-HNO3 the
prevailing species are carboxyl groups and carboxylic anhydrides
Table 3
Acidic surface oxides of CNF, CNF-HCl and CNF-HNO3 (according to TPD method).
Surface functional groups (␮mol/g)
Carboxylic anhydrides
Carboxyl acid groups
Lactone groups
Carboxylic anhydrides
Phenol groups
Carbonyl groups
CNF-HCl
CNF-HNO3
23
86
–
35
34
28
Negligible
76
112
597
482
340

390
O.V. Netskina et al. / Applied Catalysis A: General 467 (2013) 386–393
Table 4
Adsorption of 1,2-DCB, toluene and 2-propanol from organic phase of reaction medium in the presence of the CNF, CNF-HCl and CNF-HNO3 samples.
Sample
Concentration (mol/L)
1,2-DCB
Toluene
Initial
2-Propanol
Initial
Initial
5 mina
60 minb
5 mina
60 minb
5 mina
60 minb
CNF
CNF-HCl
CNF-HNO3
0.20
0.20
0.20
0.28
0.18
0.24
0.28
0.18
0.23
5.87
5.87
5.87
5.80
6.14
5.91
5.87
6.18
5.97
4.65
4.65
4.65
2.60
3.91
3.26
2.34
3.67
3.35
a
The beginning of process
The adsorption reaches equilibrium in 15 5 min.
b
Fig. 4. TEM image of reduced Pd/CNF (A), Pd/CNF-HCl (B), Pd/CNF-HNO3 (C) catalysts and Pd nanoparticle size distribution (D).
mineral acids ensues that Pd is present only on the surface of CNF.
So, the changes in the surface chemistry of carbon materials have a
substantial influence not only on the adsorption properties of the
materials but also on the particle size of the active component as
reported for activated carbons [27,47].
3.4. Catalytic hydrodechlorination of 1,2-DCB
The 1,2-DCB hydrodechlorination was carried out in a medium
consisting of two immiscible liquid phases. The reactor contains a
50% (w/w) aqueous solution of KOН and an organic phase consisting
Fig. 5. TEM image (a) and EDX spectrum (b) of palladium particle in a CNF channel of Pd/CNF. Dash line indicates EDX analysis region.

O.V. Netskina et al. / Applied Catalysis A: General 467 (2013) 386–393
391
Fig. 6. The reaction medium of catalytic hydrodechlorination of 1,2-DBC.
Fig. 8. The influence of acid treatment of the CNF on the activity of palladium cat-
alysts in the liquid-phase hydrodechlorination of 1,2-DCB. Reaction medium: 4 mL
2-propanol, 7 mL toluene and 4 mL 50% (w/w) aqueous solution of KOH. The 1,2-DCB
concentration was 0.2 mol/L. Molar ratio of Pd:C–Cl = 1:100. Temperature was 50 ◦C.
Pressure of hydrogen was 0.1 MPa. Stirring rate was 1200 rpm.
of toluene, 2-propanol and 1,2-DCB. The presence of an alkali is
necessary to bind the hydrogen chloride which forms in the
course of the hydrodechlorination to prevent deactivation of the
catalyst and possibly shifting the reaction equilibrium toward
the formation of the products (Fig. 6). In the alkali-free organic
medium the rate of the reaction was higher but a 100% conversion
of 1,2-DCB could not be achieved. With the addition of a further
portion of 1,2-DCB to the alkali-free reaction medium no hydro-
dechlorination was observed. The full loss in activity of the Pd/CNF
was caused by the palladium being washed away into the reaction
medium: 1.02 wt.% Pd before the reaction and 0.18 wt.% Pd after
the reaction. In the reaction in the two-phase medium the decrease
in the catalyst’s content of palladium was inconsiderable from
1.04 wt.% to 0.93 wt.%. Therefore, even for a Pd:C–Cl molar ratio of
1:1000 the full conversion of 1,2-DCB reached after 4.5 h (Fig. 6).
