Những tiến bộ gần đây trong sản xuất hiđro từ sự phân hủy axit amin dạng formic

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 8 ) 1 e1 7

Available online at www.sciencedirect.com

ScienceDirect
journal homepage: www.elsevier.com/locate/he

Review Article

Recent progress in hydrogen production from
formic acid decomposition
Xian Wang a,b, Qinglei Meng a,d, Liqin Gao a,b,c, Zhao Jin a, Junjie Ge a,*,
Changpeng Liu a, Wei Xing a,c,**
a

Laboratory of Advanced Power Sources, Jilin Province Key Laboratory of Low Carbon Chemical Power Sources,
Changchun Institute of Applied Chemistry, Chinese Academy of Sciences, Changchun, Jilin, 130022, PR China
b
University of Chinese Academy of Sciences, Beijing, 100039, PR China
c
State Key Laboratory of Electroanalytical Chemistry, Changchun Institute of Applied Chemistry, Chinese Academy
of Sciences, Changchun, Jilin, 130022, PR China
d
University of Science and Technology of China, Hefei, Anhui, 230026, PR China

article info

abstract

Article history:

Formic acid, as the simplest carboxylic acid which can be obtained as an industrial by-

Received 13 December 2017

product, is colorless, low toxicity, and easy to transport and storage at room tempera-

Received in revised form

ture. Recently, Formic acid has aroused wide-spread interest as a promising material for

20 February 2018

hydrogen storage. Compared to other organic small molecules, the temperature for formic

Accepted 22 February 2018

acid decomposition to produce hydrogen is lower, resulting in less CO toxicant species.

Available online xxx

Lots of catalysts on both homogeneous catalysts and heterogeneous were reported for the
decomposition of formic acid to yield hydrogen and carbon dioxide at mild condition. In

Keywords:

this paper, the recent development of mechanism and the material study for both ho-

Formic acid decomposition

mogeneous catalysts and heterogeneous catalysts are reviewed in detail.

Hydrogen production

© 2018 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.

Heterogeneous catalysis
Homogeneous catalysis
Catalysis selectivity

Contents
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Homogeneous catalysts for formic acid decomposition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Ruthenium-based catalysts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Iridium-based catalysts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Iron-based catalysts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Copper-based catalysts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

00
00
00
00
00
00

* Corresponding author.
** Corresponding author. Laboratory of Advanced Power Sources, Jilin Province Key Laboratory of Low Carbon Chemical Power Sources,
Changchun Institute of Applied Chemistry, Chinese Academy of Sciences, Changchun, Jilin, 130022, PR China.
E-mail addresses: (J. Ge), (W. Xing).
/>0360-3199/© 2018 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.
Please cite this article in press as: Wang X, et al., Recent progress in hydrogen production from formic acid decomposition, International

Journal of Hydrogen Energy (2018), />

2

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 8 ) 1 e1 7

Heterogeneous catalysts for formic acid decomposition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
The mechanism of formic acid decomposition on heterogeneous catalysts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Palladium-based catalysts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Palladium-based bimetallic catalysts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Palladium-based core-shell catalyst . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Palladium-based trimetallic catalyst . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Gold-based catalysts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Platinum-based catalysts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Other catalysts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Summary and outlook . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Introduction
Traditional fossil fuels are creating serious climate and environment issues globally [1e3]. Meanwhile, due to the increase in
energy demand, the global fossil fuel consumptions rate is expected to double in the next thirty years, which makes their
doomed depletion end come earlier. Therefore, taking advantage of sustainable energy resources, such as wind and solar
energies, is imperative, and has received huge amount of
attention [4,5]. The intermittent nature of solar and wind energies necessitates for energy storage media and technique for
its efficient on demand release. Hydrogen is an ideal energy
carrier with high energy density, cleanness, and earth abundance. The energy stored in the hydrogen molecule can be
efficiently utilized through a variety of ways, among which
proton exchange membrane fuel cell (PEMFC) is highly attractive due to its high energy efficiency, environmental benign and
high energy density. There are many viable ways to product

hydrogen, such as water electrolysis [6,7] (Equation (1)), hydrogenase route [8] (Equation (2)), and extraction from biomass
such as methanol [9] and formic acid [10e12] (Equation (3)).
2H2 OðlÞ/2H2 ðgÞ þ O2 ðgÞðelectrolysisÞ

(1)

2Hþ þ 2Xreduced /H2 þ 2Xoxidised ðhydrogenaseÞ

(2)

HCOOH4H2 þ CO2 ðhomo=heterogeneous catalysisÞ

(3)

Among these solutions, hydrogen production from formic
acid (FA) is a promising route to store and release at room
temperature, with the advantages of high gravimetric (4.4 wt
%) and volumetric (53.4 g/L) H2 capacity [12]. As the simplest
carboxylic acid, FA is a colorless and low toxicity liquid at
ambient condition (density ¼ 1.22 g/mL, m.p. ¼ 281.5 K,
b.p. ¼ 373.9 K), which can be obtained as an industrial byproduct, through photoelectric catalytic CO2 reduction, and
by decomposition of biomass [13]. The liquid phase FA
decomposition (FAD) to yield hydrogen has been realized,
making the H2 production at mild condition promising for the
on demand release and utilization in hydrogen fuel cell vehicles. Selectivity is an important issue as it determines the
quality of the final H2 gas generated. Depending on the type of
catalysts used and the working condition, such as reactant
concentration and the reaction temperature, formic acid

00

00
00
00
00
00
00
00
00
00
00
00

decomposition (FAD) may happen via the following two
possible ways [14] (Scheme 1).
In reaction pathway 1, FA decomposes through dehydrogenation pathway and produces hydrogen and carbon dioxide, which is the reverse reaction process of carbon dioxide
hydrogenation. Thus hydrogen can be effectively stored in
formic acid through this cycle. At present, major efforts are
concentrating on carbon dioxide hydrogenation [15], where
several effective techniques have been developed. However,
much fewer efforts have been paid on the FAD to produce
hydrogen, which deserves more attention.
In this review, we will focus on the recent development of
FAD catalysts on both homogeneous catalysts and heterogeneous catalysts. We will also give a summary on proposed
future research direction for FAD along with possible obstacles
on the formic acid hydrogen storage that may be encountered.

Homogeneous catalysts for formic acid
decomposition
Over the past few decades, massive efforts were paid to search
for high performance homogeneous catalysts towards FAD. In

1967, Coffey reported that soluble platinum, ruthenium and
iridium phosphine complexes were efficient in selectively
decomposing formic acid into H2 and CO2 [16]. Since then,
massive research endeavors have been concentrated on the
development of highly efficient noble-metal ruthenium and
iridium complex. Meanwhile, catalysts based on non-noble
metals complex such as iron and copper were occasionally
reported [17e22].

