Bezerra2007 PDF

Available online at www.sciencedirect.
com
Journal of Power Sources 173 (2007) 891–908
Review
A review of heat-treatment effects on activity and stability of

PEM fuel cell catalysts for oxygen reduction reaction
Cicero W.B. Bezerra a,b , Lei Zhang a , Hansan Liu a , Kunchan Lee a ,
Aldaléa L.B. Marques c , Edmar P. Marques b , Haijiang Wang a , Jiujun Zhang a,∗
a Institute for Fuel Cell Innovation, National Research Council of Canada, Vancouver, BC V6T 1W5, Canada
b Department of Chemistry, Universidade Federal do Maranhão, Av. dos Portugueses, S/N 65.080-040 São Luı́s, MA, Brazil
c Department of Technology Chemistry, Universidade Federal do Maranhão, São Luı́s, MA, Brazil
Received 25 June 2007; received in revised form 10 August 2007; accepted 13 August 2007
Available online 19 August 2007
Abstract
This paper reviews over 120 papers regarding the effect of heat treatment on the catalytic activity and stability of proton exchange membrane
(PEM) fuel cell catalysts. These catalysts include primarily unsupported and carbon-supported platinum (Pt), Pt alloys, non-Pt alloys, and transition
metal macrocycles. The heat treatment can induce changes in catalyst properties such as particle size, morphology, dispersion of the metal on
the support, alloying degree, active site formation, catalytic activity, and catalytic stability. The optimum heat-treatment temperature and time
period are strongly dependent on the individual catalyst. With respect to Pt-based catalysts, heat treatment can induce particle-size growth, better
alloying degree, and changes in the catalyst surface morphology from amorphous to more ordered states, all of which have a remarkable effect on
oxygen reduction reaction (ORR) activity and stability. However, heat treatment of the catalyst carbon supports can also significantly affect the
ORR catalytic activity of the supported catalyst. Regarding non-noble catalysts, in particular transition metal macrocycles, heat treatment is also
important in ORR activity and stability improvement. In fact, heat treatment is a necessary step for introducing more active catalytic sites. For
metal chalcogenide catalysts, it seems that heat treatment may not be necessary for catalytic activity and stability improvement. More research is
necessary to improve our fundamental understanding and to develop a new strategy that includes innovative heat-treatment processes for enhancing
fuel cell catalyst activity and stability.
© 2007 Elsevier B.V. All rights reserved.
Keywords: Proton exchange membrane fuel cells; Electrocatalysis; Oxygen reduction reaction; Heat treatment; Pt catalysts; Non-noble catalysts
Contents
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 892
2. Effects of heat treatment on carbon-supported Pt (Pt/C) catalysts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 893
2.1. Platinum particle size and morphology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 893
2.2. Carbon support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 893
2.3. Mechanism study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 895
3. Effects of heat treatment on carbon-supported Pt alloy (Pt–M/C) and non-Pt–M/C catalysts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 895
3.1. Particle size and structural parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 896
3.2. Activity and stability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 898
3.3. Non-Pt–M/C catalysts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 902
4. Effects of heat treatment on non-noble catalysts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 902
4.1. Transition metal macrocycles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 902
∗ Corresponding author. Tel.: +1 604 221 3087.

E-mail address: jiujun.zhang@nrc.gc.ca (J. Zhang).
0378-7753/$ – see front matter © 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.jpowsour.2007.08.028
892 C.W.B. Bezerra et al. / Journal of Power Sources 173 (2007) 891–908
4.2. Transition metal chalcogenides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 905

5. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 905
Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 906
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 906
1. Introduction Regarding ORR catalyst synthesis, heat-treatment has been

recognized as an important and sometimes necessary step for
Proton exchange membrane (PEM) fuel cells have been rec- catalytic activity improvement [16–21]. Many heat-treatment
ognized as the most promising energy converting devices in techniques, such as traditional oven/furnace heating [14,21],
terms of low or zero emissions and high efficiency [1–5]. It microwave heat-treatment [22–24], plasma thermal treatment
has been demonstrated that PEM fuel cells are able to meet [25–28], and ultrasonic spray pyrolysis [29], have been
transportation and stationary power requirements, which has applied to prepare PEM fuel cell electrocatalysts. Among
driven further development of this technology towards commer- these, the traditional oven/furnace heating technique is the
cialization. However, in this effort, two major technical gaps most widely used. In general, it involves heating the cat-
have been identified: high cost and low reliability and dura- alyst under an inert (N2 , Ar, or He) or reducing (H2 )
bility. The fuel cell catalyst is the major contributor to these atmosphere in the temperature range of 80–900 ◦ C for 1–4 h
difficulties. In recent years, although a great deal of effort [18,21].
has been put into the synthesis of cost-effective, active, and For example, heat treatment or thermal activation for Pt-
stable fuel cell catalysts, no real breakthroughs can be seen based catalyst synthesis has been considered a necessary step,
yet. Therefore, exploring breakthrough catalysts; improving which has a significant impact on the metal particle size and
catalyst activity, stability, and durability; and reducing cata- size distribution, particle surface morphology, and metal dis-
lyst cost are the major tasks in fuel cell commercialization persion on the support [10]. The benefits of heat treatment
[1–9]. are to remove any undesirable impurities resulting from early
At the current technical stage, the most practical catalysts in preparation stages, to allow a uniform dispersion and stable
fuel cells are highly dispersed platinum (Pt)-based nanoparti- distribution of the metal on the support, and, therefore, to
cles. These Pt nanoparticles are normally supported on carbon improve the electrocatalytic activity of the synthesized catalyst
particles in order to increase the active Pt surface and improve the [18].
catalyst utilization. However, there are several drawbacks to Pt- Another typical example is the heat treatment of macrocyclic
based catalysts, such as high cost, sensitivity to contaminants, metal complexes. Although the heat-treatment effect on these
no tolerance to methanol oxidation (direct methanol fuel cell metal macrocycle properties has been well documented, a full
(DMFC) application), fewer completed four-electron reduction understanding of the mechanism of catalytic activity improve-
reactions, and Pt dissolution. ment has not yet been achieved. When the catalysts are pyrolyzed
With respect to the exploration of alternative low-cost non-Pt at a desired high temperature in a flowing argon atmosphere, the
catalysts, several other types of catalysts, including supported molecular structure of the macrocycle is partially or completely
platinum group metal (PGM) such as Pd-, Ru- and Ir-based destroyed, resulting in an catalyst with much better catalyst
catalysts, bimetallic alloy catalysts, transition metal macrocy- activity and stability than an untreated catalyst. Heat treatment
cles, and transition metal chalcogenides, have been employed can also affect the fundamental properties of the catalyst and its
for PEM fuel cell catalysis [1,4,9]. The major force driving the carbon support, including the loading level of the catalyst on the
development of non-Pt catalysts is a reduction in cost. However, support, the number of catalytic sites, the acid–base properties
these approaches are still in the research stage, as catalyst activ- of the support, and the distribution of the catalyst particles on
ity and stability are still too low to be practical when compared the support.
to Pt-based catalysts. The effect of heat treatment on catalyst properties is signifi-
One of the major approaches to improving the activity and cant in terms of catalyst activity improvement. The mechanisms
stability of Pt-based and non-Pt catalysts is to optimize the of the catalyst reaction during the heat-treatment process and
catalyst synthesis procedure. It is well known that the electro- the resulting improvement in activity are complicated and not
catalyst performance is strongly dependent on the preparation fully understood. In addition, the relationship between heat
procedures, including the addition of metal and its precursor, treatment and the characteristics of the formed catalyst is
the support type and supporting strategy, and the heat-treatment also not completely clear. In order to accelerate the process
strategy [10–16]. For example, even if the same catalyst with the of fuel cell catalyst development, a literature review would
same catalyst loading is used, the catalytic activity will differ be of great benefit to the fuel cell catalyst community. In
depending on how the catalysts were attached to the electrode this paper, different catalyst heat-treatment strategies and their
surfaces. Methods for catalyst attachment include irreversible effects on the activities of several typical kinds of catalysts are
adsorption, vacuum deposition, conducting polymer incorpo- reviewed, in order to summarize the current level of understand-
ration, porous carbon impregnation, and solvent evaporation ing of the catalyst improvement mechanisms caused by heat
[15]. treatment.
C.W.B. Bezerra et al. / Journal of Power Sources 173 (2007) 891–908 893
2. Effects of heat treatment on carbon-supported Pt resulting in well-defined cubo-octahedron crystallites, which