It should be noted that in this reaction medium the catalyst
based on untreated CNF was usually at the interface of the two
phases as shown in Fig. 7. Such a behavior was also observed for
the Pd/CNF-HCl catalyst but the granules of Pd/CNF-HNO3 catalyst
were located exclusively in the aqueous phase. It was not possible
to make Pd/CNF-HNO3 go to the organic phase even with an intense
stirring of the reaction medium. Therefore for these experimental
conditions treatment by nitric acid has a negative effect on the
hydrodechlorination rate of 1,2-DCB (Fig. 8). Treatment of CNF in
a 10% (w/w) solution of hydrochloric acid increases the activity
of the palladium system and the 1,2-DCB dechlorination was fully
complete within 30 min.
Thus, our study of the hydrodechlorination of 1,2-DCB over
Pd/CNF, Pd/CNF-HCl and Pd/CNF-HNO3 catalysts has shown that
treatment of nanofibrous carbons by inorganic acids has an influ-
ence on their activity (Fig. 8).
4. Discussion
Physicochemical properties of nanofibrous carbons have been
studied before and after their treatment by nitric and hydrochloric
acids. According to the chemical analysis data (Table 1) treatment
by inorganic acids leads to the removal of nickel from the samples
and the surface becomes enriched with oxygen-containing groups
as is seen from Boehm titration (Table 2), TPD (Table 3) and IR spec-
troscopy data (Fig. 2). It can be suggested that functionalization of
the pores surfaces leads to their narrowing and is the main rea-
son for the decreased pore volume of the samples (Fig. 1) as was
described in [48]. A second reason may be the blocking of channels
by anions of the acids [49]. The presence of chlorine and nitrate-ions
in the acid-treated CNFs was confirmed by the chemical analysis
(Table 1) and IR spectroscopy (Fig. 2). It appears that these chloride-
and nitrate-ions may fill the channels of the CNF and thus prevents
active component precursor penetration into these channels at the
stage of the catalyst synthesis. This explains the absence of palla-
dium particles in the channels of the acid-treated CNFs (Fig. 4B and
C).
The enrichment of CNF surfaces with oxygen-containing func-
tional groups may be the reason for the differing dispersions of
the active component in the Pd/CNF, Pd/CNF-HCl, Pd/CNF-HNO3
catalysts (Fig. 4D). Thus, in Refs. [23,27] the particle size of palla-
dium on activated carbon surfaces has been shown to be dependent
on the content and nature of oxygen-containing functional groups,
with the dispersion of the active component being strongly influ-
enced by the carbon treatment conditions. [28]. According to TEM
and chemisorption data [23] the dispersion of Pd increases as the
surface becomes richer in weakly acidic oxygen-containing groups
(phenolic, lactonic and carbonyl groups). At the same time, predom-
ination on activated carbon surfaces of the more acidic phenolic
and carboxylic groups, as in the case of 5.6 M HNO3 treatment,
shifts the particle size distribution to larger sizes. In our work, it
Fig. 7. The liquid-phase 1,2-DCB hydrodechlorination in different reaction medium:
(ꢀ) the organic phase; (ꢁ) the organic phase and 4 mL 50% (w/w) solution of KOH.
The organic phase: 4 mL 2-propanol, 7 mL toluene and 2.5 mL 1,2-DCB. Molar ratio of
Pd:C–Cl = 1:1000. Temperature was 50 ◦C. Pressure of hydrogen was 0.1 MPa. Stirring
rate was 1200 rpm.

392
O.V. Netskina et al. / Applied Catalysis A: General 467 (2013) 386–393
The somewhat lower activity of the Pd/CNF system as compared
with Pd/CNF-HCl may be explained by the fact that some of the pal-
ladium particles become blocked in the channels of the CNF, which
makes them inaccessible to the reagents. At the same time, the
low activity of the Pd/CNF-HNO3 catalysts is determined not by the
particle size of palladium but is the result of the hydrophilic char-
acter of its surface caused by a high concentration of acidic surface
groups. As for palladium catalysts based on oxide supports Al2O3
and SiO2 [3,33,52,53] hydrophobic organic compounds interact
weakly with the hydrophilic surface of CNF-HNO3 (Tables 2 and 3).
Therefore, the CNF-HNO3 and the Pd/CNF-HNO3 catalyst mainly
remain in the aqueous phase, as in the case of H2O/hexane mixture
[45]. The Pd/CNF-HNO3 catalyst does not go to the organic phase
even after an intense stirring. This makes the Pd/CNF-HNO3 catalyst
inaccessible for the 1,2-DCB molecules and the hydrodechlorinat-
ion in the water–organic liquid medium does not take place (Fig. 8).
