Solid Molecular Phosphine Catalysts for Formic Acid Decomposition in the Biorefinery

Article abstract of DOI:10.1002/anie.201510681

The co-production of formic acid during the conversion of cellulose to levulinic acid offers the possibility for on-site hydrogen production and reductive transformations. Phosphorus-based porous polymers loaded with Ru complexes exhibit high activity and

Full text of DOI:10.1002/anie.201510681

Angewandte
Communications
Chemie
International Edition: DOI: 10.1002/anie.201510681
German Edition: DOI: 10.1002/ange.201510681
Biomass Utilization
Solid Molecular Phosphine Catalysts for Formic Acid Decomposition
in the Biorefinery
Peter J. C. Hausoul,* Cornelia Broicher, Roberta Vegliante, Christian Gçb, and
Regina Palkovits*
Abstract: The co-production of formic acid during the
conversion of cellulose to levulinic acid offers the possibility
for on-site hydrogen production and reductive transforma-
tions. Phosphorus-based porous polymers loaded with Ru
complexes exhibit high activity and selectivity in the base-free
decomposition of formic acid to CO and H . A polymeric
analogue of 1,2-bis(diphenylphosphino)ethane (DPPE) gave
the best results in terms of performance and stability. Recycling
tests revealed low levels of leaching and only a gradual
decrease in the activity over seven runs. An applicability study
revealed that these catalysts even facilitate selective removal of
formic acid from crude product mixtures arising from the
synthesis of levulinic acid.
pounds find application as fuel additives, monomers, solvents,
and fragrance intermediates. Likewise, the esters of LA are
[
8]
also useful as fuel components. The synthesis of LA from
cellulose proceeds via acid-catalyzed hydrolysis with the
intermediacy of glucose and hydroxymethylfurfural (HMF).
In the conversion of HMF to LA, formic acid (FA) is
2
2
[9]
coproduced stoichiometrically. The selective dehydrogen-
ation of FA to H and CO provides an on-site hydrogen
2
2
supply for reductive transformations, possibly in a one-pot
[10]
fashion. Catalytic FA decomposition is well known and can
be achieved using a variety of transition metal complexes and
supported metal catalysts. Particularly Ru/PR and Ir/bipyr-
3
idine complexes are highly active and selective for the
[
11]
synthesis and decomposition of FA.
They are typically
C
urrently, many different routes for the sustainable produc-
tion of fuels and chemicals from renewable resources such as
studied in the context of hydrogen storage and fuel-cell
[
12]
applications. Nevertheless, for the continuous decomposi-
tion of large volume streams of aqueous FA, heterogeneous
catalysts are required. Recent reviews highlight that sup-
biomass, CO , and solar energy are under intensive inves-
tigation. Perhaps, one of the most commercially viable routes
is the conversion of cellulosic waste to levulinic acid (LA)
2
ported metal catalysts typically exhibit lower activities and
[
1]
[13]
(
e.g. the Biofine process) (Scheme 1).
LA is a useful platform chemical that offers several
selectivities than homogeneous catalysts.
Therefore, the
immobilization of these complexes appears the most promis-
ing strategy for catalyst heterogenization. Zhao et al. studied
[
2,3]
possibilities for valorization.
valerolactone (GVL), which can be further converted to
Hydrogenation yields g-
[
4]
Ru and Pd catalysts immobilized on SiO functionalized with
2
[
5]
[6]
[7]
[14]
2
-MeTHF, pentanoic acid, and butenes. These com-
different heteroatom donor ligands (e.g., N, S, and P).
Turnover frequencies (TOFs) up to 960 h were achieved
at 808C for the Pd/S catalyst. Laurenczy et al. tethered Ru/
À1
[15]
TPPTS on phosphine-modified MCM41 and reported TOFs
À1
II
up to 2780 h at 1108C. Stathi et al. immobilized Fe on
À1
phosphine-modified SiO and obtained TOFs up to 7600 h
2
[16]
at 938C.
Bavykina et al. reported a biphenyl/pyridine-
based CTF (covalent triazine framework) loaded with [Cp*Ir-
À1
(
OH)] for FA decomposition. At 808C, TOFs up to 27000 h
[17]
were obtained. These studies clearly demonstrate the high
potential of immobilized catalysts. However, application
under industrially relevant conditions requires materials
that possess high corrosion resistance over a wide temper-
[18]
ature range. In this respect, porous organic polymers are
clearly superior to organic/inorganic hybrid materials. For this
purpose we have prepared phosphine-based cross-linked
polymers as solid ligands for Ru complexes. The loaded
materials were tested in the base-free aqueous decomposition
of FA.
Scheme 1. Production of LA, its subsequent valorization, and on-site
hydrogen production from FA.
[
*] Dr. P. J. C. Hausoul, C. Broicher, R. Vegliante, C. Gçb,
Prof. Dr. R. Palkovits
Polymeric analogs of PPh (pTPP), dppe (pDPPE), and
3
Institut für Technische und Makromolekulare Chemie
RWTH Aachen University
Worringerweg 1, 52074 Aachen (Germany)
E-mail: palkovits@itmc.rwth-aachen.de
1
,2-(diphenylphosphino)benzene (pDPPBe) were prepared
[19]
using a previously developed method (Scheme 2).
The
polymers were characterized by thermogravimetric analysis
TGA), X-ray diffraction (XRD), N physisorption, trans-
(
Supporting information and the ORCID identification number(s) for
the author(s) of this article can be found under http://dx.doi.org/10.
2
mission electron microscopy (TEM), and solid-state NMR
spectroscopy (see the Supporting Information). The physical
1
002/anie.201510681.
Angew. Chem. Int. Ed. 2016, 55, 5597 –5601
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5597