Ruthenium-based catalysts
In 2000, Puddephatt and co-workers investigated the binuclear Ru complex for the dehydrogenation of FA [23]. The
dissolved [Ru2(m-CO)(CO)4(m-dppm)2] catalyst in acetone

Scheme 1 e Possible ways for the formic acid
decomposition.

Please cite this article in press as: Wang X, et al., Recent progress in hydrogen production from formic acid decomposition, International
Journal of Hydrogen Energy (2018), />

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 8 ) 1 e1 7

solution, was found efficient for the reversible reaction between HCOOH and CO2/H2. For the first time, the binuclear
homogeneous catalyst was found not only effective in catalyzing FAD but also the hydrogenation of CO2 to form FA.
Beller and co-workers studied the efficient generation of
hydrogen from FA by using the [RuCl2(benzene)]2 [24],
RuBr3$xH2O [25] and [RuCl2(PPh3)3] [26] as the catalyst precursor. With the in situ generated [RuCl2(benzene)]2/6 equiv. dppe,
the catalysts were shown stable and continuously working for
the FAD with the turnover frequencies (TOF) and turnover
number (TON) at 900 hÀ1 and 260000 at mild conditions [24].
High catalytic activity was originated from the properly tuned

adducts and their concentrations. With RuBr3$xH2O, 3.4 equiv.
PPh3 catalyst system, the best activity (TOF up to 3630 hÀ1 after
20 min) was observed for hydrogen generation by using
5HCOOH/2NEt3 adduct at room temperature [25]. By using the
[RuCl2(PPh3)3], they reported that the production of hydrogen
from FA amine adducts exhibited the initial TOF of 2688 hÀ1 at
room temperature [26]. All the catalytic systems exhibit high
selectivity over the H2/CO2 path and no CO is detected in the
final mixture gas at mild conditions, demonstrating that high
quality H2 was generated and can be directly served as fuel in
H2/O2 fuel cell after removal of CO2. Later, the same group reported that light could significantly accelerate the production
of hydrogen from FA by using the ruthenium-catalysts [27]. The
catalytic performance strongly depends on the catalyst precursors and ligands used, as shown in Fig. 1.
Laurenczy et al. reported a novel hydrophilic rutheniumbased catalysts which was produced from the water-soluble
ligand
meta-trisulfonated
triphenylphosphine
with
[Ru(H2O)6]2þ and RuCl3 [28]. Owing to the addition of formate
salt, the conversion rate of the catalytic systems at all temperature was 90e95% (Fig. 2). Almost the same time, the same
group studied the water-soluble sulfonato aryl- and alkyl-/

3

arylphosphine ligands in ruthenium(II) aqueous for FAD and
found the monosulfonato triphenylphosphine and di(m-sulfonato)triphenyl phosphine with good activity [29]. They had
confirmed that the ligand basicity and steric effect were the
main parameters that determined the catalytic activity.
In 2009, Wills and co-workers reported several Ru(II) and
Ru(III) catalyst precursors for FAD in triethylamine at 393 K,

with no adding of phosphine ligands [30]. As expected, the
high FAD activities were achieved at such high temperatures
(TOF up to 1.8 Â 104 hÀ1). Regrettably, the concentrations of CO
surpassed 200 ppm for all the catalysts. They suggested that
all the precursors formed the [Ru2(HCO2)2(CO)4] as the active
species under these reaction conditions. Interestingly, all the
catalysts showed slight increase activity during each reuse,
indicating the continuous formation of active catalyst species.
Later, the same group used [Ru2Cl2(DMSO)4]/triethylamine
system to decompose FA without acid accumulation at a rate
approaching the catalyst's maximum activity in this system
[31].
In 2016, Huang et al. studied a rationally designed ruthenium catalyst for FAD with high activity and selectivity under
mild condition (Fig. 3) [32]. Recently, they investigated a
ruthenium complex containing an N,Nʹ-diimine ligand for
formic acid decomposition without formation of CO [33]. The
TOF and TON were 12000 hÀ1 and 3500000 at 90  C, respectively. They suggested that Ru complex [Ru(p-Cymene)(2,20 biimidazoline)Cl]Cl showed a good activity towards FAD and
realized the high-pressure hydrogen production from formic
acid.
The FAD processes in these multiple catalysis systems all
take place in a mixture solution. Meanwhile, the produced
hydrogen is always accompanied with production of carbon
dioxide and traces of vaporized solvent and the complex
separation process hinders their commercial applications.

Fig. 1 e Different catalyst precursors and ligands showed the different catalytic performance [27].
Please cite this article in press as: Wang X, et al., Recent progress in hydrogen production from formic acid decomposition, International
Journal of Hydrogen Energy (2018), />

4

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 8 ) 1 e1 7

Fig. 2 e The conversion of formic acid at different temperature by using the novel hydrophilic ruthenium-based catalysts
[28].

Moreover, the organic solvents which were used in the catalysts are mainly subjected to emission regulations and require
the extra exhaust gas cleaning steps. In order to avoid solvent
evaporation, ionic liquids (ILs) can be used as the reaction
medium. For the first time, Deng et al. tested the effect of the
IL on the decomposition of FA [34]. They tested their catalytic
performances by using ruthenium-based catalyst and a series
of amine-functionalized ILs. With the mixture of iPr2NEMimCl
and HCOONa, The TOF was up to 627 hÀ1 at 313 K. Dupont and
co-workers used the same ruthenium complex, [{RuCl2(pcymene)}2], for the dehydrogenation of formic acid [35]. The Ru
complex was dissolved in the ionic liquid (IL) [Et2NEMim]Cl at
353 K, and the TOF reaches 1540 hÀ1. In 2011, Wasserscheid
and co-workers investigated a novel and efficient IL-based
FAD system, which was formed by the RuCl3 and nonfunctionalized ionic liquids as catalyst precursors [36]. They
have confirmed that the most efficient system was RuCl3
dissolved in [EMMIM][OAc], while the release H2 and CO2 were
obtained as the products with no CO formation, with TOF
recorded as 850 hÀ1 at 120  C.

