Production of CO-free H2 from formic acid. A comparative study of the catalytic behavior of Pt metals on a carbon support

Article abstract of DOI:10.1016/j.jcat.2011.01.023

The vapor-phase decomposition of formic acid was investigated over Pt metals supported on inert carbon Norit with the aim of producing CO-free H ₂. FTIR spectroscopic studies revealed that formic acid dissociated on Pt metals at 220-240 K, but the formate species formed was stable only below 300-350 K. Decomposition of formic acid started at and above 350 K on all catalysts and was complete at 473-523 K. Kinetic studies on Pt/Norit demonstrated that the decomposition was a zero-order process with an activation energy of 70.7 kJ/mol. Although dehydrogenation was the predominant process at lower temperatures, CO-free H₂ was not produced on any catalyst. The highest selectivity of 98.3-99% for H₂ formation was attained on Ir/Norit. For all catalysts, the selectivity was improved considerably by the addition of water to formic acid. The production of CO-free H₂ was achieved only on supported Ir at 383-473 K. Similarly, as regards the H ₂ yield the outstanding catalyst was Ir/Norit, followed by Pt/Norit.

Full text of DOI:10.1016/j.jcat.2011.01.023

Journal of Catalysis 279 (2011) 213–219
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Production of CO-free H2 from formic acid. A comparative study
of the catalytic behavior of Pt metals on a carbon support
⇑
F. Solymosi , Á. Koós, N. Liliom, I. Ugrai
Reaction Kinetics Research Group, Chemical Research Centre of the Hungarian Academy of Sciences, Department of Physical Chemistry and Materials Science,
University of Szeged, P.O. Box 168, H-6701 Szeged, Hungary
a r t i c l e i n f o
a b s t r a c t
Article history:
The vapor-phase decomposition of formic acid was investigated over Pt metals supported on inert carbon
Norit with the aim of producing CO-free H2. FTIR spectroscopic studies revealed that formic acid dissoci-
ated on Pt metals at 220–240 K, but the formate species formed was stable only below 300–350 K.
Decomposition of formic acid started at and above 350 K on all catalysts and was complete at 473–
523 K. Kinetic studies on Pt/Norit demonstrated that the decomposition was a zero-order process with
an activation energy of 70.7 kJ/mol. Although dehydrogenation was the predominant process at lower
temperatures, CO-free H2 was not produced on any catalyst. The highest selectivity of 98.3–99% for H2
formation was attained on Ir/Norit. For all catalysts, the selectivity was improved considerably by the
addition of water to formic acid. The production of CO-free H2 was achieved only on supported Ir at
383–473 K. Similarly, as regards the H2 yield the outstanding catalyst was Ir/Norit, followed by Pt/Norit.
Ó 2011 Elsevier Inc. All rights reserved.
Received 17 December 2010
Revised 21 January 2011
Accepted 21 January 2011
Available online 24 February 2011
Keywords:
Formic acid decomposition
Reforming of formic acid
Production of CO-free H2
Pt metals
Carbon support
1. Introduction
Renewed interest in the catalytic chemistry of formic acid
[24–27] resulted from the observations that formate species are
Great efforts have been made during the last decade to develop
an efficient catalytic process for the production of H2 [1–3]. The
most frequently used source is an alcohol, and particularly ethyl
alcohol, but elimination of large amount of CO and CxHyO com-
pounds formed causes great difficulty, not to mention the poison-
ing effect of acetate [4–6]. Unfortunately, the decomposition of CH4
and other hydrocarbons is less applicable due to carbon deposition
and early deactivation of the catalyst [7–9]. In the field of hetero-
geneous catalysis, formic acid as a source of H2 has received atten-
tion only recently [10–13], although it has been considered as a
possible basic material for the production of H2 for fuel cells
[14,15]. Note that formic acid may be obtained from biomass con-
version. The decomposition of formic acid was widely used as a
model reaction in the 1950s and 1960s to test the roles of the elec-
tronic properties of metals, alloys and pure and doped oxides in
heterogeneous catalysis [16–23]. It was demonstrated that the
electronic properties of both the metals and the oxides play
an important part in the catalysis of formic acid decomposition
[15–23]. The study of this reaction provided the first convincing
evidence of the importance of the electronic interactions between
Ni and semiconducting supports (TiO2, NiO and Cr2O3) in carrier
effect [19,20].
important reaction intermediates in several catalytic processes,
e.g. CH3OH synthesis [28] and reforming [29], the water–gas shift
reaction [30] and the hydrogenation of CO2 [31–35] and CO
[36,37]. This also initiated great interest in the study of the inter-
actions of formic acid with metal single crystals on an atomic scale
[38–40]. In addition, great number of papers dealt with the decom-
position of formic acid in solution using various metal complexes.
