Efficient production of hydrogen from formic acid using a covalent triazine framework supported molecular catalyst

Article abstract of DOI:10.1002/cssc.201403173

Abstract A heterogeneous molecular catalyst based on Ir^IIICp (Cp=pentamethylcyclopentadienyl) attached to a covalent triazine framework (CTF) is reported. It catalyses the production of hydrogen from formic acid with initial turnover frequencies (TOFs) up to 27 000 h^-1 and turnover numbers (TONs) of more than one million in continuous operation. The CTF support, with a Brunauer-Emmett-Teller (BET) surface area of 1800 m²g^-1, was constructed from an optimal 2:1 ratio of biphenyl and pyridine carbonitrile building blocks. Biphenyl building blocks induce mesoporosity and, therefore, facilitate diffusion of reactants and products whereas free pyridinic sites activate formic acid towards β-hydride elimination at the metal, rendering unprecedented rates in hydrogen production. The catalyst is air stable, produces CO-free hydrogen, and is fully recyclable. Hydrogen production rates of more than 60 molL^-1h^-1 were obtained at high catalyst loadings of 16 wt% Ir, making it attractive towards process intensification.

Full text of DOI:10.1002/cssc.201403173

DOI: 10.1002/cssc.201403173
Communications
Efficient production of hydrogen from formic acid using
a Covalent Triazine Framework supported molecular
catalyst
A. V. Bavykina, M. G. Goesten,* F. Kapteijn, M. Makkee, and J. Gascon*[a]
A heterogeneous molecular catalyst based on IrIIICp* (Cp*=
pentamethylcyclopentadienyl) attached to a covalent triazine
framework (CTF) is reported. It catalyses the production of hy-
drogen from formic acid with initial turnover frequencies
(TOFs) up to 27000 hꢀ1 and turnover numbers (TONs) of more
than one million in continuous operation. The CTF support,
two important parameters in device-based applications (for ex-
ample, fuel cells). To date, the task of developing a heterogene-
ous catalyst that performs well in hydrogen production from
formic acid has proven challenging. In comparison with homo-
geneous catalytic systems, suppression of the undesired dehy-
dration reaction (HCOOH!H2O+CO)[12] remains an obstacle,
and TOFs and corresponding rates per unit reactor volume are
often low. Whereas the majority of these attempts focus on
the use of nanoparticles,[13–17] promising reports involve the
use of molecular catalysts.[18,19] Here, particular progress has
come from the group of Laurenczy, who immobilised their ho-
mogeneous Ru-mTPPTS catalyst on various supports to devel-
op a heterogeneous molecular catalyst.[20] As support, mesopo-
rous silica worked best, with a maximum TOF of 2780 hꢀ1 at
110 8C.[21] This work demonstrates the potential of ‘heteroge-
nised’ molecular systems in catalysis.
with
a
Brunauer–Emmett–Teller (BET) surface area of
1800 m2 gꢀ1, was constructed from an optimal 2:1 ratio of bi-
phenyl and pyridine carbonitrile building blocks. Biphenyl
building blocks induce mesoporosity and, therefore, facilitate
diffusion of reactants and products whereas free pyridinic sites
activate formic acid towards b-hydride elimination at the
metal, rendering unprecedented rates in hydrogen production.
The catalyst is air stable, produces CO-free hydrogen, and is
fully recyclable. Hydrogen production rates of more than
60 molLꢀ1 hꢀ1 were obtained at high catalyst loadings of
16 wt% Ir, making it attractive towards process intensification.
Herein, we report a highly active, selective and air-stable
molecular heterogeneous catalyst based on a covalent triazine
framework (CTF),[22–24] a porous type of organic polymer syn-
thesised from inexpensive feedstocks that had earlier been
successfully applied to immobilise the molecular Periana cata-
lyst for methane-to-methanol oxidation.[25] Here, IrIIICp* (Cp*=
pentamethylcyclopentadienyl) is coordinated to a mesoporous
CTF [Brunauer–Emmett–Teller (BET) surface area: 1800 m2 gꢀ1]
constructed using a 1:2 ratio of 2,6-pyridinedicarbonitrile and
4,4’-biphenyldicarbonitrile. The coordination was confirmed by
X-ray photoelectron spectroscopy (XPS).[26–29] In a subsequent
step, the counterion Clꢀ was replaced by triflate, OTfꢀ,
a weakly coordinating anion. This was performed in DMF, a sol-
vent known for binding HꢀCl[30] instead of the conventional
method using AgCl, which would result in precipitation within
the CTF pores. The obtained heterogeneous catalyst
1 (Scheme 1), is highly active for the production of hydrogen
whilst being air stable and fully recyclable.