(Pt/C) catalysts were distinguishable from the initially irregular Pt surface [10].
As a result, the CO oxidation activity on this heat-treated Pt
In PEM fuel cells including DMFCs, reactions include the catalyst was improved significantly.
oxygen reduction reaction (ORR) at the cathode and the H2 In addition, it was found that the CO oxidation activity was
(hydrogen fuel cells) or methanol (DMFCs) oxidation reaction also dependent on the Pt particle size in the range of approxi-
(HOR or MOR) at the anode. In a hydrogen fuel cell, cathodic mately 3–12 nm. The best catalytic performance was achieved
ORR is much slower than anodic HOR; therefore, the ORR is for the catalyst with an average Pt particle size of about 5 nm.
recognized as the determining reaction. In a DMFC, the rates The effect of heat treatment on the structural characteristics
of both ORR and MOR are slow and comparable. However, of the Pt/C catalyst was also observed by Antolini et al. [41].
the slow ORR is still a problem in a DMFC. Therefore, the Several Pt/C electrocatalysts were activated by dynamic heat
ORR catalysis will be our focus in this paper. It has been deter- treatment in an Ar atmosphere with two different heating rates
mined that the electrocatalytic reduction of molecular oxygen of 15 and 45 ◦ C min−1 . They observed that the heat treatment
on a catalyst can be largely affected by catalyst particle size and could lead to a growth in Pt particle size from 1 to 20 nm, inde-
surface structure [6,13,10,17,30–38]. Therefore, heat treatment pendent of the nature of the carbon support. The catalysts treated
can effectively change ORR activity and stability by altering the at different heating rates showed the same variation of Pt particle
catalyst morphology. size, but had different Pt crystallinities. The Pt crystallinity was
decreased with an increasing heating rate. The heat treatment in
2.1. Platinum particle size and morphology the temperature range of 400–550 ◦ C improved the Pt particle-
size distribution on the support surface and showed the largest
With respect to the effect of Pt/C particle size on ORR activ- increase in Pt crystallinity.
ity, there seems to be no agreement in the literature. Several In spite of some contradictory results on the effect of heat
work reported that particle size had a significant effect on ORR treatment on catalyst particle size, the effect on the ORR activity
activity, and that the optimum particle size was around 4 nm is clear; that is, heat treatment can effectively improve the ORR
[32,36,39]. When Pt catalyst particle size was smaller than activity of a Pt/C electrocatalyst.
3.5 nm, the catalyzed ORR rate was reduced. However, Yano
et al. [37] recently investigated a series of carbon-supported Pt 2.2. Carbon support
catalysts and found that the apparent ORR rate constant and the
activation energies were independent of Pt particle size in the The catalyst support (i.e., carbon support) used in a fuel cell
range of 1–5 nm. Bett et al. [40] also showed no dependency operation environment should be corrosion resistant and highly
of Pt/C specific activity on the crystallite size in the range of electrically conductive, and have a desirable water handling
3–40 nm. capability (hydrophilic character) [43,44]. Among all kinds of
The literature suggests that catalytic ORR activity might be carbon supports (activated carbon, carbon black, graphite, and
dependent on the relative concentrations of the Pt surface atoms, graphitized materials), carbon black is the most commonly used
the coordination numbers of the localized atoms, and the ratio for electrocatalysis due to its high conductivity, excellent cor-
of preferable crystal planes on the crystallite surface [36,41]. rosion resistance, appropriate pore structure, and high-surface
However, Bett et al. [40] found that Pt atoms at the corners, area [44–46]. Normally, graphitization is required in order to
edges, and kink sites or dislocation showed less ORR activity achieve a better level of corrosion resistance [45].
than those on the crystallite faces. Watanabe et al. [34] examined In practice, carbon supports are produced by burning car-
the influence of Pt crystallite size on ORR activity and found that bonaceous materials. The properties of the carbon supports are
the effect of crystallite dimensions was less important than that strongly dependent on the nature of the starting materials and the
of the intercrystallite distances. They observed that when the preparation conditions, such as activation method, temperature,
crystallite distance was greater than approximately 18–20 nm, and heating time [43]. The surface of the formed carbon supports
high-catalytic activity could be achieved, but if the Pt crystallites contains some oxygen functional groups, such as the carbonyl,
were too close, the specific activity of the catalyst was decreased. carboxylic, phenolic, anhydride, lactonic and etheric groups.
For the heat-treatment effect on catalyst morphology, Han et These functional groups can significantly affect the manufacture
al. [10] carried out conventional heat treatment of 20% Pt/C (loading and dispersion) and performance of the electrocata-
catalysts using several temperatures from 300 to 700 ◦ C for lysts, and are also responsible for both the acid–base and the
various periods of time (3–9 h) in a nitrogen and hydrogen mix- redox properties of the carbon supports [5,16,45]. Generally,
ture atmosphere. The effect of heat treatment on CO oxidation the surface oxygen-based groups can act as active or anchoring
was investigated using the cyclic voltammetry (CV) technique. centers for the generation of highly dispersed catalyst metallic
They observed that the average particle sizes grew exponentially crystallites. The surface functional groups can also affect the cat-
with the heating temperature and linearly with the heating time alyst deposition processes, including adsorption, ion-exchange
period. This observation was reported previously by Antolini or coordination, and deposition of metallic species by a redox
[30] and Alexopoulos et al. [42], seperately. reaction with specific groups on the carbon surface. For the
Moreover, it also observed that the heat treatment could adsorption process, the carbon support surface potential of zero
induce a uniform distribution of Pt over the support particles, charge seems to be of importance. It has been recognized that the
catalyst deposition process on the support is complicated. Mixed assembly (MEA) in a PEM fuel cell. The heat-treated supports
processes mentioned above could be involved in the catalyst were heated at 600 ◦ C for 30 min in a N2 environment. The
deposition [44,45]. Several works [16,17,47–49] demonstrated results showed that the carbon surface composition, Pt particle
that the number of oxygenated surface groups on the support size, and Pt dispersion were significantly affected by the heat
could alter both the dispersion and the resistance of the sintered treatment of the carbon support. The oxygen content in the heat-
Pt/C catalysts. treated support sample was about 2.8 at.%, whereas that of the
For carbon-support heat treatment, Prado-Burguete et al. [47] untreated one was 4.68 at.%. Transmission electron microscopy
prepared a series of Pt/C catalysts using different pre-heat- (TEM) and X-ray diffraction (XRD) for both the heat-treated
treated carbon supports. The effects of the pre-heat-treatment and the untreated catalysts showed that the catalyst with the
and chemical treatment on the carbon porous texture and amount heat-treated support gave a smaller particle size and higher rela-
of oxygen surface groups were investigated. In their experi- tive content of the Pt (1 1 1) crystal face. The average Pt particle
ments, carbon black samples were first heat treated at 950 ◦ C in a sizes for the heat-treated and the untreated supports were 3.2
H2 flowing environment for 12 h. Two of the samples were then and 4.2 nm, respectively. Electrochemical measurement, using
put into solutions containing 8 and 12N H2 O2 for oxidation treat- a Johnson Matthey Pt/C catalyst as the baseline, clearly indi-
ment. According to the N2 adsorption isotherms for the supports, cated that the catalyst with a heat-treated support showed higher
neither heat treatment in H2 nor heat treatment followed by oxi- ORR activity than did the untreated one. In fact, its activity
dizing treatment with peroxide changed the porous texture of the was similar to the baseline catalyst. Since the catalyst with the
carbon support appreciably. However, further work [48] showed untreated support has a particle size close to that of the baseline
that heat treatment in a helium environment could produce a catalyst, this result suggests that the particle size may not be a
drastic change in the porous structure of the carbon support, determining factor in electrocatalytic activity.
resulting in a significant decrease in micro- and meso-porosity. Antonucci et al. [13] prepared some carbon black-supported
Pt catalysts were prepared using those pre-treated supports at dif- Pt electrocatalysts for the purpose of studying the interaction
ferent temperatures and time periods. The results showed that between Pt crystallites and the surface functional groups of the
H2 O2 treatment could introduce new oxygen-containing surface support. All catalysts studied were heat treated under reduc-
species such as carboxylic and phenolic groups. With respect to ing conditions in the temperature range of 400–900 ◦ C. They
the catalyst dispersion, the H2 O2 oxidation-treated carbon sup- showed that with an increase in the amount of oxygenated
port resulted in much better Pt dispersion than did the support species on the carbon support, the Pt surface area was decreased.
that was only heat treated. This phenomenon was attributed to This may suggest that the Pt dispersion is sensitive to the heat
the large amount of oxygen-containing surface groups on the treatment, which can produce different amounts of functional
support. These groups can increase the wettability of the car- groups on the carbon surface depending on the heat-treatment
bon particle surface, facilitating the metal precursor distribution temperature. It was also observed that with an increasing in the
throughout the particle surface at the impregnation and reduction number of mildly acidic functionalities (pKa was in the range
steps of the catalyst synthesis. In addition, the catalyst formed 5.2–6.4), the metal dispersion improved; this result agreed with
on the H2 O2 oxidation-treated support gave much better thermal other observations [47,50]. Interestingly, a relationship between
stability. the metal surface area (MSA) and the number of basic functional
Pre-graphitized carbon has also been used for the catalyst sup- groups of the support surface was also observed. Two distinct
ports. Coloma et al. [16] studied three samples of pre-graphitized trends can be clearly seen in this relationship, depending on the
carbon blacks with the same porous texture. For comparison, roles of the different basic functionalities on the MSA. These two
different amounts of oxygen-containing surface groups were trends were believed to be associated with the different natures
created by oxidation treatment with H2 O2 . They observed that of the two basic groups: C␲ sites and pyrone-type complexes. In
the presence of these oxygen-containing groups affected the final the former case, increasing the C␲ sites causes an increase in the
dispersion of metallic Pt and the resistance to particle sintering, Pt surface area, giving rise to higher metal dispersions and more
resulting in less favourable dispersion and lower sintering resis- electrochemically active sites. For pyrone-type groups, which
tance. It was observed that at the impregnation step of the catalyst have a strong basic property (produced at high-heat-treatment
synthesis, the oxygenated groups on the support particles led to temperatures), can lead to the formation of Pt clusters, causing
the best metal distribution. However, at the reduction step of Pt an increase in the particle size and a reduction in the number of
ions, the oxygen-containing groups were decomposed, causing electrochemically active Pt sites.
a surface redistribution of the adsorbed Pt towards the ␲ sites Torre et al. [50] carried out heat treatment on Pt/C catalysts
in the basal planes of the graphitic crystallites. Consequently, prepared by three different methods in a flowing N2 atmosphere
a less favourable Pt dispersion and a lower sintering resistance at 200–900 ◦ C for 45 min to approximately 4 h. Their main pur-
were produced. pose was to study the effect of the acid–base properties of the
Tian et al. [17] reported that heat treatment of the carbon carbon support on Pt dispersion. According to their results,
support plays an important role in the improvement of Pt/C heat-treatment increased the basic properties of the catalysts by
electrocatalytic activity towards ORR. Pt/C catalysts were pre- removing acidic oxygen-containing groups from the particle sur-
pared using heat-treated and untreated carbon black supports. face to result in a greater number of C␲ sites and a pyrone-like
Electrochemical performance of the Pt/C catalysts were evalu- structure [51]. Although the electrochemical activities of these
ated based on the current–voltage curves of membrane electrode catalysts were not evaluated, the active MSAs were determined
functional groups of the support and (2) changes in Pt dispersion,

