Electrocatalytic properties of the catalysts based on carbon nanofibers with various platinum contents

Article abstract of DOI:10.1007/s11172-011-0165-0

The catalysts on carbon nanofibers with various platinum contents were synthesized. The morphology, resistance to oxidation, and electrochemical behavior of the catalysts in the reactions that occur in fuel cells were studied. The dependence of the specif

Full text of DOI:10.1007/s11172-011-0165-0

Russian Chemical Bulletin, International Edition, Vol. 60, No. 6, pp. 1045—1050, June, 2011
1045
Electrocatalytic properties of the catalysts based on carbon nanofibers
with various platinum contents
E. V. Gerasimova, N. G. Bukun, and Yu. A. Dobrovolsky
Institute of Problems of Chemical Physics, Russian Academy of Sciences,
1 prosp. Akad. Semenova, 142432 Chernogolovka, Moscow Region, Russian Federation.
Fax: +7 (496) 522 5401. Eꢀmail: lyramail@mail.ru
The catalysts on carbon nanofibers with various platinum contents were synthesized. The
morphology, resistance to oxidation, and electrochemical behavior of the catalysts in the reacꢀ
tions that occur in fuel cells were studied. The dependence of the specific output of cathodes of
hydrogen—air fuel cells on the sizes of the platinum clusters was established.
Key words: carbon nanofibers, Pt/C catalysts, fuel cells, size effect.
The problem of development and investigation of the
catalyst for oxygen reduction became topical in the recent
time from both theoretical and practical points of view.
The absence of cheap corrosionproof cathodic catalysts is
a major barrier for commercialization of lowꢀtemperature
fuel cells (FC). For the development of theoretical conꢀ
cepts, it is important to study reasons for the appearance
of the size effect, to determine the particle size optimum
for catalysis, and to study the nature of their interaction
with the support.
Catalysts based on platinum and its alloys are nearly
the only active catalytic systems for oxygen electroreducꢀ
tion at low temperatures in acidic media. In the first years
of development of the technology of lowꢀtemperature FC,
Pt black was used as a cathode for the catalyst (with the
loading about 30 mg cm–2 and the typical Pt particle size
range 5—30 nm). Currently, platinum supported on carꢀ
bon black is used, which made it possible to decrease the
catalyst loading about two orders of magnitude.
It became possible to develop catalysts based on carꢀ
bon black due to the stabilization of platinum on the supꢀ
port surface as nanosized clusters (2—10 nm), whose speꢀ
cific catalytic activity depends on the particle size ("size
effect").1 There were several attempts to explain a similar
phenomenon by specific features of the electronic strucꢀ
ture of the catalyst, the number of atoms with the low
coordination number, the predominant crystallographic
orientation, and the existence of defects. Probably, the
simultaneous influence of factors of this kind on the activꢀ
ity of the catalyst is a reason for ambiguous results of the
search for the optimum relation between the catalyst morꢀ
phology and catalytic activity. It was shown2,3 for the oxiꢀ
dation of hydrogen that the optimum size of the Pt crysꢀ
tallite is 2—3 nm and that for the reduction of oxygen is
2—4 nm.4,5 However, it is shown in some works that no
size effect is observed for the platinum crystallites larger than
1.4 nm.6—8 Chronoamperometric procedures were proꢀ
posed in several works for the study of the activity of platꢀ
inum nanoclusters. It was shown that the areas of intergꢀ
rain boundaries contribute considerably to the activity.9
Nanostructured carbon materials (nanotubes (CNT),
nanofibers (CNF)) and diverse oxide supports are used
along with carbon black as catalyst supports. The low cost,
high conductivity, good mechanical properties, and the
resistance to oxidation, which is higher than that of carbon
black, are usually referred to advantages of carbon nanoꢀ
fibers. The extended structure of the CNF favors the imꢀ
provement of the properties of the catalytic layer in the
fuel cell.
The surface functional groups of the carbon supports
affect the formation, dispersity, and strength of the bond
with the support of reduced small platinum particles
(2—4 nm). In addition to the catalytic activity, an imporꢀ
tant characteristic of the active FC layer is its stability in
time. The decomposition of the active layer can be related
to the agglomeration of particular platinum clusters.
Therefore, the authors of several works10—16 believe that
the chemical modification of the carbon surface by, for
example, carboxyl and thiol groups makes it possible to
anchor the metal on the support surface. However, the
oxidation of CNF results in the formation of unbound
residual modifying molecules, which complicate a further
interpretation of the electrochemical behavior of the preꢀ
pared catalysts and which may exert a negative effect on
the electroconducting properties of the support.17—19
The purpose of this work is to study the influence of
the composition of Pt/CNF and the platinum cluster size
on the electrochemical properties of the catalyst.
Published in Russian in Izvestiya Akademii Nauk. Seriya Khimicheskaya, No. 6, pp. 1021—1026, June, 2011.
1066ꢀ5285/11/6006ꢀ1045 © 2011 Springer Science+Business Media, Inc.