The use of formic acid as a convenient material for hydrogen
storage is increasingly gaining attention in the development of
a hydrogen economy.[1,2] The main advantages of formic acid
over proposed alternatives in the frame of a sustainable
energy cycle include easy handling, refuelling and transporta-
tion,[3] price (600–1250$ per tonne), as well as the possibility of
synthesising it through electrochemical reduction of CO2 using
water as hydrogen donor.[4]
Although the decomposition of formic acid to yield hydro-
gen and CO2 is thermodynamically favourable (HCO2H!H2 +
CO2, DGo =ꢀ32.8 kJmolꢀ1), efficient H2 release is only obtained
with the use of a catalyst. The most active catalysts are homo-
geneous and transition-metal based for which impressive turn-
over frequencies (TOFs) have been documented[5,6] even for
a base metal such as iron.[7–10]
The requirement of a heterogeneous catalytic system has
been questioned,[11] yet the aim of developing an air-stable,
solid catalyst that produces hydrogen seems justified as it
would certainly present advantages in handling and recycling,
When dispersed in a 3m formic acid solution of pH 1.5 at
808C, 1 instantaneously produces a large flow of gas, which
was determined by GC analysis to be a CO-free CO2/H2 (1:1)
mixture.
The catalyst could be recycled for at least four times under
standard reaction conditions (808C) without any loss of catalyt-
ic activity.[26] During recycling, the catalyst was filtered and
stored under ambient conditions, only requiring an oxygen-
free environment during its synthesis. XPS also indicates that
the oxidation state of Ir remained +3 between runs whereas
elemental analysis indicated that Ir leaching was negligible.
At a loading of 0.2 wt% and at 808C, the catalyst reached
an initial TOF of 27000 hꢀ1, by far the highest reported for any
[a] A. V. Bavykina, M. G. Goesten, Prof. F. Kapteijn, Prof. M. Makkee,
Prof. J. Gascon
Catalysis Engineering–ChemE
Delft University of Technology
Julianalaan 136, 2628BL, Delft (The Netherlands)
E-mail: m.g.goesten@tudelft.nl
j.gascon@tudelft.nl
Supporting Information for this article is available on the WWW under
http://dx.doi.org/10.1002/cssc.201403173.
ChemSusChem 0000, 00, 0 – 0
1
ꢀ 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Communications
Table 1. Catalytic performance of 1 in production of H2 from formic acid.[a]
Catalyst Conditions
Time to 99%
Initial TOF
loading [wt%] T [8C] cformic acid [molLꢀ1] conversion [min] [hꢀ1
]
1
1
1[c]
1
1
1
1
2[e]
16
16
16
16
16
2
0
30
80
80
80
80
80
80
3
3
3
0.5
3600[b]
1100
<20
<5
160
70
120[d]
120[d]
70
1350
3960
2000
7900
21300
27000
4020
12
3
3
3
0.2
2.7
[a] Standard conditions, unless specified differently: 30 mL 3m formic acid so-
lution, 300 mg catalyst, 808C, pH 1.5; wt% refers to the metal loading on the
CTF. [b] Max. conversion 65%. [c] The initial TOF on the fourth recycle was
3930 hꢀ1. [d] Max. conversion 65%. [e] 2 is the Ruthenium(II)-based equivalent
of 1, using the benzene (h6-C6H6) ligand instead of Cp*.
After reaction, a typical IrꢀH stretching at 2050 cmꢀ1 was
identified performing diffuse reflectance infrared spectroscopy
(DRIFTS) on the spent catalyst.[33]
As indicated in Table 1, a less expensive catalytic system can
be obtained by attaching RuII(h6-C6H6), 2, to CTF instead of
IrIIICp*, but the corresponding catalytic activity is lower.