morphology, and crystallite growth on the carbon support.
The carbon support surface is very heterogeneous and con-
tains a wide variety of functional groups, which can significantly
affect the chemical properties of such a surface. It has been
widely recognized that the mostly common surface functional
groups are those species containing oxygen, as mentioned above
[5,16,40,55,56]. Thermal desorption of surface oxides, such as
CO2 , CO, and H2 O, is the most common indicator of the surface
oxygen-containing groups of carbon, which can be detected by
a mass spectrometer and/or a gas chromatograph. The detected
species evoluted at different temperatures can then be used to
infer the carbon surface functional groups [55]. Fig. 2 shows
Fig. 1. Effect of heat-treatment on specific corrosion rate of Vulcan XC 72R. that heat treatment can change carbon surface functional groups
Corrosion currents recorded at 1.0 V in 180 ◦ C H3 PO4 . [53]. primarily by reducing the amount of oxygen-containing acidic
species. As a result, the surface becomes more basic due to
in a H3 PO4 solution using cyclic voltammetry. In addition, the the increased number of ␲ bonding formation (C C) and the
Pt dispersion was also verified by TEM. Their results indicated pyrone-like structure. However, heat treatment can also induce
that the metal surface areas decreased with the increasing tem- a graphitic structure through polymerization and aromatization
perature as well as with increasing pHZPC , i.e., a pH value where of the nanocrystalline graphene [56]. In this graphitization pro-
the surface exhibits a net zero charge. Their results suggested cess, high temperatures are necessary [6,16,45,56]. During the
that 700 ◦ C could be the optimum heat-treatment temperature in polymerization process, the polycyclic aromatic molecules are
the weakly basic catalyst range. formed with the elimination of heteroatoms. At a temperature
Aricò et al. [51] evaluated the effect of Pt loading, heat of approximately 1630 ◦ C, the large aromatic molecules start
treatment and air exposure on the acid–base properties of Pt/C to condense into graphene, and when the temperature is higher
catalysts. Heat treatment was performed in a dry N2 flux, at the than about 1730 ◦ C, aromatization occurs to form a graphitic
following temperatures and heating times: 110 ◦ C, 4 h; 600 ◦ C, structure. Polymerization and aromatization can be indicated by
4 h; and 900 ◦ C, 1 h. These authors also concluded that heat average lattice parameters extracted from XRS [56]. The tem-
treatment causes changes in the catalyst surface by increasing perature range in which graphitization occurs is between 827
the number of the basic sites. and 2230 ◦ C. Temperature can significantly affect the number
It seems that the more oxygen-containing surface groups exist and the quality of structural defects in the material [56]. Kijima
on the catalyst support surface, the worse the Pt distribution is. et al. [57,58] investigated the reaction course of carbonization of
Heat treatment can change these surface groups from an acidic poly(phenylenebutadiynylenes) (PPBs) in the temperature range
status to a basic one (an increase in C␲ functionalities), leading of 200–600 ◦ C. A mechanism was proposed for the formation
to a better distribution of Pt particles on the carbon support and of nano-carbonaceous layers.
consequently resulting in better catalyst ORR activity and stabil- Regarding Pt deposition on a carbon support, the decompo-
ity, even in a fuel cell operation environment [13,14,17,39,41]. sition and redistribution of surface oxygen-containing groups
In addition, it was also observed that head treatment stabilized during heat treatment can affect the dispersion and growth of Pt
the carbon support against electrochemical corrosion [52,53]. crystallites [49]. Another recognized phenomenon is the ther-
Durability experiments conducted in a phosphoric acid fuel cell mal sintering of supported Pt even at relatively low temperatures
environment (1000 h at 191 ◦ C) showed that the disordered por- [41]. This leads to the growth and then the crystallization of Pt
tion of the carbon support was oxidized, whereas the outer particle size in the face-centered cubic structure. Several models
crystalline shell remained intact [52]. Stonehart [53] showed have been proposed to explain the mechanisms of particle-size
that heat treatment could stabilize the carbon against corrosive sintering and growth; the Smoluchowski and Ostwald ripening
degradation and increase the lifetime of fuel cell cathodes con- models are the most accepted ones [41,47]. In the first mecha-
taining carbon support-based Pt catalysts, as shown in Fig. 1. nism the metal transport on the support surface occurs through
The Pt catalysts with heat-treated carbon support were more the migration of Pt crystallites and is followed by a coalescence
stable than those with untreated supports due to the strong inter- of the particles. The second mechanism involves the evaporation
action between the Pt particles and the basic anchoring sites on of atoms from small crystallites and then condensation on larger
the carbon surface. It was observed that heat treatment at 370 ◦ C ones in a process similar to the Ostwald ripening of dispersed
in air led to an improvement in the stability of this catalyst in a precipitates in solution.
60 ◦ C sulphuric acid solution [30,52,54].
3. Effects of heat treatment on carbon-supported Pt
2.3. Mechanism study alloy (Pt–M/C) and non-Pt–M/C catalysts
With respect to the mechanism study, two effects induced by Pt alloys with various transition metals have been exten-
heat treatment should be considered: (1) changes in the surface sively studied in the effort to reduce catalyst cost and to improve
Fig. 2. Thermal decomposition products and corresponding evolution temperatures of carbon surface functional groups. Re-plotted from [55] by permission of The
Electrochemical Society.
both the resistance to sintering and the electrocatalytic activity be observed between the heat-treatment temperature and the
towards ORR [6,59,60]. There are many reports focusing on particle size: with an increase in temperature, the particle size
the synthesis and properties of Pt alloy catalysts mainly with V increases considerably due to the sintering of Pt particles.
[61,62], Co [60,63–69], Fe [67,68], Ni [63,67,68] Cu [67,68,70], The alloying of a second metal into the Pt can cause a change
Ti [8,71], Pd [72] or Cr [73] as the second metal. in the Pt lattice parameters, which can be directly observed by
Generally, this kind of catalyst is prepared by deposition of XRD. For example, a contraction in the lattice spacing or a shift
the transition metal precursor on the Pt/C material, followed by in the position of the reflection peaks is indicative of the second
chemical reduction and then alloying at high temperatures. The metal being dissolved and of the formation of an ordered Pt–M
role of heat treatment is mainly to aid the formation of the alloy phase. As seen in Table 1, when temperature is increased, the
by increasing the mobility of the supported Pt. However, the lattice constant decreases. According to Vegard’s law, the lattice
high-temperature heat treatment can have a negative effect on the parameters are a linear function of the alloy composition if the
catalytic activity due to sintering and the growth of Pt particles. temperature is fixed [73,74]. Thus, the degree of alloying can be
To avoid this, other procedures, such as simultaneous deposi- determined by the value of these lattice parameters. It can also be
tion and reduction of Pt and the second metal (M), the sol–gel seen in Table 1 that the degree of alloying formation is a function
process, and the polyol method, may be needed [59,64,69]. of the temperature. Therefore, heat-treatment temperature may
be used to adjust the degree of the alloying.
3.1. Particle size and structural parameters It is not necessary for heat treatment to change the alloy prop-
erties of a catalyst. For example, when Cambanis and Chadwick
Table 1 summarizes the literature data about the effect of [62] investigated a Pt–V alloy, not any evidence of intermetallic
heat treatment on particle size, lattice parameters, and alloy- Pt3 V formation could be found even after heat treatment at tem-
ing degree for carbon-supported Pt–M catalysts. A trend can peratures ranging from 830 to 930 ◦ C. However, Antolini et al.
Table 1
Effect of heat-treatment on structural parameters of reported Pt–M/C electrocatalysts
Second metal (M) Heat-treatment Treatment temperature/time Particle Lattice Crystal Alloying References
in Pt–M/C (Pt:M) conditions period (T (◦ C)/t (h)) sizea (nm) constant (Å) structure degree (%)
V (1:1) As-prepared – 3.4 3.8834 – 22 [61]

Furnace (Ar/H2 ) 500/2 3.9 3.8800 – 25
850/2 5.4 3.8686 fcc, Pt2 V 33
Co (3:1) Furnace (N2 /10% H2 ) 600/2.5 3.83 3.855 fcc, Pt3 M [63]
700/2.5 2.95 3.876 61b
900/2.5 6.60 3.866 79b
1100/2.5 8.63 3.868 75b
Cr (3:1) 700/2.5 4.23 3.903 39b
900/2.5 6.23 3.884 78b
1100/2.5 10.4 3.883 81b
Ni (3:1) 700/2.5 4.62 3.858 53b
900/2.5 6.35 3.841 70b
1100/2.5 10.2 3.836 75b
Coc (3:1) As-prepared – 3.4 3.925 – – [64]
Furnace 700/2 8.4 3.870 fcc, Pt3 Co –
900/2 8.0 3.867 – –
1200/2 12.0 3.864 fcc, Pt3 Co –
Cod (3:1) As-prepared – <2.5 3.927 – –
Furnace 700/2 2.5 3.894 – –
900/2 4.1 3.907 – –
1200/2 10.4 3.910 – –
Co (3:1) Furnace (H2 ) 550/3 3.9 3.911 7 [66]
900/3 6.8 3.897 35
Co (3:1) As-prepared – 1.8 3.501 79 [69]
Furnace (H2 ) 400/2 9.6 3.90
Furnace (air/H2 ) 200/1 (air) + 300/1 (H2 ) 8.3 3.90
Co (5:1) As-prepared – 3.6 – – – [67]
Furnace (Ar/10% H2 ) 200/1 4.9 – – –
900/1 6.6 – – –
Cu (3:1) As-prepared – 4.7 3.86 fcc, Pt3 Cu – [70]
Furnace (Ar/10% H2 ) 300/1 6.9 3.86 –
600/1 17.1 3.88 –
900/1 29.6 3.85 –
Fe (1:1) As-prepared – 6.2 3.848(1) Tetragonal – [68]
Furnace (Ar/10% H2 ) 500/1 6.3 3.847(1) –
600/1 6.5 3.845(2) –
800/1 6.8 3.848(1) –
Co (1:1) As-prepared – 5.5 3.782(1) –
Furnace (Ar/10% H2 ) 650/1 5.8 3.780(2) –
Tie As-prepared – 4.8 3.926 – [71]
Furnace (He) 700/2 11.0 3.916 –
900/2 12.5 3.907 Pt3 Ti
1200/2 14.4 3.906 Pt3 Ti
Tif As-prepared – <4.0 3.927 –
Furnace (He) 700/2 5.0 3.922 –
900/2 7.0 3.922 –
1200/2 28.0 3.908 Pt3 Ti
Surface area measured by CO chemisorption.

a Measured by XRD.
b According to Ref. [59].
c Acidic route: Co(OH) + HCl (pH 2).
2
d Alkaline route: Co(NO ) + NH OH (pH 11).
3 2 4
e Methanol:water solution (1:1).
f Methanol:water solution (1:1), pH 10.
[61] observed that the heat treatment of Pt–V/C at 850 ◦ C led to a result, ORR kinetics may be enhanced. For example, an
the formation of an ordered phase (fcc Pt2 V). alloyed Pd showed a reduced ORR overpotential by ∼50 mV
The particle size and lattice constant of the alloy are also at 0.2 mA cm−2 , compared to that catalyzed by pure Pd.
dependent on the preparation method. Heat-treated Pt–Co/C (2) Induced change in the Gibbs free energy for the ORR. Wang
alloys prepared in acidic media (pH 2) showed larger particle and Balbuena [82] explained that if a catalyst consists of
sizes, more significant ordering, and greater alloying degree than two metals, one with a low occupancy of d-orbitals (such
those prepared in the basic media (pH 11) [64]. as Co, Ni, Cr, or V) and the other with fully occupied d-
In the following section, the effect of the lattice structural orbitals (such as Pd, Au, and Ag), the d-orbital coupling
parameters on the catalytic activity and stability of Pt–M/C effect between them can significantly decrease the Gibbs
catalysts will be discussed. free energy for the electron transfer steps in the ORR, result-
ing in enhanced ORR kinetics.
3.2. Activity and stability (3) Pd lattice compression induced by alloy formation. A metal
d-band center can be altered by orbital overlap. The for-
The effects of electronic and geometric parameters such as mation of the alloy can induce Pd lattice compression (or
Pt–Pt distance, metal particle size, and surface structure changes reduction of bond lengths between metals), for example,
on the ORR catalytic activity and stability on the Pt/C catalysts via modification of electronic structures and orbital over-
discussed above are also applicable to the Pt–M/C catalysts. lap. This will cause a shift of the d-band center, resulting
However, some differences between Pt/C and Pt–M/C should in a change in the surface activity of the Pd sites. Den-
be emphasized [8,60,63,67–69,75–78]. sity functional theory calculations made by Hammer and
For a Pt–M/C catalyst, the catalytic activity as well as the Norskov [83] confirmed that compression of the Pd lattice
stability depends not only on the nature of Pt, but also on the in Pd alloys could downshift the energy of the d-band cen-
second metal. In terms of ORR catalytic activity and tolerance to ter. Pd lattice compression can thus enhance the catalyzed
chemical corrosion, He et al. [8] classified the binary Pt–M alloy ORR.
catalysts into four categories: (1) highly corrosive and highly (4) Facilitated O2 dissociation by alloy formation. Fernandez et
active (M = Fe, Co, V, and Mn); (2) corrosive and highly active al. [84] interpreted the enhancement in activity upon alloy-
(M = Zn, Cu, Mo, and Ni); (3) stable, but less active (M = Zr, Cr, ing using a simple thermodynamic model. They thought
and Ta); and (4) stable and active (M = W and Ti). incorporation of more active metals into Pd, such as Co,
A linear correlation between the ORR specific activity of Pt could facilitate the dissociative adsorption of O2 to form
and Pt–M catalysts and Pt–Pt nearest neighbouring distance has dissociated oxygen atoms (Oads ), and that these Oads could
been described in many works [61,65,76,79]. The smaller the migrate from the Co site to the Pd site whereupon electrore-
Pt–Pt distance, the more active is the ORR catalytic activity. Heat duction could occur with less polarization. For example,
treatment of an alloying catalyst at temperatures greater than rotating disk electrode (RDE) experiments [84] showed a
700 ◦ C has two counteracting effects: (1) formation of a better clear onset potential shift of ∼200 mV in the positive direc-
Pt–M alloy, which decreases the Pt–Pt distance and affects the tion for the ORR when Pd was alloyed with Co; and a Pd–Co
d-band vacancy of Pt, and thus improves the Pt electroactivity alloy with an atomic ratio of 80:20 could give a optimum
and (2) an increase in catalyst particle size, which decreases activity, which was very close to the activity exhibited by
the active area of Pt, resulting in a depression of catalyst mass pure Pt catalysts.
activity [62,80]. (5) Formation of the desirable “Pd shell and alloy core” struc-
Some studies showed that the d-band vacancy strongly ture. Lu and co-worker’s [85], in their recent paper on
depends on the Pt particle size induced by heat-treatment first-principles consideration in the design of Pd alloy cata-
[60,61,63]. According to current understanding, the hybridiza- lysts for ORR, proposed that the key to improving the ORR
tion of 5d state with empty states above the Fermi level can activity of Pd-based catalysts is to alloy Pd with elements of
reduce the d-electron number. However, this hybridization will smaller atomic size to form a “Pd shell and alloy core” struc-
become less favourable once the particle size is increased. ture, so as to take advantage of the lattice strain effect (lattice
With respect to a fundamental understanding of the ORR contraction induced by the incorporation of smaller atoms,
catalytic activity of alloys, Pd-based bimetallic catalysts, for such as Co, into the Pd lattice). This lattice strain effect can
example, were explored. A variety of hypotheses have been put cause a downshift in the metal d-band center, weakening
forward to address the enhancement observed in comparison the interaction between the catalyst and the adsorbate (such
with a pure metal: as O2 ). Three catalysts such as Pd, Pd–Co, and Pt showed
experimentally an ORR activity order of Pt > Pd–Co > Pd,
(1) Induced change in the density of states (DOS) at the Fermi which is consistent with the theoretical interpretation [85].
level of Pd sites by alloy formation. Ota and co-worker’s [81]
considered that the decrease in DOS induced by electron The main stability-related problem of a Pt alloy catalyst is
transfer from the second element such as Co, Ni, or Cr to its dissolution in acid environments. However, the low stability
Pd may weaken the chemisorption bonds between Pd and of alloy catalysts is caused not only by the dissolution of the
reactants such as O2 , O/OH, O2 − , and H2 O2 , reducing their metal in the electrolyte, but also by the loss of active surface
respective blocking effect in the O2 reduction process. As area partly due to the sintering [78,86].
Cambanis and Chadwick [62] prepared heat-treated Pt–V/C