Iridium-based catalysts
In 2009, Himeda reported an efficient iridium catalyst for the
decomposition of FA [37]. The TOF reached up to 14000 hÀ1 at

363 K. They had demonstrated that the pH and the electron
effect of the substituents in the bipyridine ligand could tune

the catalytic activity. Fukuzumi et al. investigated a heterodinuclear iridium-ruthenium complex catalyst, which was
highly efficient for FAD in aqueous solution with the TOF up to
423 hÀ1 at pH ¼ 3.8 [38].
A novel iridiumebisMETAMORPhos complex for FAD was
reported by Reek and co-workers in 2013 [39]. The catalysts
were active (TOF up to 3092 hÀ1, in toluene) in FAD without
external base. They utilized the ligand to form anion as an
internal base to develop the “base-free” catalytic system, and
the reaction is free from CO formation.
Xiao et al. investigated a well-defined N^C cyclometallated
iridium(III) complexes catalyst for FAD to produce H2 and CO2
with the TOF up to 147000 hÀ1 at 313 K [40]. Interestingly, this
catalytic system involved the metal center and the NH functionality to explain the possible way for dehydrogenation of FA.
They suggested that the formation of H2 was facilitated by
HCOOH-mediated proton hopping (Fig. 4). The remote NH
functionality was vital to this catalytic system, without which
there was no decomposition. They suggested that FA played a
double role, showing both as the proton source and as the
proton shuttle. Ikariya and co-workers studied a Ir complexes
catalyst
which
was
produced
from
N-triflyl-1,2diphenylethylenediamine for the decomposition of FA

Fig. 3 e A new class of PN3eRu complexes [32].
Please cite this article in press as: Wang X, et al., Recent progress in hydrogen production from formic acid decomposition, International
Journal of Hydrogen Energy (2018), />

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 8 ) 1 e1 7

Fig. 4 e Proposed catalytic cycle for the dehydrogenation of
HCOOH [40].

showing high catalytic performance with the TOF up to
6000 hÀ1 without base additives at ambient temperature [41].
They had confirmed that the hydrido-Ir complex could be
determined and isolated as a crucial catalytic intermediate. In
addition, they suggested that the proton-relay processes
mediated by the NH proton and water had great potential for
efficient H2 production from FA.

Iron-based catalysts
The first light-driven iron based catalyst which was in situ
formed from Fe3(CO)12, 2,20 :60 200 -terpyridine or 1,10phenanthroline, and triphenylphosphine for FAD under the
irradiation of visible light at ambient condition, was studied
by Ludwig, Beller and co-workers [19]. The TOF was up to
200 hÀ1 at 60  C, while the Fe3(CO)12 as the precursor and 6,600 (phenyl)-2,20 :6,200 -terpyridine and PPh3 as the ligands.
Depending on the experimental and theoretical (density
functional theory, DFT) studies, the author confirmed that
triphenylphosphine played an active role in the catalytic cycle
and N-ligands enhanced the stability for this catalytic system.
Almost the same time, the same group investigated a new iron
phosphine catalyst which presented a higher catalyst activity
than the iron/triphenylphosphine system [18]. With the tribenzylphosphine and benzyldiphenylphosphine as the ligands, the catalyst exhibited significant increase in both
catalyst activity and stability (TON up to 1266). The author
attributed the improved catalyst activity and stability to the
ortho-metalated iron species from Fe(PBn3). Beller and Laurenczy and co-workers [20] later used [Fe(BF)4)2]∙6H2O,
[FeH(PP3)]BF4, [FeH(H2)(PP3)]BF4, [FeH(H2)(PP3)]BPh4 and

[FeCl(PP3)]BF4 as the precursors, tris[(2-diphenylphosphino)
ethyl]phosphine as the ligand in propylene carbonate to
fabricate highly active FAD catalysts. All the iron precursors
showed great activity for the decomposition of FA except for
[FeCl(PP3)]BF4. While using 0.005 mol percent of [Fe(BF)4)2]
∙6H2O and PP3 in propylene carbonate at 80  C without further

5

additives or base, the TOF was up to 9425 hÀ1 and the TON was
more than 92000. Based on the experimental (in situ 13C and 31P
NMR) and theoretical (DFT) studies, they suggested that
[FeH(PP3)]þ was the common complex for the two competing
catalyst cycles.
An interesting iron catalyst system which used the Lewis
acid (LA) as co-catalyst showing high activity for FAD (TON
above to 1000000) was observed by Hazari, Schneider and coworkers [17]. According to their studies, the LA is suggested
to assisting the decarboxylation of a key iron formate intermediate. Different LAs were used to promote the catalytic
activity and they suggested that the highest TOF and TON
were obtained with alkali or alkaline earth metal salt cocatalysts (especially LiBF4). Importantly, the enhancement of
activity is associated with the chemical affinity for
carboxylate.
Zell et al. investigated a highly efficient iron complex
catalyst system for FA dehydrogenation with TON up to
100000 in the presence of trialkylamines at 313 K [22]. Based on
their experiments, they observed that protonation of the iron
dihydride catalyst, followed by dihydrogen liberation, led to
an unsaturated species that was transformed into a
hydridoeformate complex. Owing to the elimination of CO2,
the iron dihydride catalyst was regenerated. According to the

DFT calculations, this process was forecasted to proceed by a
novel, non-classical intramolecular b-H elimination.

Copper-based catalysts
Recently, using simple copper complexes catalysts for FAD to
yield H2 in a HCOOH/amine mixture solution was reported by
Ravasio and co-workers [21]. While in the presence of
Cu(OAc)2 and 5:2 HCOOH/NEt3 adduct (NEt3 ¼ triethylamine),
the evolution gas which was tested by gas chromatography
showed that H2 and CO2 were formed in a 1:1 ratio with traces
of CO (<150 ppm). When they decreased the HCOOH/NEt3
ratio, they found that the amine concentration, the higher the
conversion. In addition, they observed an interesting and
obvious influence by varying the amine. Here the basicity
played an important role. Particularly, in the whole process
the higher basicity of amine was, the higher activity of the
catalysts had showed.

Heterogeneous catalysts for formic acid
decomposition
The decomposition of FA in presence of heterogeneous catalysts has been reported dating back to 1930s [42]. Many heterogeneous systems have been reported in the gas phase over
catalysts including metals, metal oxides, and metal supported
on carbon or metal oxides [43,44]. However, the reaction was
generally accomplished at high temperature (373 K), which
exceeded the boiling point of formic acid and thus making the
reaction occurred in gas phase. Therefore, the research conducted thereafter were mainly focused on developing efficient
heterogeneous catalysts to catalyze liquid phase FAD at
reduced temperatures. Noteworthy, in the late 1970's, Williams and co-workers successfully used Pd/C as the heterogeneous catalysts for FAD at room temperature [45].