It is reasonable to assume that for loadings lower than
0.2 wt% even higher TOFs can be obtained, but this goes at
the cost of net productivity per unit reactor volume (or per
mass of catalyst). Although the value of the TOF as parameter
for catalytic efficiency should certainly not be underestimated,
the highest TOFs are usually obtained for very low concentra-
tions of active sites per unit mass of catalyst, which translates
to very large reactor dimensions to obtain reasonable total
production. Since intensification of hydrogen production is
clearly vital for device-based applications, we investigated the
effect of catalyst metal loading, formic acid concentration and
reactor temperature on (molar) hydrogen production rate per
unit volume. The results are shown in Figure 1. Although TOF
values decrease for high loadings (Table 1), the trend in
Figure 1 indicates that intensifying the hydrogen production
through higher metal loadings and formic acid concentrations
can be an attractive option. This is an important asset, for ho-
mogeneous systems frequently suffer from dramatic drops in
performance and/or catalyst deactivation upon increase of cat-
alyst concentration.[34]
Scheme 1. Synthesis of the CTF through triazination, with corresponding at-
tachment of IrCp*(H2O) and counterion exchange. The methyl groups of Cp*
are omitted for the sake of clarity. The CTFs are depicted schematically, and
no long-range order was observed using X-ray diffraction (XRD).
heterogeneous catalytic system to date for this reaction
(Table 1). At 308C and a loading of 16 wt%, the catalyst re-
mained active at an initial TOF of 1350 hꢀ1 and even at 08C
the catalyst still displayed an initial TOF of 70 hꢀ1.
To test its suitability for device-based application, the cata-
lyst’s durability was tested in continuous mode, in which
a highly concentrated formic acid solution (88 wt%) was fed to
a glass reactor initially containing a 30 mL 3m formic acid solu-
tion at 808C. The H2/CO2 production was measured by a volu-
metric flowmeter a in fashion similar to the experimental pro-
cedure reported by Beller et al.[8] At termination of this experi-
ment, a turnover number (TON) of 1060000 was obtained,
which demonstrates the durability of the catalyst.
Traces of methanol (<0.08 mmol) were found to be present
in the liquid phase after full conversion, a product from the
formic acid disproportionation reaction (transfer hydrogenation
of formic acid using formic acid as hydrogen donor), a reaction
recently also reported for [IrIIICp*(H2O)(bpy)] (bpy= 2,2’-bipyri-
dine).[31]
When an experiment with fully deuterated formic acid,
DCOOD, was followed in situ by means of mass spectroscopy,
the molar flow ratios H2/HD/D2 were found to be 195:46:1. The
surprisingly large value for H2 suggests a protonic hydride and
very efficient exchange of the type IrꢀDQIrꢀH, which had
been observed by Ogo et al. before for [IrIIICp*(bpy)H] at low
pH values.[32]
Scheme 2 displays a plausible molecular pathway for the
system, which consists of the three steps that have been
shown to govern this particular type of catalysis: 1) formic acid
deprotonation, 2) b-hydride elimination under decarboxylation
and 3) hydrogen evolution.
&
ChemSusChem 0000, 00, 0 – 0
www.chemsuschem.org
2
ꢀ 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Communications
was proposed for similar systems,[33] which is herein further
supported by elemental analysis indicating that weakly coordi-
nating OTfꢀ is washed out during the recycling of the catalyst,
pointing at formate/formato replacing triflate as charge-bal-
ancing anion within the scaffold.
For (2) to occur, formic acid must be deprotonated, which
only occurs under relatively basic conditions (pKa
=3.77) in
HCOOH
which Ir hydride species are too stable for efficient hydrogen
production and thus impede step (3).[35] It is for this reason
that many catalysts for hydrogen production operate in
tandem with an additional base.[3] The catalyst in this study op-
erates efficiently without any auxiliary base, which we link to
the presence of the free pyridinic sites that provide inherent
basicity within the CTF matrix (see step 1 in Scheme 2).[37]
Indeed, formic acid is well known to have strong hydrogen-
bond interactions with pyridine, which facilitates the abstrac-
tion of protons.[38] It is noteworthy that this CTF is a co-poly-
mer and that the ratio between the building blocks in
Scheme 1 can be tailored to yield an optimum between inher-
ent basicity (by 2,6-pyridinedicarbonitrile) and reactant diffu-
sion (by mesopore-inducing 4,4’-biphenyldicarbonitrile build-
ing blocks). The respective 1:2 ratio between the building
blocks reported in this work provides an ideal match between
basic functionality and mesoporosity.
Finally, it must be emphasised that we expect further im-
provement in TOFs and overall H2 productivity in this system
to come from reactor engineering improvements in which
mass transfer and catalyst dispersion are optimised, as well as
the removal of water from the formic acid feed in continuous
H2 production. For now, this Ir/CTF catalyst sets a new standard
for hydrogen production from formic acid using heterogene-
ous catalysts. The catalyst allows straightforward handling and
recycling under ambient conditions and provides flexibility for
process intensification in a broad temperature range.
Figure 1. Performance of 1 at various metal loadings in initial net production
rate of H2 per unit liquid volume. Standard conditions, unless differently
specified: 3m formic acid solution, 808C, pH 1.5. [RuII(h6-C6H6)]=2 (see text).