The results are well summarized in some excellent papers and
reviews [41–44].
We recently found that Mo2C prepared by the reaction of MoO3
with multiwall carbon and carbon Norit, which exhibited high
activities in the reactions of C2H5OH, CH3OH and (CH3)2O [45], is
also an efficient catalyst for the decomposition and reforming of
formic acid [12]. At certain temperatures, CO-free H2 can be pro-
duced in a yield of 99–100%. In the present paper, we examine
the catalytic performances of Pt metals, mainly supported on car-
bon Norit, in the reactions of formic acid, with the aim of identifying
the most active metal and establishing experimental conditions un-
der which H2 can be produced in high yield and virtually free of CO.
2. Experimental
Carbon Norit (Alfa Aesar, ROW 0.8 mm pellets, Steam activated)
was used as a support, which was purified by treatment with HCl
(10%) for 12 h at room temperature. Afterward it was washed to
have a Cl-free sample. After this treatment, the metal impurities,
⇑
Corresponding author. Fax: +36 62 544 106.
E-mail address: fsolym@chem.u-szeged.hu (F. Solymosi).
0021-9517/$ - see front matter Ó 2011 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2011.01.023

214
F. Solymosi et al. / Journal of Catalysis 279 (2011) 213–219
mainly Fe, determined by the ICP-AES method, amounted to less
than 0.002%. The surface area of the purified sample is 859 m2/g.
The supported catalysts were prepared by impregnating the carbon
Norit with solutions of metal compounds to yield a nominal 2 wt%
metal. The following salts of Pt metals were used: H2PtCl5ꢀ6H2O,
PdCl2, RhCl3ꢀ3H2O, H2IrCl6 and RuCl3ꢀ3H2O. The impregnated pow-
ders were dried at 383 K. The fragments of catalyst pellets were oxi-
dized at 673 K and reduced at 673 K for 1 h in situ. In a certain case
SiO2 (CAB-O-SIL; 198 m2/g) and Al2O3 (Degussa, P110, 100 m2/g)
were also used as a support. HCOOH was the product of BDH, with
a purity of 99.5%. Other gases were of commercial purity (Linde).
The dispersion of metals was determined by the adsorption of
H2 at room temperature. Thermal desorption measurements
(TPD) were carried out in the catalytic reactor. The catalysts were
treated with HCOOH/Ar containing ꢁ6% HCOOH at ꢁ300 K for
60 min, and flushed with Ar for 30 min. The TPD were carried out
in Ar flow (20 ml/min) with ramp at 2 K/min from 300 K to
ꢁ700 K. Desorbing products were analyzed by gas chromatogra-
phy. Fourier Transformed Infra Red (FTIR) spectra were recorded
with a BioRad FTS-155 spectrometer with a wavenumber accuracy
of 4 cmꢂ1. Catalytic reactions were carried out at a pressure of
1 atm in a fixed-bed, continuous flow reactor consisting of a quartz
tube (8 mm id) connected to a capillary tube [12]. The flow rate
was 40 ml/min. The carrier gas was Ar, which was bubbled through
the formic acid at room temperature: its content was ꢁ5–6.0%. In
general, 0.3 g of loosely compressed catalyst sample was used.
The reaction products were analyzed with a HP 4890 gas chro-
matograph equipped with PORAPAQ Q+S and 30-m long HP-PLOT
Al2O3 column. The conversion of formic acid was determined by
taking into account the amount consumed. The selectivity of
hydrogen was calculated from the ratio of CO2 concentration to
the sum of CO2 + CO. Multiplying this value with the conversion
gave the hydrogen yield.
Fig. 1A depicts the IR spectra of formic acid adsorbed on Ir/SiO2
at ꢁ200 K and heated to different temperatures under continuous
degassing. In the CAH stretching region, absorption bands were
observed at 2942, 2920 and 2847 cmꢂ1
, with intensities
which gradually attenuated as the temperature was raised. In the
low-frequency range, a very intense spectral feature appeared
at 1722 cmꢂ1, and less intense one at 1391 cmꢂ1. At around
225 K, weak new bands started to develop at 1586, 1369
and 1270 cmꢂ1. The intensity of the 1586 cmꢂ1 increased up to
253–273 K, but then attenuated, and it disappeared at around
290–300 K. The 1722 cmꢂ1 band also underwent attenuation, but
remained detectable up to 300 K. Analogous features were found
on Rh/SiO2 (Fig. 1B). In this case the band at 1590 cmꢂ1 was some-
what more stable and could be detected even at ꢁ473 K. Similar
spectral features were likewise observed on the other SiO2-
supported metals. The position of the absorption bands is given
in Table 1.