1046
Russ.Chem.Bull., Int.Ed., Vol. 60, No. 6, June, 2011
Gerasimova et al.
potential was attained). The reversible hydrogen electrode was
used as a reference electrode, and a glassy carbon plate served as
a reference electrode. The measurements were carried out acꢀ
cording to the procedure described earlier.23
Experimental
Carbon nanofibers with a diameter of 100—200 nm, a length
of 1—5 μm, and a specific surface of the 100 m2 g–1 BET (N2)
were used as platinum supports.20 According to the elemental
analysis data, the oxygen content is lower than 0.5%.
Results and Discussion
Platinum was supported on carbon fibers by the reduction of
PtCl62– in an alkaline solution of ethylene glycol (pH 11) under
the action of microwave radiation (800 W, 1 min). The syntheꢀ
sized samples of the Pt/C catalysts were filtered and dried for
10 h at 105 °C in a desiccator. Thermogravimetric (TGA) and
differential thermal (DTA) analyses were carried out using an
STA 409 Luxx thermoanalyzer with the mass spectrometric atꢀ
tachment for analysis of gaseous products (Netzsch). The TGA
and DTA curves were recorded in the linear heating regime
at 10 °C min–1 from room temperature to 1000 °C in an air
flow (flow rate 50 mL min–1). Weighed samples (5—10 mg)
were placed in a corundum crucible, and the platinum
content was determined by the incombustible residue at the
end of experiment. The samples with the weight content of platꢀ
inum from 4% to 61% were obtained by the variation of the
Pt : C ratio.
The synthesized samples Pt/CNF with a weight conꢀ
tent of platinum of 4—61% were studied by transmission
electron microscopy. The microimages of the samples with
the platinum content 4, 12, and 49% are shown in Fig. 1.
The platinum particles are uniformly distributed over the
surface of carbon fibers. It is seen that the number of large
clusters increases with an increase in the platinum conꢀ
tent. The Pt cluster size distribution (Fig. 2) calculated by
the TEM data is described by the logꢀnormal distribution
typical of freely growing particles.24,25 The averageꢀmass
distribution is broader than the averageꢀnumerical one,
and this difference increases with an increase in the platiꢀ
num content. Assuming that the nanoparticles are spheriꢀ
cal, the specific surface Sgeom was calculated for all cataꢀ
lysts (Table 1).
The prepared catalysts were also characterized by powꢀ
der Xꢀray diffraction (Fig. 3). The platinum content in the
studied samples exerts a noticeable effect on the diffracꢀ
tion patterns. All reflections of platinum in the region
studied are well seen in the samples with a high platinum
content. For the samples with a low platinum content, the
reflection intensity from the carbon nanofibers is much
higher than that from the platinum nanoparticles, and only
one peak of platinum corresponding to the reflection [111]
is observed. Based on the Xꢀray diffraction patterns obꢀ
The specific surface of materials was determined by the BET
method on a Quadrasorb Sl instrument (Quantachrome). The
pore size distribution was constructed on the basis of the analysis
of the curves of nitrogen sorption/desorption.21,22
The platinum particle sizes were determined by transmission
electron microscopy (TEM) (Zeiss LEO 912B) and using powder
Xꢀray diffraction (Brucker D8 Advances, Kα monochromator,
1
1D detector, CuKα radiation). An analysis of the microimages
1
of the catalysts made it possible to construct the averageꢀnumerꢀ
ical particleꢀsize distributions for platinum (200—500 units), acꢀ
cepting that the particles were spherical and ignoring agglomerꢀ
ates. To obtain the averageꢀmass distribution, the weight was
calculated for each measured particle.
For electrochemical studies of the catalysts, the electrodes
were prepared as follows. Carbon nanofibers with supported platꢀ
inum were dispersed in water with the addition of dispersion
Nafion® DE1020 (DuPont, USA) in an amount of 25% of the
carbon weight and were supported in the surface of the gasꢀ
diffusion layer Toray® TGPꢀHꢀ060 (Toray Industries, Inc.,
Japan). The prepared electrode was used as a working electrode
in the threeꢀelectrode cell, and a reversible hydrogen electrode
(0.5 М H2SO4, 20 °C) served as a reference. Prior to measureꢀ
ments of carbon monoxide adsorption, background voltammoꢀ
grams were detected on the electrode under study in the potenꢀ
Table 1. Characteristics of the Pt/CNF catalysts with various
platinum contents
b
Conꢀ
ja/A (mg of Pt)–1SСО Sgeom DN
DM
DXRD
tent of Pt/
(wt.%)
(0.5 V)
m2 g–1
nm
c
4
7
0.32
3.05
2.61
0.75
1.08
0.31
0.60
30 250 1.0
90 120 2.1
120 125 2.0
90 155 1.7
1.2
2.6
2.5
2.1
6.3
3.8
4.3
—
2.5(40°)
c
tial range from 0.1 to 1.2 V with a potential sweep of 0.02 V s–1
.
12
23
34
49
61
—
The true surface of the electrodes was calculated by the surface
area of the desorption peak of carbon monoxide. For this purꢀ
pose, the solution in the cell was saturated with CO for 15 min
at an electrode potential of 0.1 V, then CO was removed by
blowing with argon, and cyclic voltammograms were detectꢀ
ed. All potentials are presented relatively to the reversible
hydrogen electrode taking into account ohmic losses, which
were determined by the impulse potentiostatic method (inacꢀ
curacy 3%).
The electrodes of fuel cells were studied in the reduction of
oxygen in a liquid gasꢀdiffusion halfꢀcell (1 M H2SO4) by cyclic
and stationary voltammetry (as stationary were taken the values
of current detected 30 min after the moment at which specified
2.3(40°)
6.3(67°)
3.4(67°)
3.7(67°)
25
30
40
50
85
70
4.1
2.4
2.6
a At a potential of 0.5 V.
b Calculation by the Sherrer equation, D = 0.9•0.5406/V cosθ,
the value of 2θ is given in parentheses.
c The reflections corresponding to Pt are comparable with the
background.
Note. DN and DM are the averageꢀnumerical and averageꢀmass
sizes of the platinum clusters; and DXRD is the size of the Pt
clusters determined by Xꢀray diffraction analysis.