In all experiments, reactions were carried out in a 30 mL glass flask with
300 mg of solid catalyst.
Acknowledgements
We thank Dr. Francesc Sastre for GC analysis, Willy Rook for the
adsorption isotherms and Dr. Kristina Djanashvili for the help
with NMR spectroscopy. The work leading to these results has re-
ceived funding from the European Union Seventh Framework
Programme (FP7/2007-2013) under grant agreement number
309701, project Eco2CO2.
Keywords: covalent triazine frameworks
· formic acid ·
heterogeneous catalysis · hydrogen production · iridium
Scheme 2. Simplified scheme displaying the catalytic cycle within the CTF
polymer backbone (sketched grey). Methyl groups of the Cp* ligand and co-
ordination of the labile aqua ligand are not depicted. 1) formic acid activa-
tion, 2) b-hydride elimination, 3) hydrogen release.
[1] C. Fellay, P. J. Dyson, G. Laurenczy, Angew. Chem. Int. Ed. 2008, 47,
3966–3968; Angew. Chem. 2008, 120, 4030–4032.
[2] A. Boddien, F. Gꢁrtner, C. Federsel, P. Sponholz, D. Mellmann, R. Jack-
stell, H. Junge, M. Beller, Angew. Chem. Int. Ed. 2011, 50, 6411–6414;
Angew. Chem. 2011, 123, 6535–6538.
[3] S. Enthaler, J. von Langermann, T. Schmidt, Energy Environ. Sci. 2010, 3,
1207–1217.
[4] P. Kang, T. J. Meyer, M. Brookhart, Chem. Sci. 2013, 4, 3497–3502.
[5] J. F. Hull, Y. Himeda, W. H. Wang, B. Hashiguchi, R. Periana, D. J. Szalda,
J. T. Muckerman, E. Fujita, Nat. Chem. 2012, 4, 383–388.
Of these three steps, (2) has been shown to be rate-limiting
for a very similar system ([RhIIICp*(bpy)(H2O)][SO4]),[35] but also
for an iron phosphine system [P(CH2CH2PPh2)3FeH]+.[36]
Evidence for (2) to proceed starting from coordinated formato
ChemSusChem 0000, 00, 0 – 0
www.chemsuschem.org
3
ꢀ 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Communications
[6] R. E. Rodrꢂguez-Lugo, M. Trincado, M. Vogt, F. Tewes, G. Santiso-Qui-
nones, H. Grꢃtzmacher, Nat. Chem. 2013, 5, 342–347.
[7] A. Boddien, B. Loges, F. Gꢁrtner, C. Torborg, K. Fumino, H. Junge, R.
Ludwig, M. Beller, J. Am. Chem. Soc. 2010, 132, 8924–8934.
[8] A. Boddien, D. Mellmann, F. Gꢁrtner, R. Jackstell, H. Junge, P. J. Dyson, G.
Laurenczy, R. Ludwig, M. Beller, Science 2011, 333, 1733–1736.
[9] T. Zell, B. Butschke, Y. Ben-David, D. Milstein, Chem. Eur. J. 2013, 19,
8068–8072.
[10] The recent reports on base metals as catalysts, iron in particular, should
be encouraged. However, these involve tailored ligands, most contain-
ing increasingly scarce phosphorus, which are more costly than the
metal itself even when iridium is used. The relatively low cost of the
CTF support, and the truly heterogeneous nature of the catalyst in this
work must be emphasized from this perspective.
[11] M. Grasemann, G. Laurenczy, Energy Environ. Sci. 2012, 5, 8171–8181.
[12] CO can poison the catalyst and high CO contents make an extra sepa-
ration step a requirement.
[13] S. Zhang, ꢄ. Metin, D. Su, S. Sun, Angew. Chem. Int. Ed. 2013, 52, 3681–
3684; Angew. Chem. 2013, 125, 3769–3772.
[14] Y. Y. Cai, X. H. Li, Y. N. Zhang, X. Wei, K. X. Wang, J. S. Chen, Angew.
Chem. Int. Ed. 2013, 52, 11822–11825; Angew. Chem. 2013, 125, 12038–
12041.
[15] Z. L. Wang, J. M. Yan, Y. Ping, H. L. Wang, W. T. Zheng, Q. Jiang, Angew.
Chem. Int. Ed. 2013, 52, 4406–4409; Angew. Chem. 2013, 125, 4502–
4505.