5. Conclusion
Physicochemical properties of CNF before and after their
treatment by hydrochloric and nitric acid have been studied
by chemical analysis, low-temperature adsorption–desorption of
nitrogen, FTIR spectroscopy, Boehm titration and temperature-
programmed desorption. Summing up the results for CNF, CNF-HCl,
and CNF-HNO3 obtained by different techniques it can be concluded
that treatment of carbon nanofibers by inorganic acids leads to
changes in (I) the content of impurities, (II) the concentration of
the surface functional groups and (III) their textural characteristics
(porosity and specific surface area).
Palladium catalysts based on CNF, CNF-HCl and CNF-HNO3 have
been prepared by incipient wetness impregnation from hydrochlo-
ric solutions of palladium chloride – H2[PdCl4]. It was shown that
changes of the physicochemical properties of CNF occurring under
the action of acids have a substantial influence on palladium activity
in the liquid-phase hydrodechlorination of 1,2-dichlorobenzene. It
was shown that palladium particles dispersion and their location,
acidity of the CNF surface and its adsorption properties are the main
factors determinated the efficiency of applied palladium catalyst.
So, pretreatment of CNF by a hydrochloric acid leads to a higher
dispersion of the active component and an increased rate of the
1,2-DCB adsorption from the reaction medium as compared with
untreated Pd/CNF catalyst. This makes Pd/CNF-HCl the most active
catalyst among Pd/CNF, Pd/CNF-HCl and Pd/CNF-HNO3.
Fig. 9. Schematic representation of interactions between the oxygen surface groups
and [PdCl4]2− anion.
was found that the more acidic groups are present on the CNF
surface the larger the size of the particles of the active compo-
nent are formed. Thus, the Pd/CNF-HNO3 catalyst shows the lowest
values of Pt dispersion (Fig. 4D) and the highest concentration of
strongly acidic surface groups (Tables 2 and 3) whose IR adsorp-
tion bands are observed in the regions 1730–1720, 1670–1570 and
1400–1340 cm−1 (Fig. 2). At the same time, the predominance of
weakly acidic oxygen-containing groups (Tables 2 and 3) on the
CNF-HCl surfaces leads to a higher dispersion of Pd in the Pd/CNF-
HCl catalyst (Fig. 4D). The influence of surface functional groups
nature on the active component dispersity may be explained by the
electrostatic interaction between oxygen-containing groups and
palladium anion ([PdCl4]2−) at the impregnation step, as described
for platinum and palladium catalysts based on activated carbon
[2,24,25]. When carbon surface has been enriched with strongly
acidic oxygen-containing groups (acidic phenolic, carboxylic and
nitro groups), electrostatic repulsion of [PdCl4]2− anions from
CNF-HNO3 surface takes place. This leads to the formation of big-
ger palladium particles (Fig. 9). Enrichment of CNF-HCl surface
by lactone and carbonyl groups increases electrostatic attraction
[PdCl4]2− anions with surface (Fig. 9). These groups act as anchor-
ing centers hinter agglomeration and surface diffusion of the active
component.
Acknowledgements
This work was supported by the Russian Foundation for Basic
Research (grant nos. 10-03-00372 and 12-03-31517). The authors
are thankful to A.V. Ischenko for the HR TEM study.
A detailed study of both the initial supports CNF, CNF-HCl,
CNF-HNO3 supports and the Pd/CNF, Pd/CNF-HCl, Pd/CNF-HNO3
catalysts by physicochemical methods may help find reasons
for their differing catalytic activities in the liquid-phase 1,2-DCB
hydrodechlorination by molecular hydrogen in a water–organic
liquid medium. Thus, the high activity of the Pd/CNF-HCl cata-
lyst in the 1,2-DCB hydrodechlorination is, first of all, accounted
for by a more highly dispersed state of the catalyst whose particle
are located only on CNF-HCl external surfaces. Moreover, the high
activity of the Pd/CNF-HCl catalysts in the liquid-phase hydrode-
chlorination of 1,2-DCB can be related the increased rate of the
1,2-DCB adsorption from the reaction medium and the decreased
adsorption of 2-propanol (Table 4). This demonstrates an important
role of the surface functional groups on the carbon in the adsorption
processes during catalytic reactions carried out in a liquid phase
as described for gas phase dechlorination of chlorobenzene and
1,3-dichlorobenzene over Pd and Ni [50,51].