The interaction of formic acid with Ir/Al2O3 resulted in a differ-
ent picture. In this case, it was not necessary to perform low-
temperature experiments, as the adsorption of formic acid even at
300 K produced very strong bands at 1597, 1394 and 1381 cmꢂ1
,
and a less intense one at 1714 cmꢂ1. The latter gradually attenu-
ated at higher temperature and disappeared at 273–300 K. In con-
trast with the results observed on SiO2-supported metals, the
1597 cmꢂ1 band exhibited high thermal stability: it was elimi-
nated only above 573 K (Fig. 2). The adsorption of formic acid on
Table 1
Vibrational frequencies (in cmꢂ1) observed following the adsorption of HCOOH and
their assignment.
Assignment
Ir/SiO2
Rh/SiO2
Ir/Al2O3
m
m
m
m
m
m
m
m
m
(CH)
(CH)
(CH) (formate)
(CO) (formic acid)
2942
2920
2847
1722
1586
1391
1369
1270
2944
2917
2863
1723
1590
1375
1310
2933
2913
3. Results
1714
1597
1394
1381
1326
1079
a(OCO)
(OCO) (formic acid and formate)
s(OCO)
(CO)
3.1. Infrared spectroscopic measurements
As Pt metals on a carbon support are not transparent, IR mea-
surements were first performed with silica-supported metals.
(CH)
Fig. 1. FTIR spectra of Ir/Norit (A) and Rh/Norit (B) following the adsorption of HCOOH at ꢁ200 K and after subsequent degassing at different temperatures.

F. Solymosi et al. / Journal of Catalysis 279 (2011) 213–219
215
Ir/Norit is characterized by the following peak temperatures: H2
(Tp = 400 K) and CO (Tp = 375 K). From the pure Norit support, we
registered the desorption of formic acid with Tp ꢃ 315 K, without
the formation of H2, CO or CO2. In the case of Ir/SiO2 only the
desorption of
a very small amount of H2 was observed at
Tp ꢁ 430 K. From Ir/Al2O3, larger amounts of H2 and CO2 with
Tp = 478 K, and a smaller quantity of CO with Tp = 500 K were
released.
3.3. Catalytic measurements
3.3.1. Decomposition of formic acid
Fig. 4 illustrates the conversion in the decomposition of formic
acid and the selectivity of H2 formation on the carbon-supported
metals as a function of temperature. The reaction began at and
above 350 K on all samples and was complete at 473–523 K. The
main process was the dehydrogenation of formic acid to H2 and
CO2. CO was formed in only 2–7% at lower temperatures, but its
amount gradually increased at higher temperatures.
As concerns the catalytic behavior of the Pt metals, the decom-
position occurred to a great extent on Ir/Norit. The selectivity for
H2 production was 98–99% at 423–523 K, which diminished to
ꢁ93% at 623 K. The decomposition rate was lower on Pt/Norit. In
this case, the selectivity for H2 at 423–473 K was 98–99%, which
decreased markedly at higher temperatures. Lower conversions
were observed on Rh/Norit. The selectivity for H2 formation was
ꢁ91.8% at 423 K, when the conversion was only ꢁ40%, and the
selectivity was only 78.5% at 623 K. Pd/Norit exhibited less activ-
ity: total conversion was achieved only at 523 K. The selectivity
for H2, 95.8% at 423 K, decreased by only ꢁ5% up to 623 K. The
decomposition occurred at the lowest rate on Ru/Norit with a H2
selectivity of 97.2% at 423 K at a conversion of ꢁ22%. The selectiv-
ity decreased to 82% at 623 K. After determination of the tempera-
ture dependence of the decomposition of formic acid, we followed
the reaction in time on stream at 573 and 623 K for 10–12 h. No or
only very slight changes were experienced in the conversion and in
the product distribution on any catalyst. Some important data for
the decomposition of formic acid on various metals are presented
in Table 2.
Fig. 2. FTIR spectra of Al2O3 (a, b) and Ir/Al2O3 (c–e) following the adsorption of
HCOOH at 300 K and after subsequent degassing at different temperatures.
pure Al2O3 produced exactly the same spectral features, with the
difference that the 1597 cmꢂ1 band was more stable. Note that in
addition to the above absorption bands weak signals were also
identified at 2040–2050 cmꢂ1 for Ir samples and at 2060 cmꢂ1
for Rh/SiO2.