Please cite this article in press as: Wang X, et al., Recent progress in hydrogen production from formic acid decomposition, International

Journal of Hydrogen Energy (2018), />

6

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 8 ) 1 e1 7

Table 1 e Selected heterogeneous catalysts for the dehydrogenation of formic acid.
Catalyst

Reaction conditions

Pd@SiO2
Pd/MSC-30
Pd/N-MSC-30
PdeB/C
Pd/r-GO
Pd/CN0.25
PdO/C
Pd/S-1-in-K
Pd-M(OH)2/S-1
PdAu/CeCeO2
AuPdeMnOx/ZIF-8-rGO
PdAg@ZrO2/C/rGO
Ag0.1Pd0.9/rGO
Co0.30Au0.35Pd0.35/C
PdNiAg/C
AueZrO2
Au@Schiff-SiO2
AuePt/CeO2
Pd @ CN

CoNx/C

FA
FA
FA
FA
FA
FA
FA
FA
FA
FA
FA
FA
FA
FA
FA
FA
FA
FA
FA
FA

þ
þ
þ
þ
þ

SF

SF
SF
SF
SF

þ SF
þ SF

þ SF
þ SF
þ SF
þ SF
þ NEt3
þ SF

þ specified solvent

TOF (hÀ1)

Mass activity (molH2 gÀ1Pd hÀ1)

T (K)

Reference

70
750
8414
1184
5420

5530
3172
3027
5803
e
382
4500
2739
80
85
1590
4368
1637
71
e

0.14
3.59
7.47
3.50
6.39
6.52
8.72
8.49
11.91
2.13
1.65
e
2.31
0.05

e
e
e
e
0.67
0.019 molH2 gÀ1Co hÀ1

365
298
333
303
353
298
303
333
333
365
298
333
323
298
323
323
323
433
288
383

[51]
[63]

[82]
[57]
[56]
[74]
[86]
[80]
[87]
[46]
[72]
[89]
[66]
[54]
[62]
[92]
[93]
[101]
[103]
[106]

Fig. 5 e The surface structure of the metal particle had great influence for the formic acid decomposition [14].

Nowadays, the dominant catalysts (Table 1) are based on the
noble metals such as palladium [14,46e90], gold [91e96], and
platinum [97e101]. In addition, some researchers are interested in photocatalytic [102e105] and non-precious metal
[106]for the dehydrogenation of FA.

The mechanism of formic acid decomposition on
heterogeneous catalysts
As shown in Scheme 1, two possible ways for FAD were reported in literature, in which the dehydrogenation pathway is

Fig. 6 e (A) The data from 13C NMR spectrum about the adsorption of FA and formate on PVP-Pd nanoparticles. (B) The
percent of three different adsorbed modes [107].
Please cite this article in press as: Wang X, et al., Recent progress in hydrogen production from formic acid decomposition, International
Journal of Hydrogen Energy (2018), />

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 8 ) 1 e1 7

more favorable in comparison to the self-poisonous CO
pathway. The surface structure of the metal particles were
reported to exert great influence on the selective catalysis for
FAD (Fig. 5) [14]. Fig. 5 shows that the formate was bridged on
the flat terrace of metal M activity sites to produce H2 and CO2,
while the b process exhibited that the formate was linked on
isolated or low coordinated M sites for the liberation of CO and
H2O.
Tsang and co-workers investigated that the unequal
sharing of bonding electrons around the 13C nucleus of the
adsorbate molecule on the metal surfaces gave rise to variations in 13C chemical shift values, which was correlated to the
adsorption states [107]. There were four resonance peaks
(Fig. 6A). They suggested that the three resonances peaks at
165.42, 165.69 and 165.95 ppm were assigned to three different
adsorbed modes closely related to monodendate, multimonodentate and bridging formate species, respectively. The
bridging formate species show the highest (Fig. 6B) implied
that the process of FAD was owing to the formation of bridging
formate intermediates. According to the Sabatier principle in
chemical catalysis, it described that the interaction between
catalyst and substrate was appropriate.
Furthermore, DFT calculations were carried out to predict
the catalytic behavior of varied metal surfaces using d-band
center model. Studt, Nørskov and co-workers used a theoretical analysis to identify alternative catalyst materials for the

dehydrogenation of formic acid [108]. According to their
theoretical study, they found that Au (211) was suggested to be
less active than Pt (111) and Pt (211), because it lay far out on
the weak-binding side of the activity volcano (Fig. 7). However,
Ojeda et al. [91] and Cao et al. [92] had observed that welldispersed Au nanoparticles supported on metal oxide
exhibited superior performance for the dehydrogenation of
FA. The difference between experiment and theoretical analysis makes the FAD mechanism still open for discussion.

Palladium-based catalysts
Nowadays, heterogeneous catalysis for the decomposition of
formic acid is mainly based on the palladium-based catalysts.
While some of the endeavors were focused on modulating the
There are some works to illustrate the catalysts morphology

7

such as core-shell nanostructure of the catalysts [14] to acquire higher catalysts utilization, others e. Furthermore, there
are some catalysts studied the focusing on the alloy to reduce
the price of the catalysts. In effect addition, some catalysts
were designed to reduce the surface electron structure of
metal palladium to improve to boost the intrinsic catalytic
performance by altering the surface chemical and electronic
structure [55,57,72,74]. Xu et al. used the Pd(NH3)4Cl2 as precursor and the NaBH4 as the reducing agent in a
polyoxyethylene-nonylphenyl ether/cyclohexane reversed
micelle system to obtain the Pd@SiO2 [51]. The Pd@SiO2 catalyst showed high performance for the liberation of H2 from
aqueous solution of FA and sodium formate (SF) at 365 K. In
addition, they had observed the interactions between Pd and
silica supports for the catalytic performance. In the following
few years, their group reported the Pd nanoparticles on
nanoporous carbon MSC-30 [63], Pd nanoparticles

(diameter 1.5 nm) on the diamine-alkalized graphite oxide
(rGO) [71] and palladium nanoclusters immobilized by a
nitrogen-functionalized porous carbon [82] for FAD. The catalysts with different size of the Pd nanoparticles and supporter at mild condition showed different catalytic activities.
In other words, the activity of catalyst was affected by carriers
and metal nanoparticles.
For the first time, Cai et al. investigated a boron-doped Pd
nanocatalyst for accelerating hydrogen production from formic acid and formate solutions [57]. The boron-doped Pd
catalyst showed excellent catalytic performance with the TOF
up to 1184 hÀ1 at 303 K. In order to reveal the high activity of
PdeB/C catalyst, they used the real-time ATR-IR spectroscopy
and found that the exceptional performance of PdeB/C
correlated well with an apparently impeded COad accumulation on its surfaces. Recently, they controlled the size of
catalyst by selective addition of different alkaline solution
(Na2CO3, NH3$H2O, or NaOH) to Pd (II) solution to obtain the
size-controlled catalysts [83]. They found that the Pd/C catalysts with smaller Pd particle sizes were highly active for the
liberation of hydrogen from a FA and SF solution of pH 3.5 at
room temperature.
Cao and co-workers studied the Pd nanoparticles anchored
on graphite oxide nanosheets (r-GO) catalyst for both aqueous
formate dehydrogenation and bicarbonate hydrogenation

Fig. 7 e Theoretical activity volcanoes for (a) H2 þ CO2 production and (b) H2O þ CO production from formic acid [108].