Electrocatalytic behavior of Pt/C catalysts
Russ.Chem.Bull., Int.Ed., Vol. 60, No. 6, June, 2011
1047
a
b
c
50 nm
50 nm
50 nm
Fig. 1. Microimages of the Pt/CNF samples with the platinum content 4 (a), 12 (b), and 49 wt.% (c).
tained, the size of platinum crystallites was determined
assuming that the particles are cubic with the very narrow
size distribution.
The averageꢀmass distributions better correspond to
the Xꢀray phase analysis results.
The thermal resistance of the prepared catalysts to
oxidation was studied by thermogravimetry. The temperaꢀ
ture of the onset of sample oxidation was chosen as
a parameter that characterizes resistance. It was shown
that the resistance of the prepared samples decreases
monotonically with an increase in the platinum content
and is close to the resistance of the catalysts based on
carbon black Vulcan XCꢀ72 for the samples with the same
platinum content (Fig. 4).
The combustion of carbon unbound with Pt occurs in
the temperature range from 530 to 600 °C. As can be seen
from Fig. 4, no considerable decrease in the temperature
of the onset of carbon oxidation is observed when an inꢀ
crease in the platinum content exceeds 20%, apparently
metal particles cover the carbon surface in an optimal
way, regardless of the nature of the support (CNF, carꢀ
bon black).
In order to simulate cathodic and anodic processes in
the FC, the materials obtained were studied in a liquid
N
a
b
c
N
N
80
60
40
20
40
30
20
10
40
30
20
10
0.5
1.0
1.5
2.0 D/nm
1
2
3
4 D/nm
1
2
3
4
5
6 D/nm
N
N
N
d
e
f
30
15
15
10
5
20
10
10
5
1.0
1.5
2.0 D/nm
1.5 2.0 2.5
3.0
D/nm
1
2
3
4
5
6 D/nm
Fig. 2. Numerical (a—c) and weight (d—f) distributions of platinum clusters in the Pt/CNF samples with the platinum content 4 (a, d),
12 (b, e), and 49 wt.% (c, f) calculated by the TEM data. N is the number of clusters, and D is the cluster size.