catalysts with different Pt/vanadium ratios. The heat treatment
was performed at 830 and 930 ◦ C. They observed that heat treat-
ments at these two temperatures resulted in crystallite sizes in
the ranges of 3.3–3.5 and 4.0–4.5 nm, respectively, independent
of the vanadium content. These catalysts were tested for ORR at
180 ◦ C in a phosphoric acid solution. It was found that the cata-
lyst heat treated at 930 ◦ C had less activity, which was attributed
to the sintering effect. The results also showed that at both heat-
treatment temperatures, the catalyst activities were dependent
on the vanadium content, and the optimum vanadium content
was in the range of 15–22 at.%.
Antolini et al. [61] heat-treated dispersed Pt–V alloy catalysts
under a Ar + 10% H2 atmosphere for 2 h at 500 and 850 ◦ C in an
effort to evaluate the effect of particle size and lattice parameters Fig. 3. Variation of the real activity of Pt/C and Pt–Co/C PAFC cathode catalysts
as a function of heat-treatment temperature. Plotted according to the data in [64]
on the catalytic ORR performance in a PEM fuel cell. The cat- by permission of The Electrochemical Society.
alyst treated at 850 ◦ C showed a better ORR performance than
did Pt/C due to the formation of an ordered phase (fcc Pt2 V)
on the Pt–V alloy catalyst surface (fcc Pt2 V). XRD measure- alyst activity improvement. More work is needed to distinguish
ments indicated a decreased lattice parameter and an increased which factor is dominant.
particle size (from 3.4 to 5.4 nm) with an increase in heat- Beard and Ross [64] prepared Pt–Co/C alloys (3:1 atom ratio)
treatment temperature (see Table 1). Heat treatment can increase using two methods: one in an acidic (pH 2) and the other in
the Pt–V alloy degree, resulting in enhanced catalytic activity an alkaline (pH 11) aqueous medium. The synthesized cata-
towards ORR. lysts were then heat treated at 700, 900, and 1200 ◦ C, under an
Min et al. [63] prepared three carbon-supported alloy cata- inert atmosphere for 2 h (for structural parameters, see Table 1).
lysts (Pt–M/C, M = Co, Ni, Cr), and heat treated each of them Fig. 3 shows the specific activity under phosphoric acid fuel cell
separately at 600, 700, 900, and 1100 ◦ C in a reducing atmo- conditions, of heat-treated Pt/C and Pt–Co/C catalysts (synthe-
sphere (N2 /10% H2 ) for 2.5 h (Table 1). Their ORR catalytic sized in acidic media) as a function of treatment temperature.
activities were measured using a half-cell at an electrode poten- It can be seen that heat treatment improves the catalyst activ-
tial of 900 mV (vs. dynamic hydrogen electrode). Measurements ity in the whole temperature range, and that Pt–Co/C treated at
were carried out with pure oxygen in a 190 ◦ C H3 PO4 solution. 1200 ◦ C shows a slightly higher specific activity than Pt/C. An
The results showed that catalytic activity increased with increas- important observation was the effect of the heat-treatment on Co
ing heat-treatment temperature, and the amount of increased loss. Untreated samples synthesized by both methods showed a
activity was dependent on the nature of the second metal used in Co loss of more than 80%, whereas the samples heat treated
the alloying process, particle size and alloying degree with Pt. at 1200 ◦ C showed losses of only 15% and 28.4% for acidic
The altered Pt–Pt distance might be favourable for ORR. and alkaline routes, respectively. Therefore, heat treatment can
Santiago et al. [69] prepared a carbon-supported Pt70 Co30 effectively reduce Co loss.
nanoparticle catalyst using a polyol process. The catalyst had Watanabe et al. [65] prepared Pt–Co/C catalysts and investi-
a very small particle size of 1.9 ± 0.2 nm. A carbon power was gated the effect of heat treatment on their ORR catalytic activities
used for the support, which was heat treated at 850 ◦ C for 5 h and stability under the operating conditions of a phosphorous
in an Ar atmosphere before the catalyst anchoring. The cata- acid fuel cell. The heat treatment was performed in a reduc-
lyst formed was heat treated again in different ways. The first ing atmosphere (N2 + 10% H2 ) at several temperatures ranging
sample was treated at 400 ◦ C for 2 h in an H2 atmosphere. The from 400 to 900 ◦ C for different time intervals under 20 h. They
other sample was first treated at 200 ◦ C for 1 h in air, followed obtained well-defined alloys with ordered and disordered struc-
by a 300 ◦ C treatment for 1 h in an H2 atmosphere. The ORR tures. The highest ordered structure was achieved with catalysts
performance catalyzed by both as-prepared and heat-treated heat treated at temperatures between 600 and 650 ◦ C for more
Pt–Co catalysts was better than that of commercially available than 5 h. A sample with disordered structure could be obtained
Pt/C and Pt–Co (3:1)/C catalysts. In addition, both heat-treated by heating the ordered sample to 800 ◦ C for 15 min, and then
samples showed an increase in the crystallite size compared to quenching it at a temperature lower than 400 ◦ C. With the same
the as-prepared Pt70 Co30 /C (see Table 1). The electrocatalytic heating time, the mean diameters of the Pt–Co alloy crystallites
activities, measured by MEA testing, of the heat-treated and increased linearly with increasing heat-treatment temperature.
as-prepared samples were almost the same even though the Corrosion tests were performed by exposing these catalysts to a
as-prepared sample had a significantly smaller particles size 210 ◦ C H3 PO4 solution for 0–50 h under a controlled potential
than the heat-treated sample. Furthermore, the first sample heat- of 0.8 V. It was found that Co atoms on the particle surfaces of
treated at 400 ◦ C showed a significant performance loss in PEM both alloy catalysts dissolved easily in the acid media, but the
fuel cell measurements. It is clear that the mutual functioning of disordered alloy was more corrosion resistant. The ORR perfor-
the particle size, alloying, and surface oxides plays a role in cat- mance of cathodes catalyzed by ordered and disordered catalyst
Table 2
Electrochemical characterization for Pt–Cu/C catalysts [70]
Pt–M/C (heat-treatment temperature, ◦ C) Electrochemical surface area (cm2 mg−1) At 700 mV (vs. Ag/AgCl)
Mass activity (mA mg−1 ) Specific activity (mA cm−2 )
Pt/C 679 332 0.49

Pt–Cu/C 423 323 0.76
Pt–Cu/C (300) 414 651 1.33
Pt–Cu/C (600) 333 217 0.65
Pt–Cu/C (900) 241 150 0.62
alloys in 50-h corrosion tests are shown in Fig. 4. It can be seen tions. All Pt–Cu/C catalysts showed a smaller lattice constant
that before the corrosion test, the ordered catalyst exhibits a and hence a shorter average Pt–Pt distance than did the pure
better specific activity (1.35 times higher than that of the disor- Pt catalyst, and heat treatment had negligible effect on these
dered one). However, after the corrosion test, the ordered catalyst parameters. Table 2 shows the catalyst surface areas and the
shows a decay in activity. It was concluded that the disordered ORR activities obtained by cyclic voltammetry for the catalysts
alloy is preferable from the viewpoint of structure stability and before and after heat treatment. It can be clearly seen that the
electrocatalytic activity. addition of copper caused a decrease in catalyst surface area. The
The effect of heat treatment on the structural and activity catalyst heat treated at 300 ◦ C gave the highest ORR activity. The
properties of Pt–Co/C alloy catalysts can also be seen from the authors concluded that the sample treated at 300 ◦ C might favour
work of Salgado et al. [66]. The catalysts were heat treated at O2 adsorption and electron transfer on the Pt surface.
550 and 900 ◦ C under a hydrogen atmosphere for 3 h (Table 1). Beard and Ross [71] prepared Pt–Ti/C catalysts in two solu-
Increases in particle size and contraction of the lattice were tions with different pHs. The as-prepared catalysts then were
observed, suggesting a formation of better alloying with increas- heat treated under a flowing He atmosphere for 2 h at 700,
ing temperature. Through ORR kinetic analysis in a single 900, and 1200 ◦ C. Stability tests were carried out for 6 h in O2 -
PEMFC fed with H2 /O2 , it was found that the sample treated saturated 98% H3 PO4 electrolyte (170 ◦ C) and at a polarization
at 550 ◦ C was more active than that treated at 900 ◦ C, although potential range of 0.6–0.9 V (vs. reversible hydrogen electrode,
its alloying degree was poorer. RHE). It can be seen from Table 1 that with an increase in heat-
Tseng et al. [70] synthesized Pt–Cu/C alloy catalysts and treatment temperature, the particle size increases, resulting in
studied the effects of precursor, preparation methods, and heat an abrupt jump from 4.8 nm at room temperature to 11.0 nm at
treatment on the catalyst particle size, dispersion, lattice struc- 700 ◦ C for catalysts synthesized in the acidic route, and from
ture, and ORR activity (Table 1). Due to the absence of two 7.0 nm at 900 ◦ C to 28.0 nm at 1200 ◦ C for those prepared in
characteristic diffraction peaks in the XRD spectra, the authors the alkaline route. In addition, heat could created superlattice
speculated that all prepared catalysts were disordered phases structures, suggesting the formation of the ordered alloy phase
with a face-centered cubic structure in which the Pt and Cu atoms (Pt3 Ti). The stability test data for the catalysts prepared in the
were randomly distributed at the corner and face-centered posi- acidic route are summarized in Table 3. The results indicate
significant losses of titanium, even in the heat-treated samples.
Therefore, the titanium in the catalyst could not be stabilized by
either alloying or heat treatment.
In an attempt to avoid alloy particle growth, Xiong et al. [67]
prepared carbon-supported Pt–M alloys (M = Fe, Co, Ni and
Cu) using a low-temperature (70 ◦ C) method and investigated
their ORR catalytic activity in a half-cell containing a sulphuric
acid solution and in a PEM fuel cell cathode. Two alloy cata-
lysts, Pt–Co/C and Pt–Fe/C, were separately heat-treated at 200
and 900 ◦ C in an atmosphere of a flowing mixture of Ar + 10%
H2 . They observed that the particle size increased with increas-
Table 3
Stability tests for Pt–Ti/C catalysts in phosphoric acid [71]
Pt–Ti/C (HT ◦ C) Pt/Ti ratio Lattice constant (Å) Superlattice
Before After
As-prepared 42:1 57:1 3.927 No