Please cite this article in press as: Wang X, et al., Recent progress in hydrogen production from formic acid decomposition, International
Journal of Hydrogen Energy (2018), />

8

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 8 ) 1 e1 7

[56]. The Pd/r-GO catalyst was used for the dehydrogenation of
potassium formate solution with the TOF up to 5420 hÀ1 at
353 K. When the temperature and H2 pressure changed, the
same catalyst could be used to completely reduce the KHCO3
to HCOOK. Later, their group used Pd coupled on pyridinicnitrogen-doped carbon (CNx) as the robust and efficient solid
catalyst for the liberation hydrogen from fromic acid and the
Pd/CN0.25 exhibited high performance with the TOF up to
5530 hÀ1 at 298 K [74]. Based on their experiment, the Pd/CNx
showed high performance was due to a possible electron
transfer from CNx to Pd nanoparticles. More importantly, the
data from ATR-IR spectroscopy showed that the N content of
CNx supports had a strong electronic effect on Pd
nanoparticles.
Chen et al. investigated a MotteSchottky catalyst for the
decomposition of FA solution ambient condition [103]. This
novel MotteSchottky catalyst was based on Pd nanoparticles
and g-C3N4 (Pd@CN). The carbon nitride was both semiconductive support and the stabilizer for the coupling of metal
nanoparticles
to
form
the
MotteSchottky
nanoheterojunctions. The TOF was value up to 49.8 mol H2
molÀ1Pd hÀ1 at 288 K (Fig. 8). However, when under photoirradiation (l ! 400 nm) the TOF was elevated to 71 mol H2
molÀ1Pd hÀ1 at 288 K.
In 2015, our group used the FA as reducing agent and
H2PdCl4 solution as the precursor solution to in situ generated
Pd/C catalyst in ambient conditions for both dehydrogenations of FA and FA electrooxidation [70]. While the
forming gas from FAD without CO directly used in fuel cells,
the power density of the forming gas was 80 mW cmÀ2.

Recently, we for the first time revealed the important role of
PdO in determining the FAD performance [86]. Through XPS
analysis, a positive correlation between the FAD performance
and the content of PdO has been found. To clarify the real
effect of PdO, a series of experiment was carried out (Fig. 9).
Time-evolved ATR-IR spectra show that PdO/C had an excellent antipoisoning effect than Pd/C (3.6 nm) catalyst. DFT
calculation shows that PdO can help pulling hydrogen in the
formic acid molecular to release CO2 and restraining the

Fig. 8 e Decomposition of FA over different catalysts at
288 K [103].

dehydration pathway, which not only accelerated the reactivity, but also promoted the selectivity. Besides, the chemical
block technique demonstrates the adsorption sites was Pd,
which means the FA adsorption occurs over Pd sites, and be
accelerated by the bordered PdO. Therefore, PdePdO interface
is believed as active site for FA dehydrogenation. A novel ultrasmall Pd clusters anchored on nanosized silicalite-1 zeolite
by in situ confinement was reported by Yu and co-workers [80].
They used the [Pd(NH2CH2CH2NH2)2]Cl2 as precursor by direct
hydrothermal method to synthesis well-dispersed and ultrasmall Pd clusters in nanosized silicalite-1 zeolite. The
catalyst showed excellent activity for H2 generation with no
CO formation. The TOF was value to 856 hÀ1 at 298 K and
3027 hÀ1 at 333 K, respectively. Lately, by using hydrothermal
synthesis method, they synthesized subnanometric hybrid
Pd-M(OH)2 (M ¼ Ni, Co) clusters which were encapsulated in
siliceous zeolites for FAD [87]. The catalyst performed excellent catalytic properties (TOF up to 5803 hÀ1) at 333 K without
any additive.

Palladium-based bimetallic catalysts
Our group synthesized PdeAg/C and PdeAu/C catalysts for the

effective H2 production from FA at 365 K [46]. We found that
the initial reaction rate was extraordinarily fast and reforming
gas in the first 5 min was 31% of the total reforming gas in 2 h.
Furthermore, the performance of the PdeAg/C and PdeAu/C
catalysts were accelerated greatly by co-deposition with CeO2
(H2O) x. There might be two possible reasons for the CeO2 to
promote the Pd-based catalytic activity. One was probably
that more cationic palladium species were produced to
oxidize CO in the presence of the CeO2. Another was that CeO2
(H2O) x on the Pd surface might enhance the reaction 1. Later,
we had investigated the promotion effect of three rare earth
elements (Dy, Eu, and Ho) on the PdeAu/C catalysts [48]. The
PdeAueDy/C was the most active catalyst with the rate of
1198 mL minÀ1 gÀ1 Pd and the TOF of 269 ± 202 hÀ1 than the
PdeAueEu/C and PdeAueHo/C at 365 K. We suggested that
the promotion effect was likely due to the capability of rare
earth elements to provide abundant oxygen species to act
with the poisonous intermediates.
A novel metal organic framework immobilized AuePd
nanoparticles for decomposition of FA was studied by Xu and
co-workers in 2011 [49]. Owing to its large pore sizes, window
sizes and hybrid pore surface, MIL-101 was chosen as the
support for encapsulation of metal nanoparticles. The AuePd/
ethylenediamine-grafted MIL-101 showed high catalytic activity at 363 K. The addition of Au improved the high tolerance
of AuePd catalysts to CO poisoning. In the following few years,
monodisperse AuPd alloy nanoparticles with controlled
composition [52] and nitrogen-doped graphene as the carrier
to support the AuPdeCeO2 [60] for the dehydrogenation of
formic acid was reported by their group. In 2015, Yan et al.
reported a ZIF-8-reduced-graphene-oxide (ZIF-8erGO) bisupport to immobilize AuPd-MnOx nanocomposite for FAD

at room temperature [72]. They used a wet-chemical method
to synthesis the AuPd-MnOx/ZIF-8-rGO catalyst (Scheme 2).
The catalyst exhibited excellent catalytic activity and the
initial TOF was value to 382.1 mol H2 mol catalystÀ1 hÀ1.
Compared to the AuPdeMnOx/C, the AuPd-MnOx/ZIF-8-rGO
showed higher catalytic activity was due to strong metal-

Please cite this article in press as: Wang X, et al., Recent progress in hydrogen production from formic acid decomposition, International
Journal of Hydrogen Energy (2018), />

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 8 ) 1 e1 7

9

Fig. 9 e The In situ physical characterization and the possible kinetic calculation for the whole reaction process and the
calculation about different metals for FAD under this conditions [86].