3.2. Thermal desorption measurements
TPD spectra for various products after the adsorption of formic
acid on different samples at ꢁ300 K are presented in Fig. 3. The
weakly adsorbed formic acid was released from every sample
immediately above 300 K. Desorption of other compounds from
In order to establish what adsorbed species exist on the surface
under dynamic conditions, we performed FTIR spectroscopic
studies in situ during the catalytic reaction in a flow of a gaseous
Fig. 3. TPD spectra following the adsorption of HCOOH on Ir/Norit (A), Ir/SiO2 (B) and Ir/Al2O3 (C) at 300 K.

216
F. Solymosi et al. / Journal of Catalysis 279 (2011) 213–219
Fig. 4. Conversion of formic acid (A) and the selectivity of H2 formation (B) on Pt metals supported by carbon Norit as a function of temperatures.
Table 2
Some characteristic data for the decomposition of formic acid on Norit supported Pt metals.
Catalyst
Dispersion
(%)
Work function of
the metals (eV) [46]
NH (ꢄ102)
NH (ꢄ102)
423 K
473 K
2
2
at 373 K
at 423 K
H2 selectivity (%)
H2 yield (%)
H2 selectivity (%)
H2 yield (%)
Ir
Pt
Pd
Ru
Rh
26.1
23.5
8.6
5.6
24.3
5.76
5.70
5.12
4.71
4.98
9.6
6.4
4.4
1.3
0.9
41.2
37.8
36.3
25.3
12.8
99.0
98.0
95.1
97.3
91.8
79.2
55.6
27.7
15.2
37.2
98.3
99.1
95.5
94.9
84.8
98.3
89.8
60.8
52.5
75.7
ꢀ
ꢁ
Molecules formed
Metal site ðsÞ
NH ¼ turnover numbers given by :
.
2
Fig. 5. In situ IR spectra registered on silica supported Pt metals during the decomposition of formic acid at different temperatures.
mixture of HCOOH + Ar. This study was carried out in the DRIFT
cell. Beside the intense absorption band due to formic acid at
reaction temperature their intensities remained practically unal-
tered. In the CO stretching region, weak absorption bands were
also detected at 2030–2050 cmꢂ1. When the stream of HCOOH + Ar
was changed to pure Ar, all the absorption bands disappeared at
1720–1730 cmꢂ1
, weak spectral features appeared at 1595,
1395–1400 and ꢁ1200 cmꢂ1 for all samples. With the rise of the

F. Solymosi et al. / Journal of Catalysis 279 (2011) 213–219
217
Table 3
once, with the exception of that due to CO. Selected regions of IR
spectra are displayed in Fig. 5.
Some characteristic data for production of H2 in steam reforming of HCOOH over
supported Pt metals.
On the 2% Pt/Norit catalyst, we performed kinetic measurements
in the temperature range 380–425 K. The partial pressure of formic
acid was varied, with the total flow rate kept at 40 ml minꢂ1 by the
addition of Ar ballast to the system. The decomposition of formic
acid under these conditions followed zero-order kinetics. The
Arrhenius plots yielded 70.7 kJ/mol 3 kJ/mol for the activation
energy of the formation of H2. This value is consistent with previous
data, 75 kJ/mol on Rh/Al2O3 [25], 72 kJ/mol on Pt/Al2O3 [10] and
67 kJ/mol on Pd/C [11].