Fig. 4. The relative specific activities at 900 mV vs. RHE on the ordered and 700 – – 3.928 No
disordered alloy catalysts before and after the corrosion tests for 50 h at 0.8 V 900 – – 3.910 Very weak
and at 210 ◦ C in 105% H3 PO4 . Reproduced from [65] by permission of The 1200 29:1 48:1 3.907 Yes
Electrochemical Society.
Wei et al. [88] demonstrated that heat treatment plays a key

role in obtaining a catalyst with proper corrosion resistance and
a high-surface area. In their work, as-prepared Pt/C and Pt–Fe/C
were heat treated separately at various temperatures from 95 to
900 ◦ C in a flowing gas stream (N2 + 7% H2 ) for 40 min. Their
results are shown in Table 4. It can be seen that the alloying
degree and particle size are strongly dependent on the heat-
treatment temperature. As usual, particle size increases with
increasing the temperature. More interestingly, the Pt alloy cat-
alysts exhibited a better resistance to sintering than did the pure
Pt catalyst if the treatment temperature was higher. In addition,
heat treatment could improve the stability of the catalysts in
terms of dissolution reduction.
Roh et al. [86] prepared a carbon black-supported ternary
Pt–Cu–Fe catalyst with a metal atomic ratio of 2:1:1. This as-
prepared catalyst was then heat treated at 900 ◦ C in a reducing
Fig. 5. Effect of the heating temperature on the extent of atomic ordering (I/If ) atmosphere (N2 + 10% H2 ) for various periods from 0 to 5 h.
for the Pt–Fe/C (I1 1 0 /I1 1 1 ) and Pt–Co/C (I1 0 0 /I1 0 1 ). Re-plotted from [68,87] The results indicated that when the heating time was increased,
by permission of The Royal Society of Chemistry and The Electrochemical
the ordered alloy extent increased accordingly, and the lattice
Society.
structure of Pt changed from a face-centered cubic to a con-
tracted face-centered tetragonal structure. After 2.5 h of heat
ing heat-treatment temperature and time. The ORR activities treatment, the extent of the ordered alloy became larger. With
of these two catalysts were improved by heat treatment at the respect to stability, these electrocatalysts were tested in a 200 ◦ C
lower temperature (200 ◦ C). However, both catalysts treated at H3 PO4 (100%) medium under constant stirring in air for 5 h. It
the higher temperature of 900 ◦ C catalysts showed a dramatic observed that more than 80% of both Cu and Fe in the non-heat-
degradation in ORR activity, which was attributed to the increase treated catalysts was dissolved during this test. In comparison,
in particle size. ordered alloys showed more stability than did the disordered
In a recent work, Xiong and Manthiram [68] synthesized ones. Regarding ORR performance in a fuel cell operation,
several carbon-supported Pt–M alloys by first precipitating the ordered alloys produced by heat treatment showed improved
Mn+ hydroxide (M = Fe, Co, Ni or Cu) on commercially avail- catalytic activity compared to the pure Pt catalyst. The ORR
able Pt/C particles, and then heat treating them at 900 ◦ C in a activity of as-prepared catalyst alloys was lower than that of the
flowing mixture of Ar + 10% H2 . The ORR catalytic activities baseline Pt/C catalyst, while heat-treated alloys showed slightly
were validated in a PEM fuel cell. Pt–Fe/C and Pt–Co/C, con- higher activity at the heat-treatment period of 2.5 h.
trary to other catalysts, showed ordered structures and enhanced In summary, heat-treatment of as-prepared Pt–M/C catalysts
ORR catalytic activity compared to Pt and disordered Pt alloys. increases the catalyst particle size, and at the same time, changes
The effect of heat-treatment temperature on the properties of the catalyst lattice structure from a disordered one to a partially
Pt–Fe/C [68] and Pt–Co/C [87] was studied by heating these ordered one. It seems that the lattice structure change effectively
two catalysts at various temperatures from 500 to 900 ◦ C for improves the ORR catalytic activity and the catalyst stability,
1 h. The results are shown in Fig. 5. It can be seen that although the catalyst active area is reduced by the growth of
heat-treatment temperature of 600 and 650 ◦ C produce the particle size. Therefore, the negative effect of the catalyst particle
best ordering structure and consequently the highest catalytic size increase on the ORR activity may be overwhelmed by the
activity. positive effect of the lattice structure change.
Table 4
Effect of heat-treatment on stability and particle size of powder catalysts [88]
Catalyst Heat-treatment temperature (◦ C) Crystal detected by XRD Dissolution of Pt after ageinga (%) Metal particle size (nm)
Before ageing After ageing
Pt/C 95 Pt 28.7 7.5 12.2

750 2.1 35.8 36.3
900 0.2 46.7 46.7
Pt67 Fe33 /C 500 – 11.4 23.4 30.1
650 Pt, Pt3 Fe 10.4 20.6 21.4
750 Pt3 Fe 9.2 13.5 12.7
900 4.6 22.3 21.9
a 0.3 g immersed in 20 g 105% H3 PO4 , saturated with air at 204 ◦ C, for 5 h and under open-circuit conditions.
catalysis is transition metal N4 -chelates (macrocycles). These