Scheme 2 e The whole process to synthesis the AuPdeMnOx/ZIF-8erGO composite [72].

Please cite this article in press as: Wang X, et al., Recent progress in hydrogen production from formic acid decomposition, International
Journal of Hydrogen Energy (2018), />

10

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 8 ) 1 e1 7

support interaction between ZIF-8erGO and active nanoparticles. Recently, Song et al. reported a well-dispersed PdAg
catalyst for the dehydrogenation of FA [89].They chose the
zirconia/porous carbon/reduced graphene oxide nanocomposite derived from metal organic framework/graphene
oxide to anchor PdAg nanoparticles. The PdAg@ZrO2/C/rGO

showed high catalytic performance for FAD and the TOF
valued up to 4500 hÀ1 at 333 K.
In 2013, Sun and co-workers investigated a monodisperse
AgPd alloy nanoparticles the dehydrogenation of HCOOH at
323 K [55]. Different composition of the AgPd nanoparticles
were synthesized by changing the molar ratio of Ag/Pd. The
Ag42Pd58 showed the highest catalytic performance and the
TOF was 382 hÀ1. Based on their experiment, they suggested
that different composition of AgPd exhibited different performance was due to the drastic alloy effect. Almost the
same time, Ag0.1Pd0.9/rGO was reported by Yan et al. [53] and
Xu et al. [66] Yan et al. used a simple co-reduction method to
obtain Ag0.1Pd0.9 nanoparticles assembled on rGO [53]. The
synergistic coupling between Ag0.1Pd0.9 and rGO made the
catalyst with the high catalytic activity (TOF up to 105.2 mol
H2 molÀ1 catalyst hÀ1) for FAD at ambient temperature. Xu
et al. used a non-noble metal sacrificial approach to
immobilize the AgPd alloy nanoparticles on rGO [66]. The
cobalt compound was co-precipitated during the reduction
of precursors to prevent the noble metal nanoparticles from
aggregation and then the non-noble metal was sacrificed by
acid to obtain the Ag0.1Pd0.9/rGO catalyst. The catalyst
showed high performance with the initial TOF up to
2739 hÀ1 at 323 K. In 2015, Zahmakiran and co-workes used
a facile impregnation method followed by sodium borohydride reduction to get the PdAg alloy and MnOx nanoparticles supported on amine-grafted silica catalyst [64]. The
catalyst was for the liberation of H2 from FA with high activity (330 mol H2 mol catalystÀ1hÀ1) without any additives
at ambient condition. Recently, the amine-functionalized
UiO-66 modified AgePd alloy and AgPdeMnOx supported
on carbon nanospheres for the production of H2 from FA
were studied by Wang and co-workers [76]. Owing to the
different carrier, the AgPd alloy showed different catalytic

performance for the dehydrogenation of FA. The interaction
between alloy and carriers makes it possible to synthesis
kind of catalysts to applicate in high-performance metal
nanocatalysts.

Palladium-based core-shell catalyst
Our group reported a novel PdAu@Au/C core-shell catalyst for
FAD at 365 K in 2010 [47]. This special nanostructure was
synthesized by a facile reduction method in the absence of
stabilizer in Fig. 10. The stable core-shell structure was formed
by both the high miscibility of Pd and Au and the proper molar
ratio of metal precursors. The catalytic performance was
tested in a test tube which contained 5 mL of 6.64 M formic
acid, 6.64 M sodium formate, and 60 mg of catalyst at 365 K
and found that the PdAu@Au/C showed highest catalytic activity (Fig. 10). The reforming gas was tested by FT-IR spectroscopy and found that the CO was determined to 30 ppm.
An interesting AgePd core-shell nanocatalyst for FAD at
room temperature was studied by Tsang and co-workers in
2011 [14]. By using the wet chemical synthesis, the ultrathin
Pd shell on Ag core nanoparticles was obtained to enhance the
H2 production from HCOOH without CO generation at ambient
condition (Fig. 11). As the temperature increased, the rate of
reaction would be increased. However, the concentration of
CO would be more than 74 ppm when the solution was heated
above 343 K. More importantly, the author used the atom
probe tomography to confirm the coreeshell configuration
and found that the shell contained between 1 and 10 layers of
Pd atoms. Based on their experiments, they suggested that the
electronic promotion by underlying Ag had a short range of
few atomic distances.
TiO2-supported AgPd@Pd nanocatalysts were studied by

Hattori and co-workers for formic acid dehydrogenation to
produce H2 in 2015 [67]. The formation of AgePd bimetallic
nanocatalysts were synthesized by a two-step microwavepolyol method with an average diameter of 4.2 ± 1.5 nm.
Compared to the AgPd@Pd, the AgPd@Pd/TiO2 showed higher
catalytic activity and the hydrogen production rate was
16.0 ± 0 0.89 L gÀ1 hÀ1 at 300 K. Based on their experiments,
they suggested that the higher catalytic performance of the
AgPd@Pd in the presence of TiO2 was owing to the strong
electron-donating effects of TiO2 to Pd shells leading to
enhance the adsorption of formate to the catalysts surface
and decomposition from formate. Furthermore, they considered that the formate was adsorbed on the catalysts to form
the bidentate formate and then the bidentate formate
decomposed to CO2* þ H* after that the recombination of H*
and CO2* formed the H2 and CO2. Almost the same time, they
used the similar method to synthesis a series of AgPd@Pd/TiO2

Fig. 10 e The formation of core-shell structure catalysts (a), using the core-shell structure catalysts for the decomposition of
FA at 92  C (b) [47].

Please cite this article in press as: Wang X, et al., Recent progress in hydrogen production from formic acid decomposition, International
Journal of Hydrogen Energy (2018), />

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 8 ) 1 e1 7

11

Fig. 11 e The data from the atom probe tomography to confirm the coreeshell configuration [14].

catalysts with a lower Pd content at lower temperature [68].
The catalytic activity of Ag100-xPdx@Pd/TiO2 (x ¼ 7, 10, and 15)

were higher than Ag82Pd18@Pd/TiO2 at 300 K. Noteworthy,
with the decrease of Pd content, the rate of hydrogen production increased. They suggested that the extent of alloying
of Ag and Pd in the AgPd would have an influence on the
catalytic activity.