Catalyst
Temp. (K)
Conv. (%)
H2 sel. (%)
H2 yield
2.0% Ir/Norit
383
423
473
50
88
100
100
99.5
99.5
50.0
87.5
99.5
5% Ir/Norit
2% Ir/SiO2
383
423
473
87
100
ꢁ100
38
48
95
100
100
100
87.0
100
100
383
423
473
100
99.0
98.0
38.0
47.5
93.1
2% Pt/Norit
2% Rh/Norit
2% Pd/Norit
2% Ru/Norit
383
423
473
25
52
100
99.2
99.0
98.6
24.8
51.4
98.6
3.3.2. Reforming of formic acid
The addition of water to the formic acid (HCOOH:H2O = 1:1)
influenced the conversion of formic acid at different temperatures
only slightly, but affected the product distribution and enhanced
the selectivity of H2 formation to various extents. As dehydrogena-
tion predominated at lower temperatures, the reforming of formic
acid was followed in time on stream at 383, 423 and 473 K. The
selectivity data determined for various catalysts demonstrated
that, apart from minor fluctuations, little change occurred in time
on stream (Fig. 6). CO-free H2 was obtained on 2% Ir/Norit at
383
423
473
15
26
100
ꢁ86.0
97.0
97.0
12.9
25.2
97.0
383
423
473
15
31
95
99.5
98.7
96.6
14.9
30.5
91.7
383
423
473
18
25
66
98.0
97.5
95.0
17.6
24.3
62.7
383 K. SH decreased to ꢁ99.5% at 423 and 473 K. On 5% Ir/Norit,
2
SH remained 100% in the temperature range 383–473 K. Very high
2
values of 99.5–99.2% were measured for Pd/Norit and Pt/Norit at
383 K, when that for Ru/Norit was ꢁ98%, which diminished to
97.5% and 95.0% at 423 and 473 K, respectively. Rh/Norit exhibited
different behaviors as the highest selectivity, 96.0–97.0%, was
manifested at 423–473 K and the lowest at 383 K. This feature
was reproducible. In order to obtain more information on the effi-
ciency of the catalyst in the generation of H2 we calculated the
yield of H2 formation. The results on various samples are to be seen
in Table 3.
4. Discussion
4.1. Interaction of HCOOH with Pt metals
At 200–250 K, on Ir/SiO2, all the absorption bands characteristic
of molecularly adsorbed formic acid appeared in the IR spectrum
(Fig. 1). The assignments of the bands are shown in Table 1. When
the sample was heated, the intensities of all the bands gradually
Fig. 6. Selectivity of hydrogen formation in the reforming of HCOOH on Pt metals supported by carbon Norit at different temperatures (HCOOH:H2O = 1:1).

218
F. Solymosi et al. / Journal of Catalysis 279 (2011) 213–219
HCOOH = H2 + CO2:
and
D
G = ꢂ48.4 kJ/mol
ð4Þ
ð5Þ
decreaseꢂd1. Besides the most intense band of formic acid at
1369 cmꢂ1 at and above 225 K. The appearance of these vibrations
strongly suggests the dissociation of formic acid and the formation
of formate species, very likely bonded to Ir:
1722 cm
new spectral features were detected at 1586 and
HCOOH = H2O + CO:
D
G = ꢂ28.5 kJ/mol
If the space velocity is high, and the temperature is low enough, the
occurrence of secondary reactions can be excluded, such as the
hydrogenation of CO2 to CH4, which is well catalyzed by supported
Pt metals [31–35]. All the Pt metals are active in the decomposition
of formic acid, but with appreciable differences in catalytic effi-
ciency (Fig. 4A). The differences in catalytic behavior show up in
the selectivity of H2 formation, particularly at higher temperatures
(Fig. 4B). The conversion data in Fig. 4A indicate the activity se-
quence Ir > Pt > Rh > Pd > Ru.
A huge number of data initially suggested the importance of the
electronic factor in the decomposition of formic acid on different
metals and oxides [16–21]. Those catalysts were found to be active
which were able to accept electrons from formic acid or its dissoci-
ation products. Later, as the role of the electron theory of catalysis
declined, however, more emphasis was given to the formation and
stability of formate on catalyst surfaces [21–27,47,48]. Those metals
were considered to be the best catalysts on which this surface com-
plex was formed easily, but was not very stable. Plots of the rate of
decomposition at the same temperature as a function of the heat of
formation of the corresponding formate yielded ‘‘volcano-shaped’’
curves [47]. The Pt group proved to be the best of 13 metals, but
the activities of the metals were not related to the number of surface
metals atoms, so this activity sequence cannot be considered as a
reliable one [47]. A more reasonable relationship was presented
by Barteau [40,51], who found a linear correlation between the
decomposition temperature of formates on metal single crystals
and the heat of formation of the corresponding metal oxides. It is
to be pointed out that no data for the reactions of formic acid or
formate on Ir catalyst were given in either works [21,22,40,47,48].