1048
Russ.Chem.Bull., Int.Ed., Vol. 60, No. 6, June, 2011
Gerasimova et al.
I (arb. units)
I (arb. units)
4000
4
1
5
600
500
400
3000
3
2
300
6
38
40
42 2θ/deg
7
2000
1000
3
2
4
1
39
40
41
2θ/deg
Fig. 3. XRD peak corresponding to the reflection from the plane [111] of the Pt crystallites of the synthesized catalysts Pt/CNF conꢀ
taining 4 (1), 7 (2), 12 (3), 23 (4), 34 (5), 49 (6), and 61 wt.% (7) of platinum. Curves (1)—(4) in the magnified scale are shown in inset.
gasꢀdiffusion halfꢀcell, which makes it possible to study
the processes of oxygen reduction and hydrogen oxidation
that occur independently in a fuel cell.23 The multistep
reaction of oxygen reduction is rateꢀdetermining in a hyꢀ
drogen—air fuel cell. Therefore, to reveal the influence of
the morphology of catalytic particles on the rate of the
cathodic process, we recorded the stationary voltammetꢀ
ric curves of oxygen reduction on the prepared catalysts.
The current densities on the catalysts in the reaction of
oxygen reduction at a potential of 0.5 V were calculated to
compare the working efficiency of the catalysts studied
(see Table 1). It is shown that the dependence of the
characteristics of the oxygen process on the platinum conꢀ
tent in the catalyst can be described by a curve with the
maximum, and the optimum Pt content is about 8—12%.
However, the extreme mode of the obtained dependences
is the manifestation of the size effect (Fig. 5), which
depends on both the nature of surface defects and kinetic
regularities of the reaction on the catalyst surface.
The true surface of platinum was also determined by
the materials Pt/CNF from the surface area of the CO
desorption peak (see Table 1). A comparison of the values
of SCO with Sgeom calculated from the size distributions of
the surfaces showed that the results of the both methods
for surface determination are poorly consistent when they
are used for the estimation of platinum particles with the
diameter lower than 2 nm. The same observation was made
by the authors of Ref. 4. It is known that the specific
catalyst surface increases sharply with a decrease in
the particle size, which becomes especially pronounced
for the crystallites with the sizes smaller than 2 nm.
Thus, it could be expected that the electrochemical acꢀ
tivity, which is directly affected by the number of surface
atoms, should increase sharply with a decrease in the
particle size. However, this does not correspond to the
data presented in Fig. 5. It is possible that a part of the
cluster surface is inactive, because the particles are incorꢀ
porated in the pores of the carbon support and/or the
T/°C
600
1
550
2
500
450
400
350
10
20
30
40
50
60 CPt (%)
Fig. 4. The temperature of the onset of combustion of the cataꢀ
lysts based on the carbon nanofibers Pt/CNF (1) and carbon
black Pt/XCꢀ72 (2) vs platinum content (CPt) in the material.