promising cathodic catalysts have shown reasonable activity and
remarkable selectivity towards ORR. They can catalyze a direct
four-electron reduction of O2 to produce water, and are inert
to methanol oxidation [9,11,29]. However, practical catalysts in
PEM fuel cells require strong activity and high stability, major
barriers for this kind of catalyst.
It is generally recognized that the activity and stability of
transition metal macrocycles for ORR catalysis could be signif-
icantly improved by heat treatment. In the past several decades,
heat treatment has been employed to treat carbon-supported
macrocyclic catalysts, and the attained catalysts showed con-
siderable improvement in ORR stability and activity. However,
long-term stability remains a problem. A possible cause might
be the hydrogen peroxide produced in the ORR process, which
Fig. 6. Single scan voltammograms for Pt/C, Pd/C and Pd–Co/C alloys before could attack the catalyst and degrade its performance.
and after heat-treatment at various temperatures in an oxygen saturated 0.5 M Jahnke et al. [96] showed that heat treatment had a signif-
H2 SO4 solutions at 25 ◦ C. Potential scan rate: 5 mV s−1 . Reproduced from [90] icant effect on the ORR catalytic activity of carbon-supported
by permission of Elsevier. CoTAA (dihydrodibenzotetraazaannulene). The results, shown
in Fig. 7, indicate that all of the heat-treated CoTAA samples
3.3. Non-Pt–M/C catalysts showed better ORR performance than did the untreated one. It
seems that the best ORR activity was obtained at 600 ◦ C, and the
Non-Pt alloy catalysts have been investigated for many years best stability was obtained at 800–900 ◦ C. When heat-treatment
in an attempt to overcome the high cost and limited resources temperature goes up over 1000 ◦ C, both stability and activity are
of Pt [89]. Typical examples are palladium (Pd)–M/C catalysts, dropped sharply.
which have the advantage of being methanol-tolerant ORR elec- Among several other works dealing with CoTAA as
trocatalysts [90–93]. an ORR electrocatalyst [97–101], Biloul et al. [98] tested
The effect of heat treatment of Pd–M/C catalysts on the the performance of cobalt (Co) porphyrins (cobalt tetram-
morphology and ORR activity were reported by several groups ethoxyphenylporphyrin (TpOCH3 PPCo) and its derivative
(M = Co [90,93]; Cu [94]; Fe [95]). Recently Zhang et al. [90] (CF3 PPCo)) and CoTAA as ORR electrocatalysts under fuel cell
verified that heat-treatment is necessary to enhance the cat- conditions. All of these macrocycles were vacuum-deposited on
alytic activity towards ORR of a carbon-supported Pd–Co alloy carbon black (Printex XE2) or on active charcoal (Norit SX ultra)
(Fig. 6). The maximum ORR activity was found when the to form supported catalysts. Then the resulting carbon-supported
sample was treated at 300 ◦ C. However, different synthesis pro- catalysts were heat treated at 500, 600 and 800 ◦ C under an
cedures can affect the catalyst properties. Zhang et al. [90] used N2 or Ar atmosphere for 2 h. The effects of supports, catalyst
an impregnation method and carried out heat-treatment with a loadings, and heat-treatment temperatures on the ORR activity
temperature ramping rate of 5 ◦ C min−1 under flowing N2 , at were investigated. According to polarization curves, if the treat-
temperatures ranging from 300 to 700 ◦ C for 2 h, while Wang et ment temperature was the same, the active charcoal-supported
al. [93] prepared the same catalysts via a modified polyol reduc- TpOCH3 PPCo showed a better performance than did the carbon
tion followed by a heat treatment for 2 h at 400–700 ◦ C in an black-supported TpOCH3 PPCo. The optimal CoTAA loading
atmosphere of flowing N2 /10% H2 . Both works observed that
with an increase in temperature, the catalyst lattice contraction,
degree of alloying, and particles size were increased.
A similar effect on morphology and ORR electrocatalytic
activity was observed when Cu [94] or Fe [95] was used as the
second metal and the catalyst was heat treated in the temperature
range of 300–800 ◦ C. When the Pd:Cu ratio in Pd–Cu/C was
1:3, the best treatment temperature in terms of ORR activity
improvement was 600 ◦ C. For Pd–Fe/C, the best temperature
was 500 ◦ C.
4. Effects of heat treatment on non-noble catalysts
4.1. Transition metal macrocycles
Another kind of the fuel cell catalyst which has been exten- Fig. 7. Dependence of thermal activation on temperature (half-cell, electrolyte:
sively explored as a potential replacement for Pt for ORR 4.5N H2 SO4 ). Reproduced from [96] by permission of Springer.
on the active charcoal support was found to be about 13 wt% at treatment was performed in an inert atmosphere at different tem-
800 ◦ C. If other conditions such as catalyst type, loading, and peratures ranging from 150 to 1100 ◦ C for different durations
support type were kept constant, the treated catalyst with the ranging from 0 to 8 h. When the heat-treatment time was short,
best ORR performance was the one treated at 800 ◦ C. To dis- the catalyst activities initially increased with the increase in heat-
criminate the durability of different types of catalysts, a fuel cell treatment temperature, and a peak activity was observed. Among
with different cathode catalysts was tested at a current density of all samples, the best active catalyst was CoTAA/C heat treated
100 mA cm−2 for less than 300 h. The observed durability order at a temperature of 600–700 ◦ C. The metal-free catalyst showed
was CF3 PPCo < TpOCH3 PPCo CoTAA/SX. visible activity only when the heat-treatment temperature was
In an earlier work, Gouèrec et al. [102] observed that the greater than 1000 ◦ C.
optimal heat-treatment temperature of these carbon-supported The positive effect of heat treatment on the ORR stability and
tetraphenylporphyrins for ORR was 600 ◦ C. Fig. 8 shows activity of CoTAA/C was also observed by Gouérec et al. [103].
the electron spin resonance spectra of TpOCH3 PPCo/SX The samples were heat treated under an N2 or Ar atmosphere
and TpCF3 PPCo/SX catalysts before and after the heat- for 2 h at a temperature range of 500–900 ◦ C. For the durability
treatment at 500, 600, and 800 ◦ C. It can be seen that tests, the electrode modified by this heat-treated CoTAA/C was
the untreated samples show the presence of Co porphyrins: polarized at 500 mV (vs. RHE) for 100 h. It was found that the
TpOCH3 PPCo(II)/TpOCH3 PPCo(III)O2 and TPCF3 PPCo(II) 600 ◦ C samples showed the best performance in terms of activity
(Fig. 8a and d, respectively). With an increase in heat-treatment and durability. The authors also showed that Co-involved bond-
temperature, the concentration of Co species in these samples ing was different in these heat-treated catalysts. According to the
decreases (Fig. 8b and e). Interestingly, even there is no visible results of X-ray photoelectron spectroscopy (XPS), the catalyst
Co present in the catalyst (800 ◦ C-treated sample), ORR activity heat treated at 800 ◦ C showed a complete demetallation. In order
can also be observed. to discriminate the Co bonding in these materials, experiments
Franke et al. [101] investigated the effect of heat treatment with time-of-flight secondary ion mass spectrometry (ToF-
on the ORR catalytic activity of three transition metal macrocy- SIMS) were carried out on unsupported catalysts (CoTAA and
cles, CoTAA/C, FeTAA/C, and the metal-free H2 TAA/C. Heat H2 TAA), carbon-supported catalysts (active charcoal-supported
Fig. 8. ESR spectra of TpOCH3 PPCo/SX Ultra on thermal treatment: (a) non-thermally treated; (b) 500 ◦ C treated; (c) 800 ◦ C treated, and TpCF3 PPCo/SX (d)
non-thermally treated; (e) 500 ◦ C treated; (f) 800 ◦ C treated. Reproduced from [102] by permission of Elsevier.
CoTAA (CoTAA/C) and H2 TAA (H2 TAA/C)), which were heat ing heat-treatment temperature. The authors believed that the
treated at different temperatures. On the positive side of the Co metal and/or the Co-containing fragments were responsible
spectrum, pure CoTAA exhibited the parent ion (C18 H14 N4 Co+ ) for catalytic activity enhancement in the temperature range of
and its fragments such as Co+ , C8 H9 N+ , C9 H9 N+ , C7 H4 NCo+ , 600–700 ◦ C.
C11 H9 N2 + , C7 H5 N2 Co+ , and C8 H4 N2 Co+ . On the negative side Since the structure of these N4 -chelate compounds could be
of the spectrum, several ions containing N and/or Co were partially or completely destroyed during pyrolysis at different
detected, including CN− , C2 N− , C3 N− , C2 N2 Co− , C7 H5 N2 − , heat-treatment temperatures, it can be expected that this type of
and CN4 Co− . However, after the CoTAA was deposited on the electrocatalyst could be synthesized using common and inex-
carbon, neither the parent ion nor the characteristic fragments pensive starting materials [1,9,108–111].
could be seen on the positive side of the spectrum. On the Lefèvre et al. prepared a series of Co-based [112] and
negative side of the spectrum, only some fragments could be Fe-based [108–110] ORR electrocatalysts by mixing the corre-
detected, including CN− , C2 N− , C3 N− , CNO− , and C2 N2 Co− . sponding metal precursors. During the synthesis in a gas mixture
The spectra of heat-treated and non-heat-treated CoTAA/C were of NH3 + H2 + Ar, the other experimental conditions, including
similar, except the intensities of ions containing N and Co were ligands (acetate and porphyrin), metal loading (0.2% Co or Fe),
decreased, and the intensities of Co and oxygen associated and support (obtained by pyrolysis of perylene tetracarboxylic
ions such as CoOH+ , CoO− , CoOH− , CoO2 − , CoO2 H− , and dianhydride (PTCDA)), were kept constant. The synthesized cat-
CoO3 − were increased for heat-treated CoTAA/C. At 700 ◦ C, alysts were then heat treated at the temperatures ranging from
for example, neither CoN4 nor CoNx were detected, suggest- 400 to 1000 ◦ C. It was observed that regardless of the precursor
ing a complete decomposition of the macrocycle. However, used, two catalytic sites for Fe-based catalysts were detected:
the intensity of the oxygenated Co fragments was significantly FeN4 /C and FeN2 /C, which were represented by FeN4 C4 + and
enhanced. Due to fact that the optimum ORR activity was found FeN2 C4 + ions, respectively. Their relative weights in the catalyst
at 600 ◦ C for the CoTAA/C catalyst, the detected fragments con- were dependent on the heat-treatment temperature. At tempera-
taining Co with N and/or O suggested that a stronger interaction tures above 700 ◦ C, only FeN2 /C was more abundant, no matter
between the metal in the CoNx with oxygen surface groups on which kind of Fe precursor was used. The correlation between
the support might play a role in catalyst activity enhancement. the catalytic activity and the relative abundance of FeN2 /C at dif-
When the temperature went up to 800 ◦ C, the macrocycle started ferent heat-treatment temperatures suggested that the FeN2 /C
to destruct, a favourable bonding of the nitrogen atoms, and site might be more active than the FeN4 /C site. In contrast to
the oxygen-containing group on the support surface appeared, Fe-based catalysts, Co-based electrocatalysts had no dominant
resulting in a catalyst surface with a high content of Co oxide catalytic site such as CoNx Cy + . The best ORR catalytic activities
species. for both types of catalysts were found to be at the heat-treatment
Weng et al. [104] studied the carbon-supported Co phthalo- temperature range of 500–800 ◦ C.
cyanine (CoPc) heat treated at different temperatures using A positive effect of heat-treatment on the ORR electrocat-
ToF-SIMS coupled with XPS and TEM. The most active ORR alytic activity of nanosized RuNx catalysts was observed by
CoPc/C catalyst was that heat treated at 600 ◦ C under an inert Liu et al. [113]. This catalyst was supported on carbon par-
atmosphere for 2 h [105], as illustrated in Fig. 9. Further stud- ticles using RuCl3 and propylene diammine as the Ru and
ies [106,107] showed that the SIMS intensities of Co+ and all N precursors. The catalysts were then heat treated in an Ar
other Co-containing organic fragments decreased with increas- atmosphere at a temperature range of 600–900 ◦ C. As-prepared
Ru-based chelate did not exhibit an ORR activity, suggesting
that the active catalyst site was not formed. According to the
authors’ observation, a temperature of 700 ◦ C gave the best ORR
activity. As the temperature was increased beyond this point,
the Ru crystallite size was gradually increased, forming cata-
lyst agglomeration, which then negatively affected the catalyst
activity.
In addition, some new heat-treatment techniques for synthe-
sizing highly active macrocycle catalysts by creating favourable
morphology and increasing surface area have been validated. Liu
et al. [29] recently developed an ultrasonic spray pyrolysis (USP)
technique to synthesize CoTMPP/C catalysts. This heat-method
led to spherical, porous, and uniform CoTMPP/C particles with
a high-surface area of 834 m2 g−1 . The ORR activity of the pre-
pared catalyst was double that of the conventional heat-treated
catalyst in an rotating ring-disk electrode (RRDE) measurement.
As shown in Fig. 10, the results of H2 –air PEM single fuel cell
Fig. 9. Evolution of the catalyst electroactivity for the reduction of oxygen as testing showed that the cell performance of the USP-derived cat-
a function of the heat-treatment temperature. Half-cell measurements, H2 SO4 alyst was much higher than that of the conventional heat-treated
solution, pH 0.5. Reproduced from [104] by permission of Elsevier. one.
no further heat treatment is needed [1,124–126]. However,