Taking into account the dispersity of Norit-supported metals, the
specific activities of the metals in terms of turnover frequencies
HCOOHðaÞ = HCOOðaÞ + HðaÞ
ð1Þ
This conclusion is supported by the results on pure SiO2, when this
band was missing at both low and high temperatures. This is in har-
mony with the experience that formic acid does not dissociate to
formate on SiO2 [21,22,25,47,48]. The absorption bands due to for-
mate, however, were detected only up to 290–300 K, indicating the
decomposition of formate species on Ir:
HCOOðaÞ = CO2ðgÞ + HðaÞ
ð2Þ
In addition to the formate bands, weak absorption peak appeared at
2040 cmꢂ1, which can be attributed to CO linearly bonded to an Irx
cluster [49]. This suggests another route of decomposition of sur-
face formate:
HCOOðaÞ = COðaÞ + OHðaÞ
ð3Þ
Interestingly, there was no indication of absorption bands at ꢁ2107
and ꢁ2037 cmꢂ1 due to the surface dicarbonyl complex Ir+(CO)2.
Similar measurements were carried out on other samples. As shown
in Fig. 1B, the adsorption of formic acid on Rh/SiO2 at ꢁ200 K pro-
duced similar spectral features, but the formate bands were some-
what more stable. In this case, the CO bands developed at
2060 cmꢂ1. Vibrations due to Rh+(CO)2 at 2030 and 2100 cmꢂ1 were
again missing. Previous FTIR and STM studies revealed that dicar-
bonyl complex is formed in the CO-induced oxidative disruption
of Rh and Ir nanoparticles [49,50]. The absence of this species is
probably due to the presence of H2, which hinders the disruption
processes [50]. We obtained very similar results on other SiO2-
supported metals. The stability of the formate bands depended only
slightly on the nature of the metals, and the bands were eliminated
at 290–373 K. The low stability of formate on Pt metals is in har-
mony with the results obtained on single crystals of these metals
by various electron spectroscopic methods [38–40].
A completely different picture was observed for Ir/Al2O3 (Fig. 2).
In this case, strong absorption bands due to formate species
appeared even following the adsorption of formic acid at 300 K,
and they were detectable up to 573 K. As the adsorption of formic
acid on Al2O3 produced the same IR spectra, it can be concluded that
the formate is located not on the Ir, but rather on the Al2O3. A com-
parison of the thermal behavior of the formate band at 1597 cmꢂ1 on
Ir/Al2O3 and on pure Al2O3 revealed that this species was apprecia-
bly more unstable on Ir/Al2O3. A possible reason is that Ir promotes
the decomposition of formate that diffuses to the metal/oxide
interface.
(NH , rates per surface metal atom) were calculated at 473 K (Table
2
2). This gave the following sequence: Ir, Pt, Pd, Ru, Rh. This order is
nearly the same as obtained by Barteau [40] on metal single crystal
surfaces. As concerns a possible relationship of these data with the
electric properties of the metals, their work functions are also given
in Table 2. This shows that the specific activity increases with in-
crease in the work function of the metals, i.e. their ability to accept
electrons. This may support the early assumption concerning the
role of the electric structure of the metals in the decomposition of
formic acid.
The situation is more complex as regards the selectivity of H2
formation, for its temperature dependence varies with the metal.
On most of the catalysts, this parameter decreased significantly
as the temperature was raised. It is important to note that pure
H2 was not generated by the decomposition of formic acid on
any of the catalysts. The selectivity for H2 at 423 K decreased in
the sequence Ir, Pt, Ru, Pd, Rh.
The results of the TPD are in harmony with the IR studies. The
desorption of formic acid observed at ꢁ300 K very likely originates
from the formic acid molecularly adsorbed on SiO2 or carbon Norit
supports. The H2 and a small amount of CO released at higher tem-
peratures are the products of the decomposition of formate at
ꢁ300 K, which remained adsorbed on the metals. As CO2 is not
chemisorbed on Pt metals at 300 K, no desorption of CO2 was
established by TPD. In harmony with the IR spectroscopic studies,
the situation was different for Ir/Al2O3. Besides H2 and CO, a large
amount of CO2 also evolved as a result of the decomposition of sta-
ble formate on the Al2O3 at higher temperatures.
Higher selectivities for H2 measured in the reforming reaction
are probably a result of the occurrence of the water–gas shift
reaction:
H2O + CO = H2 + CO2
ð6Þ
As the primary aim of this study was to generate CO-free H2, the
formation of even a few tenths of one percent of CO makes the cat-
alyst less attractive and should be avoided. H2 was produced with
98.0–99.5% selectivity on Norit-supported Pt, Pd and Ru at 383–
423 K. H2 completely free of CO was obtained only on Ir catalyst.
Moreover, a H2 yield of 100% was attained only on the Ir/Norit. Irid-
ium remained a highly selective catalyst at 383 K, even when it was
deposited on a SiO2 support, though the H2 yield was much less
4.2. Decomposition and reforming of formic acid
Formic acid is a convenient source of H2, and its decomposition
is the simplest reaction via which to produce H2. Only two reaction
pathways need to be considered:

F. Solymosi et al. / Journal of Catalysis 279 (2011) 213–219
219
}
[7] F. Solymosi, Gy. Kutsán, A. Erdohelyi, Catal. Lett. 11 (1991) 149.
than that on Ir/Norit (Table 3). Ojeda and Iglesia [10] have recently
shown that the selectivity for hydrogen is close to unity on gold
nanoparticles supported on Al2O3. Ross et al. [11] reported a selec-
tivity of 95–99% on Pd/C catalyst.