Electrocatalytic behavior of Pt/C catalysts
Russ.Chem.Bull., Int.Ed., Vol. 60, No. 6, June, 2011
1049
j/A (mg of Pt)–1
decreases monotonically with an increase in the amount
of platinum. The dependence of the specific power of the
cathode based on the prepared catalysts on the platinum
cluster size can be described by the curve with the maxiꢀ
mum. The presence of micropores on the support surface
was shown to exert a substantial effect on the electroꢀ
chemical behavior of small platinum clusters (d ≤ 2 nm).
The effect indicates that some metal particles with the size
smaller or comparable with the pore size of the carbon
fiber are blocked by fibers. The results obtained in the
work suggest that the Pt/C electrocatalysts with the platiꢀ
num content 7—12 wt.% and the particle size 2—3 nm are
most promising among the catalysts based on carbon
nanofibers.
3.0
2.5
2.0
1.5
1.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0 D/nm
Fig. 5. Current density (j) vs size of the metal clusters (D) for the
catalysts Pt/CNF at a potential of 0.5 V.
The carbon nanofibers were kindly presented by A. A.
Volodin (Institute of Problems of Chemical Physics, Rusꢀ
sian Academy of Sciences). The authors are grateful to
I. S. Bushmarinov (N. A. Nesmeyanov Institute of Orgaꢀ
noelement Compounds, Russian Academy of Sciences)
for Xꢀray diffraction studies.
This work was financially supported by the Russian
Foundation for Basic Research (Project No. 09ꢀ03ꢀ
01157a).
surface platinum atoms are deactivated by the surface
groups of the support.
To check this assumption, we measured nitrogen
adsorption and desorption on the initial CNF and on the
CNFꢀbased catalyst. The plots of the pore volume vs diꢀ
ameter are presented in Fig. 6. The results of the analysis
show that the most part of pores in the initial nanofibers
have the size from 1 to 2 nm, whereas these pores disapꢀ
pear in the fibers with supported platinum. This suggests
that nucleation starts in pores of the carbon support and
continues until the cluster size would be comparable
with the pore size. As a result, the particle that formed
becomes incapable of efficient catalyzing electrochemical
processes.
In the present work, the samples of Pt/CNF with various
platinum content were prepared by the polyol method.
Their morphology, thermal stability, and electroactivity
in the reaction of oxygen reduction were studied. It was
found that the resistance of the materials to oxidation
References
1. Catalysis and Electrocatalysis at Nanoparticle Surfaces, Eds
A. Wieckowski, E. R. Savinova, C. G. Vayenas, Marcel Dekꢀ
ker, New York, 2003, 970 pp.
2. J. Lipkowski, P. N. Ross, Electrocatalysis, WileyꢀVCH, New
York, 1998, p. 197.
3. P. H. Lewis, J. Phys. Chem., 1963, 67, 2151.
4. K. Makino, M. Chiba, T. Koido, Meet. Abstr. — Electroꢀ
chem. Soc., 2010, 1002, 647.
5. L. Gan, H. Du, B. Li, F. Kang, New Carbon Mater., 2010,
25, 53.
6. M. Watanabe, H. Sei, P. Stonehart, J. Electroanal. Chem.,
1989, 261, 375.
7. M. Watanabe, S. Saegusa, P. Stonehart, Chem. Lett., 1988,
17, 1487.
Vp•10–2/cm3 g–1
8. M. Watanabe, S. Saegusa, P. Stonehart, J. Electroanal. Chem.,
1989, 271, 213.
1.6
1
2
9. O. V. Cherstiouk, A. N. Gavrilov, L. M. Plyasova, I. Yu.
Molina, G. A. Tsirlina, E. R. Savinova, J. Solid State Electroꢀ
chem., 2008, 12, 497.
1.2
10. Zh. Liu, X. Lin, J. Y. Lee, W. Zhang, M. Han, L. M. Gan,
Langmuir, 2002, 18, 4054.
11. H. Hiura, T. W. Ebbesen, K. Tanigaki, Adv. Mater., 1995,
7, 275.
12. D. S. Bag, R. Dubey, N. Zhang, J. Xie, V. K. Varadan,
D. Lal, G. N. Mathur, Smart Mater. Struct., 2004, 13, 1263.
13. X. Li, I.ꢀM. Hsing, Electrochim. Acta, 2006, 51, 5250.
14. Y.ꢀT. Kim, K. Ohshima, K. Higashimine, T. Uruga, M. Taꢀ
kata, H. Suematsu, T. Mitani, Angew Chem., Int. Ed., 2006,
45, 407.
0.8
0.4
1.5
2.0
2.5
3.0
3.5 Dp/nm
Fig. 6. Micropore size distribution for the initial CNF (1) and
catalyst Pt/CNF with the platinum content 4 wt.% (2).
15. J. K. Lim, W. S. Yun, M. Yoon, S. K. Lee, C. H. Kim,
K. Kim, S. K. Kim, Synth. Met., 2003, 139, 521.

1050
Russ.Chem.Bull., Int.Ed., Vol. 60, No. 6, June, 2011
Gerasimova et al.
16. X. Sun, R. Li, D. Villers, J. P. Dodelet, S. Désilets, Chem.
Phys. Lett., 2003, 379, 99.
17. D. Sebastián, I. Suelves, R. Moliner, M. J. Lázaro, Carbon,
2010, 48, 4421.
23. E. A. Astaf´ev, Yu. A. Dobrovolsky, E. V. Gerasimova, A. V.
Arsatov, Al´ternativnaya energetika i ekologiya [Alterꢀ
native Power Engineering and Ecology], 2008, No. 2, 86
(in Russian).
18. D. Pantea, H. Darmstadt, S. Kaliaguine, L. Sümchen,
C. Roy, Carbon, 2001, 39, 1147.
24. L. B. Kiss, J. Söderlund, G. A. Niklasson, C. G. Granqvist,
Nanotechnology, 1999, 10, 25.
19. D. Pantea, H. Darmstadt, S. Kaliaguine, C. Roy, Appl. Surf.
Sci., 2003, 217, 181.
25. C. G. Granqvist, R. A. Buhman, J. Appl. Phys., 1976,
47, 2200.
20. A. A. Volodin, P. V. Fursikov, Yu. A. Kasumov, I. I. Khodos,
B. P. Tarasov, Izv. Akad. Nauk, Ser. Khim., 2005, 2210 [Russ.
Сhem. Bull., Int. Ed., 2005, 54, 2281].
21. P. Tarazona, U. Marini Bettolo Marconi, R. Evans, Mol.
Phys., 1987, 60, 573.
22. J. Jagiello, D. Tolles, in Fundamentals of Adsorption, Ed.
F. Meunier, Elsevier, Paris, 1998, vol. 6, p. 629.
Received March 4, 2011;
in revised form June 6, 2011