several works reported a subsequent stage of heat treatment
[123,127,128]. For example, Reeve et al. [123] prepared carbon-
supported Mox Ruy Sz , Rhx Ruy Sz , and Rex Ruy Sz by refluxing the
respective carbonyls with sulphur and carbon black in xylene for
20 h under nitrogen. The products were filtered, washed with
acetone, dried, and finally heat treated at 350 ◦ C for 2 h under
nitrogen.
Regarding the effect of the heat-treatment on these cata-
lysts, Rodrı́guez et al. [127] compared the ORR electrocatalytic
activity of the as-prepared catalyst with the 500 ◦ C heat-treated
carbon-supported Mox Ruy Sez (predominant phase Mo4 Ru2 Se8 )
catalysts. The as-prepared and heat-treated catalysts were all
stable in acidic media, but the untreated sample exhibited much
higher activity compared to the heat-treated one, which might
be attributed to the loss of Se during the heat-treatment. Heat-
Fig. 10. Polarization and power density curves for H2 –air PEM fuel cells with treatment at over 250 ◦ C for unsupported Ru-based catalysts
conventional and ultrasonic spray pyrolysis (USP) derived CoTMPP/C catalysts. could cause the release of carbon monoxide or carbon dioxide,
Reproduced from [29] by permission of Elsevier.
resulting in a lower ORR catalytic activity [126].
Cheng et al. [128] observed that heat treatment under H2 at
Bogdanoff et al. [25] developed another new heat-treatment 400 ◦ C for 1 h improved the ORR catalytic activity and stabil-
technique, the plasma thermal treatment method, to suppress ity of an RuSe electrocatalyst. They observed that increasing
particle aggregation and increase surface area of macrocycle the temperature beyond 400 ◦ C would not lead to better per-
catalysts during the synthesis process. In this way, the electro- formance. According to this study, the positive effect of heat
chemical activity of the prepared catalysts was also effectively treatment on the properties of such catalysts may be attributed
improved. to the following: (1) removal of the residual substances from
With respect to the applications of such transition metal the synthesis and emptying the pores, which may be utilized by
macrocycle catalysts in cathode catalysis, although the ORR new active sites formed from interactions between Ru, S, and C
catalytic activity can be improved by with the assistance of during heat treatment; (2) crystallisation that occurs during heat
techniques such as heat treatment, their stability in a fuel cell treatment, which causes a structural transition from an amor-
operation remains a challenge [11,96,98,103,110,113–117]. phous to a metallic state and reduces the electrical resistance; (3)
reduction of the number of oxygen groups on the support sur-
4.2. Transition metal chalcogenides face, which enhances metal sintering. Such a reduction could
decrease material sintering and, thus, increase the stability of
Transition metal chalcogenides, which have high-catalytic these catalysts. As Ru can easily form Ru oxides in air, Se plays
activity towards ORR and high-methanol tolerance [118–120], an important role in stabilizing the RuSe active site against fur-
represent one type of non-Pt catalyst for fuel cell applications. ther oxidation; and (4) formation of a catalytic site. As heat
Two typical chalcogenides are pseudo-binary selenides contain- treatment has a significant effect on the catalyst composition, it
ing an Ru-substitute (Mo6−x Rux Se8 , where x = 1.8 and 2.3) and has been assumed that the presence of new catalytic sites such
ternary Ni-based tellurides (Ni0.85 Mo6 Te8 ) [119]. The Ru-based as RuSe after heat-treatment can act as an electron reservoir and
compounds were prepared by sintering the mixtures of pure ele- provide the adsorption sites for oxygen. The formation of such
ments in a sealed quartz tube at high temperature (∼1200 ◦ C) an active site is responsible for the higher activity and stability
for several hours. The Ni samples were prepared by reaction of the catalyst, although its real structure has to be identified.
between binary Mo6 Te8 and elemental Ni powder at 1000 ◦ C
[119–121]. It seems that there are few articles focusing on the 5. Conclusions
effect of heat treatment on the ORR electrocatalytic properties
of these materials. This is probably because the synthesis pro- Heat treatment of fuel cell catalysts plays an important
cess is a high-temperature process and further heat-treatment role in the improvement of ORR electrocatalytic activity and
may not be necessary [122]. stability. These catalysts primarily include unsupported and
However, some novel chalcogenide materials have been syn- carbon-supported Pt, Pt alloys, non-Pt alloys, and transition
thesized at low temperatures (<200 ◦ C) by refluxing a metal metal macrocycles. The major effects of heat treatment on cata-
carbonyl compound (such as Mo(CO)6 , Ru3 (CO)12 , Os3 (CO)12 ) lyst properties, including particle size, morphology, dispersion
and the corresponding chalcogen. The temperatures used in the of the metal on the support, alloying degree, active site forma-
synthesis were limited by the boiling point of the organic solvent tion, catalytic activity, and catalytic stability, were reviewed in
used. This faster, simple, and economic low-temperature proce- this paper. The optimum heat-treatment temperature and time
dure can produce cluster-like materials, which have improved period with respect to ORR activity and stability improvement
electroactivity and chemical stability [122,123]. In most cases are strongly dependent on the individual catalyst.
With respect to Pt-based catalysts, heat treatment can induce [11] P. Vasudevan, Santosh, M. Neelam, S. Tyagi, Transition Met. Chem. 15
particle-size growth, better alloying degree, and changes in the (1990) 81–90.
catalyst surface morphology from amorphous to more ordered [12] J. Luo, N. Kariuki, L. Han, L. Wang, C.J. Zhong, T. He, Electrochim.
Acta 51 (2006) 4821–4827.
states, which have remarkable effects on ORR activity and sta- [13] P.L. Antonucci, V. Alderucci, N. Giordano, D.L. Cocke, H. Kim, J. Appl.
bility. In addition, heat treatment of catalyst carbon supports can Electrochem. 24 (1994) 58–65.
also affect the ORR catalytic activity of the supported catalyst [14] S.J. Mukerjee, App. Electrochem. 20 (1990) 537–548.
significantly. However, if the heat-treatment temperature is too [15] A. Elzing, A. van Der Putten, W. Visscher, E. Barendrecht, J. Electroanal.
high (>1000 ◦ C), even though the stability of the catalysts is Chem. 200 (1986) 313–322.
[16] F. Coloma, A.S. Escribano, J.L.G. Fierro, F. Rodrı́guez-Reinoso, Lang-
improved, the catalytic activity could be degraded. muir 10 (1994) 750–755.
Heat treatment is also important for ORR activity and [17] J.H. Tian, F.B. Wang, Z.Q. Shan, R.J. Wang, J.Y. Zhang, J. Appl. Elec-
stability improvement of non-noble catalysts, in particular tran- trochem. 34 (2004) 461–467.
sition metal macrocycles. In fact, heat treatment is necessary [18] G. Hinds, Preparation and characterisation of PEM fuel cell electrocata-
for introducing more active catalytic sites. However, if the lysts: a review, NPL Report, United Kingdom, 2005, pp. 10–12.
[19] M. Kang, Y.S. Bae, C.H. Lee, Carbon 43 (2005) 1512–1516.
temperature is greater than about 800 ◦ C, it could have a neg- [20] M. Mazurek, N. Benker, C. Roth, H. Fuess, Fuel Cells 6 (2006) 208–
ative impact on the catalytic properties of these materials. 213.
For metal chalcogenide catalysts, it seems that heat treat- [21] H. Cheng, W. Yan, K. Scott, Fuel Cells 7 (2007) 16–20.
ment may not be necessary for catalytic activity and stability [22] R. Minami, Y. Kitamoto, T. Chikata, S. Kato, Electrochim. Acta 51 (2005)
improvement. 864–866.
[23] M. Nie, P.K. Shen, M. Wu, Z. Wei, H. Meng, J. Power Sources 162 (2006)
The mechanisms of heat treatment, including induced cata- 173–176.
lyst surface structure change and surface active site formation, [24] Z.F. Ma, X.Y. Xie, X.X. Ma, D.Y. Zhang, Q. Ren, N. Heß-Mohr, V.M.
seem complicated. More fundamental research is necessary in Schmidt, Electrochem. Commun. 8 (2006) 389–394.
order understand these mechanisms and to develop new strate- [25] P. Bogdanoff, I. Herrmann, M. Hilgendorff, I. Dorbandt, S. Fiechter, H.
gies that include heat-treatment processes for enhancement in Tributsch, J. New Mater. Electrochem. Sys. 7 (2004) 85–92.
[26] I. Herrmann, V. Bruser, S. Fiechter, H. Kersten, P. Bogdanoff, J. Elec-
fuel cell catalyst activity and stability. trochem. Soc. 152 (2005) A2179–A2185.
[27] N.A. Savastenko, V. Brüser, M. Brüser, K. Anklam, S. Kutschera, H.
Acknowledgements Steffen, A. Schmuhl, J. Power Sources 165 (2007) 24–33.
[28] H. Shioyama, K. Honjo, M. Kiuchi, Y. Yamada, A.I. Ueda, N. Kuriyama,
T. Kobayashi, J. Power Sources 161 (2006) 836–838.
The authors thank the financial support of the Institute [29] H. Liu, C. Song, Y. Tang, J. Zhang, J. Zhang, Electrochim. Acta 52 (2007)
for Fuel Cell Innovation, National Research Council Canada 4532–4538.
(NRC-IFCI). Professor Cicero W.B. Bezerra (Grant holder [30] E. Antolini, J. Mater. Sci. 38 (2003) 2995–3005.
from CNPq-Brazil) also wants to thank the financial support [31] M.M. Shaijumon, S. Ramaprabhu, N. Rajalakshmi, Appl. Phys. Lett. 88
(2006) 253105.
from Conselho Nacional de Desenvolvimento Cientı́fico e Tec-
[32] M. Peuckert, T. Yoneda, R.A.D. Betta, M. Boudart, J. Electrochem. Soc.
nológico, Brazil (Proc. 202353/2006-0). The authors are also 133 (1986) 944–947.
grateful to the CNPq (ETACOMB Proc. no. 505167/2004-2) [33] Y. Takasy, N. Oshashi, X.G. Zhang, Y. Murakami, H. Minagawa, S. Sato,
and FINEP (DEFCCA Conv. no. 520186-2005-2) for financial K. Yahikozawa, Electrochim. Acta 41 (1996) 2595–2600.
support. [34] M. Watanabe, H. Sei, P. Stonehart, J. Electroanal. Chem. 261 (1989)
375–387.
[35] T.J. Schmidt, V. Stamenkovic, M. Arenz, N.M. Markovic, P.N. Ross Jr.,
References Electrochim. Acta 47 (2002) 3765–3776.
[36] K. Kinoshita, J. Electrochem. Soc. 137 (1990) 845–847.
[1] L. Zhang, J. Zhang, D.P. Wilkson, H. Wang, J. Power Sources 156 (2006) [37] H.Y. Yano, J. Inukai, H. Uchida, M. Watanabe, P.K. Babu, T. Kobayashi,
171–182. J.H. Chung, E. Oldfield, A. Wieckowski, Phys. Chem. Chem. Phys. 8
[2] H. Liu, C. Song, L. Zhang, J. Zhang, H. Wang, D.P. Wilkinson, J. Power (2006) 4932–4939.
Sources 155 (2006) 95–110. [38] A. Kabbabi, F. Gloaguen, F. Andolfatto, R. Durand, J. Electroanal. Chem.
[3] K. Lee, J. Zhang, H. Wang, D.P. Wilkinson, J. Appl. Electrochem. 36 373 (1994) 251–254.
(2006) 507–522. [39] M.L. Sattler, P.N. Ross, Ultramicroscopy 20 (1986) 21–28.
[4] R. Bashyam, P. Zelenay, Nature 443 (2006) 63–66. [40] J. Bett, J. Lundquist, E. Washinton, P. Stonehart, Electrochim. Acta 18
[5] A. Guha, W. Lu, T.A. Zawodzinski, D.A. Schiraldi, Carbon 45 (2007) (1973) 343–348.
1506–1517. [41] E. Antolini, F. Cardenilli, E. Giacomerri, G. Squadrito, J. Mater. Sci. 37
[6] S. Mukerjee, S. Srinivasan, in: W. Vielstich, A. Lamm, H.A. Gasteiger (2002) 133–139.
(Eds.), Handbook of Fuel Cells: Fundamentals, Technology and Appli- [42] A.H. Alexopoulos, A.I. Roussos, C.K. Kiparissides, Chem. Eng. Sci. 59
cations. vol. 2. Electrocatalysis, Wiley, Chicester, 2003, pp. 502– (2004) 5751–5769.
519. [43] D.S. Cameron, S.J. Cooper, I.L. Dodgson, B. Harrison, J.W. Jenkins,
[7] M. Gattrell, B. MacDougall, in: W. Vielstich, A. Lamm, H.A. Gusteiger Catal. Today 7 (1990) 113–137.
(Eds.), Handbook of Fuel Cells: Fundamentals, Technology and Appli- [44] S. Kim, S.J. Park, Electrochim. Acta 52 (2007) 3013–3021.
cations. vol. 2. Electrocatalysis, Wiley, Chicester, 2003, pp. 443–464. [45] E. Auer, A. Freund, J. Pietsch, T. Tacke, Appl. Catal. A: Gen. 173 (1998)
[8] T. He, E. Kreidler, L. Xiong, J. Luo, C.J. Zhong, J. Electrochem. Soc. 153 259–271.
(2006) A1637–A1643. [46] G. Wang, G. Sun, Z. Zhou, J. Liu, Q. Wang, S. Wang, J. Guo, S. Yang,
[9] B. Wang, J. Power Sources 152 (2005) 1–15. Q. Xin, B. Yi, Electrochem. Solid State Lett. 8 (2005) A12–A16.
[10] K.S. Han, Y.S. Moon, O.H. Han, K.J. Hwang, I. Kim, H. Kim, Elec- [47] C. Prado-Burguete, A. Linares-Solano, F. Rodrı́guez-Reinoso, C.S.-M.
trochem. Commun. 9 (2007) 317–324. Lecea, J. Catal. 115 (1989) 98–106.
[48] C. Prado-Burguete, A. Linares-Solano, F. Rodrı́guez-Reinoso, C.S.-M. [87] L. Xiong, A. Manthiram, J. Mater. Chem. 14 (2004) 1454–1460.
Lecea, J. Catal. 128 (1991) 397–404. [88] Z. Wei, H. Guo, Z. Tang, J. Power Sources 62 (1996) 233–236.
[49] M.C. Roman-Martı́nez, D. Cazorla-Amorós, A. Linares-Solano, F. [89] V. Raghuveer, P.J. Ferreira, A. Manthiram, Electrochem. Commun. 8
Rodrı́guez-Reinoso, C.S.-M. Lecea, H. Yamashita, M. Anpo, Carbon 33 (2006) 807–814.
(1995) 3–13. [90] L. Zhang, K. Lee, J. Zhang, Electrochim. Acta 52 (2007) 3088–3094.
[50] T. Torre, A.S. Aricó, V. Alderucci, V. Antonucci, N. Giordano, Appl. [91] J.L. Fernandez, V. Raghuveer, A. Manthiram, A.J. Bard, J. Am. Chem.
Catal. A: Gen. 114 (1994) 257–272. Soc. 127 (2005) 13100–13101.
[51] A.S. Aricò, V. Antonucci, L. Pino, P.L. Antonucci, N. Giordano, Carbon [92] V. Raghuveer, A. Manthiram, A.J. Bard, J. Phys. Chem. B 109 (2005)
28 (1990) 599–609. 22909–22912.
[52] G.A. Gruver, J. Eletrochem. Soc. 125 (1978) 1719–1720. [93] W. Wang, D. Zheng, C. Du, Z. Zou, X. Zhang, B. Xia, H. Yang, D.L.
[53] P. Stonehart, Carbon 22 (1984) 423–431. Akins, J. Power Sources 167 (2007) 243–249.
[54] M. Uchida, Y. Aoyama, M. Tanabe, N. Yanagihara, N. Eda, A. Ohta, J. [94] D. Myers, Proceedings of DOE Hydrogen Program Review, Arlington,
Electrochem. Soc. 142 (1995) 2572–2576. VA, 2007.
[55] K.H. Kangasniemi, D.A. Condit, T.D. Jarvi, J. Electrochem. Soc. 151 [95] M.H. Shao, K. Sasaki, R.R. Adzic, J. Am. Chem. Soc. 128 (2006)
(2004) E125–E132. 3526–3527.
[56] R. Schlögl, in: G. Ertl, H. Knözinger, J. Weifkamp (Eds.), Handbook [96] H. Jahnke, M. Schönborn, G. Zimmermann, Top. Curr. Chem. 61 (1976)
of Heterogeneous Catalysis, vol. 1, Wiley–VCH, Weinheim, 1997, pp. 133–181.
138–191. [97] K. Wiesener, G. Grünig, J. Electroanal. Chem. 180 (1984) 639–643.
[57] M. Kijima, D. Fujiya, T. Oda, M. Ito, J. Therm. Anal. Calorim. 81 (2005) [98] A. Biloul, P. Gouèrec, M. Savy, G. Scarbeck, S. Besse, J. Riga, J. Appl.
549–554. Electrochem. 26 (1996) 1139–1146.
[58] M. Kijima, H. Tanimoto, K. Takakura, D. Fujiya, Y. Ayuta, K. Matsuishi, [99] A. van Der Putten, A. Elzing, W. Visseher, E. Barendrecht, J. Electroanal.
Carbon 45 (2007) 594–601. Chem. 205 (1986) 233–244.
[59] E. Antolini, Mater. Chem. Phys. 78 (2003) 563–573. [100] A. van Der Putten, A. Elzing, W. Visseher, E. Barendrecht, J. Electroanal.
[60] E. Antolini, J.R.C. Salgado, M.J. Giz, E.R. Gonzalez, Int. J. Hydrogen Chem. 221 (1987) 95–104.
Energy 30 (2005) 1213–1220. [101] R. Franke, D. Ohms, K. Wiesener, J. Electroanal. Chem. 260 (1989)
[61] E. Antolini, R.R. Passos, E.A. Ticianelli, Electrochim. Acta 48 (2002) 63–73.
263–270. [102] P. Gouèrec, A. Biloul, O. Contamin, G. Scarbeck, M. Savy, J.M. Barbe,
[62] G. Cambanis, D. Chadwick, Appl. Catal. 25 (1986) 191–198. R. Guilard, J. Electroanal. Chem. 398 (1995) 67–75.
[63] M. Min, J. Cho, K. Cho, H. Kim, Electrochim. Acta 45 (2000) 4211– [103] P. Gouèrec, A. Biloul, O. Contamin, G. Scarbeck, M. Savy, J. Riga, L.T.
4217. Weng, P. Bertand, J. Electroanal. Chem. 422 (1997) 61–75.
[64] B.C. Beard, P.N. Ross Jr., J. Electrochem. Soc. 137 (1990) 3368–3374. [104] L.T. Weng, P. Bertrand, G. Lalande, D. Guay, J.P. Dodelet, Appl. Surf.
[65] M. Watanabe, K. Tsurumi, T. Nakamura, P. Stonehart, J. Electrochem. Sci. 84 (1995) 9–21.
Soc. 141 (1994) 2659–2668. [105] G. Tamizhamani, J.P. Dodelet, D. Guay, G. Laland, G.A. Capuano, J.
[66] J.R.C. Salgado, E. Antolini, E.R. Gonzalez, J. Phys. Chem. B 108 (2004) Electrochem. Soc. 141 (1994) 41–45.
17767–17774. [106] M. Ladouceur, G. Lalande, G. Guay, J.P. Dodelet, L. Dignard-Bailey,
[67] L. Xiong, A.M. Kannan, A. Manthiram, Electrochem. Commun. 4 (2002) M.L. Trudeau, R. Schulz, J. Eletrochem. Soc. 140 (1993) 1974–
898–903. 1981.
[68] L. Xiong, A. Manthiram, J. Electrochem. Soc. 152 (2005) A697– [107] M.C.M. Alves, J.P. Dodelet, G. Guay, M. Ladouceur, G. Tourillon, J.
A703. Phys. Chem. 92 (1992) 10898–10905.
[69] E.I. Santiago, L.C. Varanda, H.M. Villullas, J. Phys. Chem. C 111 (2007) [108] M. Lefèvre, J.P. Dodelet, P. Bertrand, J. Phys. Chem. B 104 (2000)
3146–3151. 11238–11247.
[70] C.J. Tseng, S.T. Lo, S.C. Lo, P.P. Chu, Mater. Chem. Phys. 100 (2006) [109] M. Lefèvre, J.P. Dodelet, P. Bertrand, J. Phys. Chem. B 106 (2002)
385–390. 8705–8713.
[71] B.C. Beard, P.N. Ross Jr., J. Eletrochem. Soc. 133 (1986) 1839–1845. [110] M. Lefèvre, J.P. Dodelet, Electrochim. Acta 48 (2003) 2749–2760.
[72] H. Li, G. Sun, N. Li, S. Sun, D. Su, Q. Xin, J. Phys. Chem. C 111 (2007) [111] M. Bron, J. Radnik, M.F. Erdmann, P. Bogdadoff, S. Fiechter, J. Elec-
5605–5617. troanal. Chem. 535 (2002) 113–119.
[73] J. Liu, M.M. Maye, V. Petkov, N.N. Kariuki, L. Wang, P. Njoki, D. Mott, [112] M. Lefèvre, J.P. Dodelet, P. Bertrand, J. Phys. Chem. B 109 (2005)
Y. Lin, C.J. Zhong, Chem. Mater. 17 (2005) 3086–3091. 16718–16724.
[74] A.R. Denton, N.W. Ashcroft, Phys. Rev. A 43 (1991) 3161–3164. [113] L. Liu, H. Kim, J.W. Lee, B.N. Popov, J. Electrochem. Soc. 154 (2007)
[75] E. Antolini, J.R.C. Salgado, E.R. Gonzalez, J. Power Sources 160 (2006) A123–A128.
957–968. [114] Z.F. Ma, X.Y. Xie, X.X. Ma, D.Y. Zhang, Q. Ren, N.H. Mohr, V.M.
[76] V. Jalan, E.J. Taylor, J. Electrochem. Soc. 130 (1983) 2299–2302. Schmidt, Electrochem. Commun. 8 (2006) 389–394.
[77] S. Mukerjee, S. Srinivasan, M.P. Soriaga, J. McBreen, J. Electrochem. [115] G. Lalande, G. Faubert, R. Côté, D. Guay, J.P. Dodelet, L.T. Weng, P.
Soc. 142 (1995) 1409–1422. Bertrand, J. Power Sources 61 (1996) 227–237.
[78] H.R. Colón-Mercado, B.N. Popov, J. Power Sources 155 (2006) 253– [116] G. Faubert, R. Côté, D. Guay, J.P. Dodelet, G. Dénès, P. Bertrand, Elec-
263. trochim. Acta 43 (1998) 345–353.
[79] M.T. Paffett, J.G. Beery, S. Gottesfeld, J. Electrochem. Soc. 135 (1988) [117] S. Baraton, C. Coutanceau, C. Roux, F. Hahn, J.M. Léger, J. Electroanal.
1431–1436. Chem. 577 (2005) 223–234.
[80] H. Yang, W. Vogel, C. Lamy, N.C. Alonso-Vante, J. Phys. Chem. B 108 [118] N. Alonso-Vante, H. Tributsch, Nature 323 (1986) 431–432.
(2004) 11024–11034. [119] N. Alonso-Vante, W. Jaegermann, H. Tributsch, W. Hönle, K. Yvon, J.
[81] K. Lee, O. Savadogo, A. Ishihara, S. Mitsushima, N. Kamiya, K.-I. Ota, Am. Chem. Soc. 109 (1987) 3251–3257.
J. Electrochem. Soc. 153 (1) (2006) A20–A24. [120] N. Alonso-Vante, B. Schubert, H. Tributsch, A. Perrin, J. Catal. 112 (1988)
[82] Y. Wang, P.B. Balbuena, J. Phys. Chem. B 109 (2005) 18902–18906. 384–391.
[83] B. Hammer, J.K. Norskov, Adv. Catal. 45 (2000) 71–129. [121] C. Fischer, N. Alonso-Vante, S. Fiechter, H. Tributsch, J. Appl. Elec-
[84] J.L. Fernandez, D.A. Walsh, A.J. Bard, J. Am. Chem. Soc. 127 (2005) trochem. 25 (1995) 1004–1008.
357–365. [122] N. Alonso-Vante, in: W. Vielstich, A. Lamm, H.A. Gusteiger (Eds.),
[85] Y. Suo, L. Zhuang, J. Lu, Angew. Chem. Int. Ed. 46 (2007) 2862–2864. Handbook of Fuel Cells: Fundamentals, Technology and Applications.
[86] W. Roh, J. Cho, H. Kim, J. Appl. Electrochem. 26 (1996) 623–630. vol. 2. Electrocatalysis, Wiley, Chicester, 2003, pp. 534–543.
[123] R.W. Reeve, P.A. Christensen, A.J. Dicknson, A. Hamnett, K. Scott, [126] H. Tributsch, M. Bron, M. Hilgendorff, H. Schulenburg, I. Dorbandt,
Electrochim. Acta 45 (2000) 4237–4250. V. Eyert, P. Bogdanoff, S. Fiechter, J. Appl. Electrochem. 31 (2001)
[124] O. Solorz-Feria, K. Ellmer, M. Giersig, N. Alonso-Vante, Electrochim. 739–748.
Acta 39 (1994) 1647–1653. [127] F.J. Rodrı́guez, P.J. Sebastian, O. Solorza, R. Pérez, Int. J. Hydrogen
[125] V.L. Rhun, N. Alonso-Vante, J. New. Mater. Electrochem. Syst. 3 (2000) Energy 23 (1998) 1031–1035.
333–383. [128] H. Cheng, W. Yuan, K. Scott, Electrochim. Acta 52 (2006) 466–473.