[8] M. Belgued, H. Amariglio, P. Pareja, A. Amariglio, J. Sain-Just, Catal. Today 13
(1992) 437.
[9] T. Koerts, M.J.A.G. Deelen, R.A. van Santen, J. Catal. 138 (1992) 101.
[10] M. Ojeda, E. Iglesia, Angew. Chem. Int. Ed. 48 (2009) 4800.
[11] D.A. Bulushev, S. Beloshapkin, J.R.H. Ross, Catal. Today 154 (2010) 7.
[12] Á. Koós, F. Solymosi, Catal. Lett. 138 (2010) 23.
[13] X. Zhou, Y. Huang, W. Xing, C. Liu, J. Liao, T. Lu, Chem. Commun. (2008) 3540.
[14] S. Ma, R. Larsen, R.I. Masel, J. Power Sources 144 (2005) 28, and references
therein.
[15] J.H. Choi, K.J. Jeong, Y. Dong, J. Han, T.H. Lim, J.S. Lee, Y.E. Sung, J. Power Sources
163 (2006) 71, and references therein.
[16] G.-M. Schwab, Discuss. Faraday Soc. 8 (1950) 166;
G. Reinäcker, G. Techel, Z. Anorg. Chem. 304 (1960) 58.
[17] D.D. Eley, P. Luetic, Trans. Faraday Soc. 53 (1957) 1476.
[18] Z.G. Szabó, F. Solymosi, Acta Chim. Hung. 25 (1960) 145. 161.
[19] Z.G. Szabó, F. Solymosi, in: Actes Congr. Intern. Catalyse 2e Paris, 1961, p. 1627.
[20] F. Solymosi, Catal. Rev. 1 (1968) 233.
[21] G.C. Bond, Catalysis by Metals, Academic, London, 1962.
[22] P. Mars, J.J.F. Scholten, P. Zweitering, Adv. Catal. 14 (1963) 35.
[23] J.M. Trillo, G. Munuera, J.M. Criado, Catal. Rev. 7 (1972) 51, and references
therein.
As regards the occurrence of the decomposition of formic acid
on supported Pt metals, we accept the early proposal that it pro-
ceeds through the formation and decomposition of formate spe-
cies. In situ IR spectroscopic studies revealed that formate species
exists on the catalyst surfaces during the reaction at 383–473 K
(Fig. 5). We assume that the slow step in the decomposition is very
likely to be the cleavage of the CAH bond in the adsorbed formate.
This step may occur more easily on the metals as compared with
rupture of the CAO bond in formate, which may be one of the rea-
sons for the preference of the dehydrogenation of formic acid. The
large work function of Ir, i.e. its ability to accept electrons from the
substrate, may also contribute to its unequaled activity and selec-
tivity. It may be noted that supported Ir is likewise one of the best
catalysts in the decomposition of hydrazine [52].
[24] E. Iglesia, M. Boudart, J. Catal. 81 (1983) 214.
[25] F. Solymosi, A. Erdohelyi, J. Catal. 91 (1985) 327.
}
[26] D.E. Fein, I.E. Wachs, J. Catal. 210 (2002) 241.
[27] W.-J. Chun, K. Tomishige, M. Hamakado, Y. Iwasawa, J. Chem. Soc. Faraday
Trans. 91 (1995) 4161, and references therein.
5. Conclusions
[28] A. Deluzarche, R. Kieffer, A. Muth, Tetrahedron Lett. 18 (1977) 3357.
[29] G.A. Olah, Á. Molnár, Hydrocarbon Chemistry, Wiley-Interscience, New York,
2003.
(i) FTIR spectroscopic measurements on SiO2-supported Pt
metals demonstrated the dissociative adsorption of formic
acid at 225–230 K. The formate species formed, however,
was detectable only up to 300–373 K. In situ IR studies
revealed that formate exists on the catalyst surface even
during the catalytic reaction at 383–473 K.