References
[1] E. Auer, A. Freund, J. Pietsch, T. Tacke, Appl. Catal., A 173 (1998) 259–271.
[2] F. Rodríguez-Reinoso, Carbon 36 (1998) 159–175.
[3] F.J. Urbano, J.M. Marinas, J. Mol. Catal. A 173 (2001) 329–345.
[4] E.G. Furuya, H.T. Chang, Y. Miura, K.E. Noll, Sep. Purif. Technol. 11 (1997) 69–78.
[5] Z.M. de Pedro, E. Diaz, A.F. Mohedano, J.A. Casas, J.J. Rodriguez, Appl. Catal. B
103 (2011) 128–135.
[6] V.B. Fenelonov, M.S. Melgunov, N.A. Baronskaya, React. Kinet. Catal. Lett. 63
(1998) 305–312.
[7] V.B. Fenelonov, V.A. Likholobov, A.Yu Derevyankin, M.S. Mel’gunov, Catal.
Today 42 (1998) 341–345.
[8] V.V. Shinkarev, V.B. Fenelonov, G.G. Kuvshinov, Carbon 41 (2003) 295–302.
[9] P. Serp, M. Corrias, P. Kalck, Appl. Catal., A 253 (2003) 337–358.
[10] G.G. Kuvshinov, YuI. Mogilnykh, D.G. Kuvshinov, D.Yu. Yermakov, M.A. Yer-
makova, A.N. Salanov, N.A. Rudina, Carbon 37 (1999) 1239–1246.
[11] T.V. Reshetenko, L.B. Avdeeva, Z.R. Ismagilov, V.V. Pushkarev, S.V. Cherepanova,
A.L. Chuvilin, V.A. Likholobov, Carbon 41 (2003) 1605–1615.
[12] H.P. Boehm, Carbon 32 (1994) 759–769.

O.V. Netskina et al. / Applied Catalysis A: General 467 (2013) 386–393
393
[13] P. Albers, K. Deller, B.M. Despeyroux, A. Schäfer, K. Seibold, J. Catal. 133 (1992)
467–478.
[14] J.L. Figueiredo, M.F.R. Pereira, M.M.A. Freitas, J.J.M. Órfão, Carbon 37 (1999)
1379–1389.
[15] W. Shen, Z. Li, Y. Liu, Recent Patents Chem. Eng. 1 (2008) 27–40.
[16] A. Derylo-Marczewska, B. Buczek, A. Swiatkowski, Appl. Surf. Sci. 257 (2011)
9466–9472.
[33] G. Yuan, M.A. Keane, Appl. Catal. B 52 (2004) 301–314.
[34] S.L. Goertzen, K.D. Thériault, A.M. Oickle, A.C. Tarasuk, H.A. Andreas, Carbon 48
(2010) 1252–1261.
[35] A.M. Oickle, S.L. Goertzen, K.R. Hopper, Y.O. Abdalla, H.A. Andreas, Carbon 48
(2010) 3313–3322.
[36] C. Amorim, G. Yuan, P.M. Patterson, M.A. Keane, J. Catal. 234 (2005)
268–281.
[17] F. Yu, J. Ma, Y. Wu, J. Hazard. Mater. 192 (2011) 1370–1379.
[18] T. Karanfil, J.E. Kilduff, Environ. Sci. Technol. 33 (1999) 3217–3224.
[19] L. Li, P.A. Quinlivan, D.R.U. Knappe, Carbon 40 (2002) 2085–2100.
[20] B. Cabal, T. Budinova, C.O. Ania, B. Tsyntsarski, J.B. Parra, B. Petrova, J. Hazard.
Mater. 161 (2009) 1150–1156.
[37] S. Brunauer, L.S. Deming, W.E. Deming, E. Teller, J. Am. Chem. Soc. 62 (1940)
1723–1732.
[38] G.R. Heal, L.L. Mkayula, Carbon 26 (1988) 803–813.
[39] L.B. Okhlopkova, A.S. Lisitsyn, V.A. Likholobov, M. Gurrath, H.P. Boehm, Appl.
Catal., A 204 (2000) 229–240.