[21] W. Gan, P. J. Dyson, G. Laurenczy, ChemCatChem 2013, 5, 3124–3130.
[22] P. Kuhn, M. Antonietti, A. Thomas, Angew. Chem. Int. Ed. 2008, 47,
3450–3453; Angew. Chem. 2008, 120, 3499–3502.
[23] P. Kuhn, A. Forget, D. Su, A. Thomas, M. Antonietti, J. Am. Chem. Soc.
2008, 130, 13333–13337.
[24] P. Kuhn, A. Thomas, M. Antonietti, Macromolecules 2009, 42, 319–326.
[25] R. Palkovits, M. Antonietti, P. Kuhn, A. Thomas, F. Schꢃth, Angew. Chem.
Int. Ed. 2009, 48, 6909–6912; Angew. Chem. 2009, 121, 7042–7045.
[26] See the Supporting Information (available online).
[27] J. R. Pels, F. Kapteijn, J. A. Moulijn, Q. Zhu, K. M. Thomas, Carbon 1995,
33, 1641–1653.
[28] S. Kato, J. Jung, T. Suenobu, S. Fukuzumi, Energy Environ. Sci. 2013, 6,
3756–3764.
[29] F. Angersbach-Bludau, C. Schulz, J. Schçffel, P. Burger, Chem. Commun.
2014, 50, 8735–8738.
[30] M. G. Goesten, P. C. M. M. Magusin, E. A. Pidko, B. Mezari, E. J. M.
Hensen, F. Kapteijn, J. Gascon, Inorg. Chem. 2014, 53, 882–887.
[31] A. J. M. Miller, D. M. Heinekey, J. M. Mayer, K. I. Goldberg, Angew. Chem.
Int. Ed. 2013, 52, 3981–3984; Angew. Chem. 2013, 125, 4073–4076.
[32] T. Abura, S. Ogo, Y. Watanabe, S. Fukuzumi, J. Am. Chem. Soc. 2003, 125,
4149–4154.
[33] S. Ogo, R. Kabe, H. Hayashi, R. Harada, S. Fukuzumi, Dalton Trans. 2006,
4657–4663.
[34] T. Van Gerven, A. Stankiewicz, Ind. Eng. Chem. Res. 2009, 48, 2465–2474.
[35] S. Fukuzumi, T. Kobayashi, T. Suenobu, ChemSusChem 2008, 1, 827–
834.
[16] J. H. Lee, J. Ryu, J. Y. Kim, S.-W. Nam, J. H. Han, T.-H. Lim, S. Gautam,
K. H. Chae, C. W. Yoon, J. Mater. Chem. A 2014, 2, 9490.
[17] W. Zhang, H. Huang, F. Li, K. Deng, X. Wang, J. Mater. Chem. A 2014, 2,
19084–19094.
[18] K. Maeda, K. Sekizawa, O. Ishitani, Chem. Commun. 2013, 49, 10127.
[19] M. Trincado, D. Banerjee, H. Grꢃtzmacher, Energy Environ. Sci. 2014, 7,
2464.
[36] X. Yang, Dalton Trans. 2013, 42, 11987.
[37] J. Roeser, K. Kailasam, A. Thomas, ChemSusChem 2012, 5, 1793–1799.
[38] M. J. Fernandez-Berridi, J. J. Iruin, L. Irusta, J. M. Mercero, J. M. Ugalde, J.
Phys. Chem. A 2002, 106, 4187–4191.
Received: October 24, 2014
[20] W. Gan, P. J. Dyson, G. Laurenczy, React. Kinet. Catal. Lett. 2009, 98, 205–
213.
Published online on && &&, 0000
&
ChemSusChem 0000, 00, 0 – 0
www.chemsuschem.org
4
ꢀ 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

COMMUNICATIONS
H2 and triazine frameworks: The syn-
thesis, characterization, and catalytic
performance of a new heterogeneous
molecular catalyst based on IrIIICp*
(Cp*=pentamethylcyclopentadienyl) at-
tached to a covalent triazine framework
is reported. It catalyses the selective
production of hydrogen from formic
acid with very high initial turnover fre-
quencies and turnover numbers in con-
tinuous operation. The catalyst is air
stable, produces CO-free hydrogen and
is fully recyclable.
A. V. Bavykina, M. G. Goesten,* F. Kapteijn,
M. Makkee, J. Gascon*
&& – &&
Efficient production of hydrogen from
formic acid using a Covalent Triazine
Framework supported molecular
catalyst
ChemSusChem 0000, 00, 0 – 0
www.chemsuschem.org
5
ꢀ 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