Bezerra2007 PDF

Uploaded by

Copyright:

Available Formats

Bezerra2007 PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Bezerra2007 PDF

Uploaded by

Copyright:

Available Formats

Available online at www.sciencedirect.

Journal of Power Sources 173 (2007) 891–908

A review of heat-treatment effects on activity and stability of

∗ Corresponding author. Tel.: +1 604 221 3087.

4.2. Transition metal chalcogenides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 905

1. Introduction Regarding ORR catalyst synthesis, heat-treatment has been

2. Effects of heat treatment on carbon-supported Pt resulting in well-defined cubo-octahedron crystallites, which

functional groups of the support and (2) changes in Pt dispersion,

V (1:1) As-prepared – 3.4 3.8834 – 22 [61]

Surface area measured by CO chemisorption.

Cambanis and Chadwick [62] prepared heat-treated Pt–V/C

Mass activity (mA mg−1 ) Specific activity (mA cm−2 )

Pt/C 679 332 0.49

Pt–Ti/C (HT ◦ C) Pt/Ti ratio Lattice constant (Å) Superlattice

As-prepared 42:1 57:1 3.927 No

Wei et al. [88] demonstrated that heat treatment plays a key

Before ageing After ageing

Pt/C 95 Pt 28.7 7.5 12.2

catalysis is transition metal N4 -chelates (macrocycles). These

4. Effects of heat treatment on non-noble catalysts

4.1. Transition metal macrocycles

no further heat treatment is needed [1,124–126]. However,

You might also like