[21] H. Tamon, M. Okazaki, Carbon 34 (1996) 741–746.
[22] D.J. Suh, P. Tae-Jin, I. Son-Ki, Carbon 31 (1993) 427–435.
[23] V.Z. Radkevich, T.L. Senko, K. Wilson, L.M. Grishenko, A.N. Zaderko, V.Y. Diyuk,
Appl. Catal., A 335 (2008) 241–251.
[24] L. Calvo, M.A. Gilarranz, J.A. Casas, A.F. Mohedano, J.J. Rodríguez, J. Hazard.
Mater. 161 (2009) 842–847.
[25] L. Calvo, M.A. Gilarranz, J.A. Casas, A.F. Mohedano, J.J. Rodríguez, Ind. Eng. Chem.
Res. 44 (2005) 6661–6667.
[26] E.J.A.X. van de Sandt, A. Wiersma, M. Makkee, H. van Bekkum, J.A. Moulijn, Appl.
Catal., A 173 (1998) 161–173.
[27] V.I. Simagina, O.V. Netskina, E.S. Tayban, O.V. Komova, E.D. Grayfer, A.V.
Ischenko, E.M. Pazhetnov, Appl. Catal., A 379 (2010) 87–94.
[28] A.L. Dantas Ramos, P. da Silva Alves, D.A.G. Aranda, M. Schmal, Appl. Catal., A
277 (2004) 71–81.
[29] P. Albers, K. Deller, B.M. Despeyroux, G. Prescher, A. Schafer, K. Seibold, J. Catal.
150 (1994) 368–375.
[40] B. Viswanathan, P. Indra Neel, T.K. Varadarajan, Methods of Activation and Spe-
cific Applications of Carbon Materials, National Centre for Catalysis Research,
Indian Institute of Technology, Madras, 2009.
[41] U.J. Kim, C.A. Furtado, X. Liu, G. Chen, P.C. Eklund, J. Am. Chem. Soc. 127 (2005)
15437–15445.
[42] D.B. Mawhinney, V. Naumenko, A. Kuznetsova, J.T. Yates, J. Liu, R.E. Smalley, J.
Am. Chem. Soc. 122 (2000) 2383–2384.
[43] R.A. Marcus, J.M. Fresco, J. Chem. Phys. 27 (1957) 564–568.
[44] D. Forney, W.E. Thompson, M.E. Jacox, J. Chem. Phys. 99 (1993) 7393–7403.
[45] T.G. Ros, A.J. van Dillen, J.W. Geus, D.C. Koningsberger, Chem. Eur. J. 8 (2002)
1151–1162.
[46] T.J. Bandosz, in: P. Serp, J.L. Figueiredo (Eds.), Carbon Materials for Catalysis,
John Wiley & Sons, Inc., Hoboken, New Jersey, 2009, pp. 45–92.
[47] C. Prado-Burguete, A. Linares-Solano, F. Rodriguez-Reinoso, C. Salinas-Martinez
de Lecea, J. Catal. 115 (1989) 98–106.
[48] A. Gil, G. de la Puente, P. Grange, Microporous Mater. 12 (1997) 51–61.
[49] H.P. Boehm, A. Vass, R. Kollmar, Proc. 18th Biennial Conf. on Carbon, July 19–24,
1987, pp. 86–87.
[30] P. Albers, R. Burmeister, K. Seibold, G. Prescher, S.F. Parker, D.K. Ross, J. Catal.
181 (1999) 145–154.
[31] M. Gurrath, T. Kuretzky, H.P. Boehm, L.B. Okhlopkova, A.S. Lisitsyn, V.A.
Likholobov, Carbon 38 (2000) 1241–1255.
[32] A.G. Gentsler, V.I. Simagina, O.V. Netskina, O.V. Komova, S.V. Tsybulya, O.G.
Abrosimov, Kinet. Catal. 48 (2007) 60–66.
[50] S. Jujjuri, M.A. Keane, Chem. Eng. J. 157 (2010) 121–130.
[51] X. Wang, N. Perret, M.A. Keane, Chem. Eng. J. 210 (2012) 103–113.
[52] G. Yuan, M.A. Keane, J. Catal. 225 (2004) 510–522.
[53] W. Wu, J. Xu, R. Ohnishi, Appl. Catal. B 60 (2005) 129–137.