[30] Y. Amenomiya, J. Catal. 57 (1979) 64.
}
[31] F. Solymosi, A. Erdohelyi, T. Bánsági, J. Catal. 68 (1981) 371.
}
[32] F. Solymosi, A. Erdohelyi, in: Proceedings of the 7th International Congress on
Catalysis, Tokyo, 1980, Kodansha/Elsevier, Tokyo/Amsterdam, 1981, p. 1448.
}
[33] F. Solymosi, A. Erdohelyi, M. Kocsis, J. Chem. Soc. Faraday Trans. I 77 (1981)
1003.
(ii) Pt metals deposited on carbon Norit were found to be effec-
tive catalysts in the vapor-phase decomposition of formic
acid to generate H2, though CO-free H2 was not obtained
by the decomposition on any of the metals. Ir/Norit proved
to be the best and the most selective catalyst.
[34] E. Ramarson, R. Kieffer, A. Kiennemann, J. Chem. Soc. Chem. Commun. (1982)
645.
[35] A. Deluzarche, J. Cressely, R. Kieffer, J. Chem. Res.-S (1979) 136;
A. Deluzarche, J. Cressely, R. Kieffer, J. Chem. Res.-M (1979) 1656.
[36] R.A. Dalla Betta, M. Shelef, J. Catal. 49 (1977) 383.
[37] F. Solymosi, I. Tombácz, M. Kocsis, J. Catal. 75 (1982) 78.
[38] R.J. Madix, Adv. Catal. 29 (1980) 1.
(iii) The addition of water to the formic acid allowed the produc-
tion of H2 completely free of CO on Ir catalysts at 383–423 K.
[39] F. Solymosi, J. Kiss, I. Kovács, Surf. Sci. 192 (1987) 47.
[40] M.A. Barteau, Catal. Lett. 8 (1991) 175.
[41] A. Boddien, F. Gärtner, R. Jackstell, H. Junge, A. Spannenberg, W. Baumann, R.
Ludwig, M. Beller, Angew. Chem. Int. Ed. 49 (2010) 8993.
[42] C. Fellay, P.J. Dyson, G. Laurenczy, Angew. Chem. Int. Ed. 47 (2008) 3966.
[43] S. Enthaler, ChemSusChem 1 (2008) 801, and references therein.
[44] S. Enthaler, J. von Langermann, T. Schmidt, Energy Environ. Sci. 3 (2010) 1207,
and references therein.
Acknowledgment
This work was supported by the grant OTKA under Contact
Number K 81517.
[45] F. Solymosi, R. Barthos, A. Kecskeméti, Appl. Catal. A Gen. 350 (2008) 30, and
references therein.
[46] R.C. Weast, M.J. Astle (Eds.), CRC Handbook of Chemistry and Physics, 64th ed.,
CRC Press, 1983.
[47] R.E. Eischens, W.A. Pliskin, in: Actes Congr. Intern. Catalyse 2e Paris, 1961, p.
789.
[48] W.M.H. Sachtler, J. Fahrenfort, Actes du Deuxieme Congres Internationale de
Catalyse (1961) 831.
[49] F. Solymosi, É. Novák, A. Molnár, J. Phys. Chem. 94 (1990) 7250.
[50] A. Berkó, G. Ménesi, F. Solymosi, J. Phys. Chem. 100 (1996) 17732, and
references therein.
References
[1] G. Sandstede, T.N. Veziroglu, C. Derive, J. Pottier (Eds.), Proceedings of the
Ninth World Hydrogen Energy Conference, Paris, France, 1972, p. 1745.
[2] A. Haryanto, S. Fernando, N. Murali, S. Adhikari, Energy Fuels 19 (2005) 2098.
[3] F. Marino, M. Boveri, G. Baronetti, M. Laborde, Int. J. Hydrogen Energy 26
(2001) 665.
[4] C. Díagne, H. Idriss, A. Kiennemann, Catal. Commun. 3 (2002) 565.
[5] D.K. Liguras, D.I. Kondarides, X.E. Verykios, Appl. Catal. B Environ. 43 (2003)
345.
[51] M. Mavrikakis, M.A. Barteau, J. Mol. Catal. A Chem. 131 (1998) 135.
[52] Y.B. Jang, T.H. Kim, M.H. Sun, J. Lee, S.J. Cho, Catal. Today 146 (2009) 196.
}
[6] A. Erdohelyi, J. Raskó, T. Kecskés, M. Tóth, M. Dömök, K. Baán, Catal. Today 116
(2006) 367.