Palladium-based trimetallic catalyst
In order to improve the catalyst activities and decrease the
costs of catalysts, the polymetallic nanomaterials containing
first-row transition metals and noble metals were synthesized
for the FA dehydrogenation. Jiang and co-workers reported an
efficient CoAuPd/C catalyst for H2 generation FA at ambient
condition [54]. The author used the fresh NaBH4 aqueous solution for the reduction of the mixture solution containing
Co2þ, Au3þ, and Pd2þ to obtain the Co0.30Au0.35Pd0.35/C catalysts. In addition, they also synthesized a series of different
raw material ratio catalysts. However, the Co0.30Au0.35Pd0.35/C
showed the best catalytic activity (the initial TOF up to 80 hÀ1
at 298 K).
Trimetallic PdNiAg nanoparticles supported on activated
carbon catalytic system was obtained by Yurderi et al. [62] The
trimetallic catalyst was synthesized by wet-impregnation
followed by simultaneous reduction method without any
stabilizer at room temperature. The PdNiAg/C showed high
catalytic (TOF ¼ 85 hÀ1) and selectivity (nearly to 100%) at

323 K. More importantly, the PdNiAg nanoparticles showed
the exceptional stability to make it possible for the reusing of
FA dehydrogenation. Based on their quantitative kinetic
studies, they suggested that the trimetallic catalytic system
for FA dehydrogenation was first-order in catalyst concentration, half-order in sodium formate concentration and zeroorder in FA. Later, the same group reported the amine grafted
silica supported CrAuPd for FAD [73]. They used the same
synthetic method to synthesis the compound catalyst with 3aminopropyltriethoxysilane functionalized silica as the carrier to support CrAuPd alloy nanoparticles. The catalyst

showed high catalytic performance with an original TOF up to
730 mol H2 mol catalystÀ1 hÀ1.
Luo and co-workers investigated the monodisperse
CoAgPd nanoparticles assembled on graphene for yielding H2
with no CO formation from HCOOH at ambient condition [81].
Using the 2-methylpyridine borane as the reducing agents, the
mixture solution containing Co2þ, Agþ, and Pd2þ was reduced
to obtain the monodisperse trimetallic CoAgPd alloy nanoparticles at 363 K. Different compositions of the trimetallic
CoAgPd nanoparticles were obtained to test for the catalytic
performance. The Co1.6Ag62.2Pd36.2/graphene showed highest
catalytic performance and the TOF was up to 110 hÀ1. They
had also studied different molar ratio of FA/SF by the trimetallic CoAgPd catalyst and found that the catalyst showed
highest activity with 9:1 FA/SF molar ratio.

Please cite this article in press as: Wang X, et al., Recent progress in hydrogen production from formic acid decomposition, International
Journal of Hydrogen Energy (2018), />

12

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 8 ) 1 e1 7

Scheme 3 e The possible pathways for FAD on well-dispersed Au catalyst [91].

Gold-based catalysts
In 2009, Ojeda and Iglesia reported an Au-Based Catalyst for
the decomposition of FA at near-ambient condition (about
350 K) [91]. They observed that Au/Al2O3 showed better catalytic activity for FAD to produce H2 and CO2 with a trace of CO
(<10 ppm) when used different carriers to carry the welldispersed Au species. Based on the kinetic isotope effects,
they suggested that there were two possible mechanisms On
Au/Al2O3 (Scheme 3). According to their data, they suggested

that the mechanism A1, for which rate constants reflect rates
of DCOO* and H* (or HCOO* þ D*) in HD formation, was more
likely the pathways for FAD on dispersed Au species.
In 2012, Cao et al. had reported a catalytic system, the
subnanometric gold and an acid-tolerant oxide support for the
production of CO-free H2 for FAD, which showed high catalytic
activity (TOF up to 1590 hÀ1 and TON more than 118400 at
323 K) [92]. The whole reaction could be induced both at
ambient condition and higher temperature. The H2 evolution
can be obtained at room temperature. Based on their preliminary mechanism studies, they suggested that the reaction
was unimolecular in nature and proceeded through a unique
amine-assisted formate decomposition mechanism on Aue
ZrO2 (Scheme 4). In the process for FAD, the cooperation between Au clusters and ZrO2 matrix played an important role
for the essential AuÀformate activation under mild conditions. Later, the same group investigated that Au nanoparticles supported on ZrO2 in FA-dimethylethanolamine
system could be efficient for H2 evolution (TOF up to

1166 hÀ1) at 333 K [94]. The amine as the proton scavenger
could promote the OeH bond cleavage in the crucial step of FA
deprotonation in the interface of AueZrO2 leading in high
catalytic performance. They also studied that the amine
concentration could affect the hydrogen generation rate.
There was low catalytic performance for H2 production
without amine and an increasing concentration was beneficial
for H2 evolution.
A novel Schiff base modified Au catalyst for efficient H2
evolution from FA was studied by Liu and co-workers in 2015
[93]. The catalytic system showed high catalytic activity for H2
production in high-concentration FA in the absence of any
additives. The TOF could value to 4368 hÀ1 in high concentration FA solution, and was as high as 2882 hÀ1 even in nearpure FA at 323 K. For the first time, the rational designed
Au@Schiff-SiO2 catalyst was directly used in the pure FA.

Recently, Kegnæs and co-workers investigated that using
zeolite incorporated gold nanoparticles for hydrogen evolution from vapor phase decomposition of formic acid [95]. By
pressure assisted impregnation and reduction, the catalyst
system was prepared and resulted in a uniform distribution of
small gold nanoparticles that were incorporated in zeolite
silicalite-1 crystals. All the catalysts were tested in vapor
phase for the decomposition of FA to produce H2 and CO2 with
no formation of CO. Almost the same time, the same group
studied that silica encapsulated and amine functionalized
gold nanoparticles showed highly catalytic activity in vapor
phase for the evolution of H2 from FA [96]. In order to control
the particle size distribution and shell thickness, the Au@SiO2
core-shell catalysts were prepared in a reverse micelle system.
The smallest gold nanoparticles showed better catalytic activity (TOF up to 958 hÀ1) at 403 K. Based on the kinetic isotope
labelling experiments, they suggested that H-assisted formate
decomposition into CO2 and H2 was the rate limiting step. Due
to the high reaction temperature, formic acid was heated in
vapor phase. The protection atmosphere (Ar or N2) might be
pumped into the pure formic acid vapor which might limit the
development of FAD.

Platinum-based catalysts

Scheme 4 e The possible reaction pathway over the
Au/ZrO2 catalyst for hydrogen production from the TEAF
system [92].

A solid carbon supported PtRuBiOx catalyst for Liquid-phase
FAD to yield CO-free H2 was observed by Chan et al., in 2012
[97]. Different FA concentrations and temperatures were

tested for the kinetic experiments. Under the investigated
conditions, they found that the kinetics had half-order to FA
and first-order dependence for formate. Based on their

Please cite this article in press as: Wang X, et al., Recent progress in hydrogen production from formic acid decomposition, International
Journal of Hydrogen Energy (2018), />

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 8 ) 1 e1 7

experimental results, two possible mechanisms with
adsorption of FA or formate were suggested. The two possible
mechanisms of HCOOH or HCOOÀ adsorption could be used
for the explanation of the variation in apparent activation
energy at different FA concentrations.
In 2012, Bulushev and co-workers studied Pt nanoclusters
stabilized by N-doped carbon nanofibers for H2 evolution
from vapor-phase FA [98]. The N-doped catalysts with low Pt
contents showed higher catalytic activity than the undoped
catalysts. They suggested that the increased catalytic activity were as the follow: the higher dispersion of Pt on N-doped
material, the appearance of new active sites on metalsupport interface and the strong interaction on metalesupport. The N-doped catalysts and the undoped catalysts
were also used for the ethylene hydrogenation, but it
exhibited diametrically different results that the undoped
catalysts showed greater catalytic activity. Later, the same
group used different noble metals such as Pd, Pt, and Ru as
the precursor to synthesize a series of single atoms catalysts
for the decomposition of vapor-phase FA [99]. Among all the
single atoms catalysts, the N-doped carbon nanofibers supporting single atoms of Pt-group metals showed the best
catalytic performance. The N-doped carbon nanofibers were
considered as a stable and efficient macro-ligand, which
possessed great chelate properties and successfully combined the advantages of both heterogeneous and homogeneous catalysts.

An interesting PteCu single-atom alloys for the selective
formic acid dehydrogenation was investigated by Marcinkowski et al. [100] They studied the reaction of formic acid on
single-atom alloys which were consisted by single Pt atoms
substituted into a Cu lattice. In their studies they found that
single atom alloys of PteCu improved the conversion of formic
acid to formate and maintained the high selectivity of Cu. In
addition, they also had demonstrated the spillover of formate
species from Pt sites to Cu. Furthermore, the author had
confirmed that single Pt atoms could be able to lower the
barrier to dehydrogenation of formic acid to formate at low
temperature, but it did not increase the decomposition of
formate to CO2 and H2.
Recently, wang and co-workers reported an AuePt alloy
catalytic systems which showed not only great selectivity for
FAD but also good selectivity for the hydrogenolysis of
carboneoxygen species in the hydrodeoxygenation processes at 423 K [101]. The AuePt alloy nanoparticles supported on CeO2 (AuePt/CeO2) exhibited excellent catalytic
activity (TOF up to 1637 hÀ1) at 433 K. In addition, the catalysts showed great selectivity (up to 99.8%) for the reduction
of vanillin to 2-methoxy-4-methylphenol. In their reaction
systems, the FA was the hydrogen sources to provide enough
hydrogen on the catalysts surface for the reduction of
vanillin.

Other catalysts
In 2012, Halasi and co-workers studied N-modified TiO2 catalytic systems for the photolysis of HCOOH for the production
of pure H2 [102]. The pure, N-doped and Rh-promoted TiO2
were used for the photo-induced vapor-phase decomposition
of FA. Owing to the N incorporation lowering the bandgap of

13

TiO2, the catalysts promoted the photolysis of FA. Based on
their experiments, they found that the N-doped TiO2
increased the photo activity in the FAD at 300 K. Furthermore,
the extent of the photodecomposition of FA was enhanced
with the depositing of Rh on pure or N-doped TiO2.
An interesting cadmium sulfide quantum dots catalytic
system used for the photocatalytic formic acid with controllable selectivity for H2 or CO at room temperature was
observed by Reisner and co-workers in 2015 [104]. In the
presence of ligand-capped and cobalt co-catalyst, the TOF was
À1
value up to 116 ± 14 mmol H2 gÀ1
cath . However, in the absence
capping ligands, the same cadmium sulfide generated up to
À1
102 ± 13 mmol CO gÀ1
cath . In addition, the author also calculated the turnover numbers and the turnover numbers were
more than 6 Â 105 and 3 Â 106, respectively. Although formic
acid was used as the hydrogen storage material, the photocatalytic production of CO based on CdS nanocrystals also
allowed formic acid to act as a renewable storage material
for CO.
In order to reduce the cost of the catalyst, more and more
researchers begin to study the non-noble metal catalyst for
the catalytic decomposition of FA. Recently, Beller and coworks studied a series of cobalt supported on carbon heterogeneous catalysts for the FAD [106]. Owing to the strong coordination capability of cobalt, they chose Co(OAc)2$4H2O as
the cobalt source and a series of Nitrogen-containing organic
compounds as the N ligands to obtain the precursor, followed
with high temperature calcination to obtain the final catalysts. The highest catalytic performance achieved was
À1
423.3 mL gÀ1
cat h , which is the first report of using non-noble
metals for the heterogeneous FAD catalysis. According to

their XPS results and poisoning experiments, they claimed
that the CoNx centers supported on C were the active sites for
the decomposition of formic acid. This newly generated
cobalt-based catalyst opened a new pathway for developing
FAD heterogeneous catalysts.

Summary and outlook
Owing to the high gravimetric and volumetric hydrogen capacity, FA is a promising route for hydrogen energy room
temperature storage and release. Both homogeneous catalysts and heterogeneous catalysts are used for the dehydrogenation of formic acid. The homogeneous catalysts such as
ruthenium, iridium, iron and copper based catalysts show
promising FAD performance. However, the homogeneous
catalysts encountered difficulty in industrial hydrogen production as the reaction process take place in liquid solventacid mixture. It is hard to reuse and recycle the catalysts
for the whole reaction. In addition, in order to obtain the
homogeneous catalysts, lots of organic solvents containing
phosphine needs to be used which is harmful to the environments. Compared to the homogeneous catalysts, the
heterogeneous catalysts will be easy for reusing and recycling. However, the heterogeneous catalysts show lower
catalytic performance. Furthermore, the FAD mechanism on
heterogeneous catalysts is still unclear. Therefore, understanding of reaction mechanism at the interface is necessary
in the following researches. In addition, TOF is now

Please cite this article in press as: Wang X, et al., Recent progress in hydrogen production from formic acid decomposition, International
Journal of Hydrogen Energy (2018), />