Angewandte
Communications
Chemie
Figure 1. Reactor pressure of the Ru-catalyzed decomposition of aque-
ous formic acid at 1608C (conditions: see Table 1.)
[a]
Table 1: Ru-catalyzed formic acid decomposition.
Catalyst
FA conv. TOF
[CO](g) [Ru](aq) Ru loss
[ppm] [ppm] [%]
À1
[
%]
[h
]
[
b]
blank
Ru/C
2
24
23
–
810
750
2740
15050
22900
13650
n.d.
7880 5.13
3270
<10
67 0.22
13 0.19
142 0.80
–
–
34.2
–
[d]
[c]
[
[
RuCl (p-cymene)]
–
–
2
2
RuCl
2
(p-cymene)(PPh
3
)] 30
99
–
Ru@pTPP
Ru@pDPPE
Ru@pDPPBe
0.7
0.6
1.4
99
99
Scheme 2. Synthesis of phosphine-based cross-linked polymers.
[
(
[
a] Conditions: 50 mL Hastelloy autoclave, 15 mL aqueous solution
10 wt% FA, 2.17m, 32.6 mmol FA), 0.022 mol% Ru, 1608C.
b] 100 min. [c] 2.25 mg [RuCl (p-cymene)] , 80 min. [d] 15 mg Ru/C
2
2
and chemical properties of pDPPE and pDPPBe are quite
(5 wt%, 0.022 mol% Ru), 80 min.
31
similar to those reported previously for pTPP. The P NMR
spectra show that all phosphorus centers are trivalent and the
1
3
C NMR spectra confirm the proposed structures. All
The monophosphine complex [RuCl (p-cymene)(PPh )] was
2 3
polymers are insoluble in common organic solvents, thermally
stable (up to 3408C), and XRD-amorphous, fluffy powders.
The BET surface areas of pDPPE and pDPPBe range
also tested and gave a slightly better conversion of 30% with
a considerably improved H /CO selectivity (CO < 10 ppm).
2
2
In contrast, the Ru-loaded polymers reached full con-
version within 15–30 min in all cases (Figure 1). Following
a short induction period of 4–6 min the pressure increased
almost linearly, suggesting pseudo-zero-order kinetics. TOFs
were estimated by using the pressure plots. The TOFs of
2
À1
between 33 and 42 m g and are somewhat smaller than
2
À1
that of pTPP (150 m g ). TEM analysis shows that the
materials are composed of polydisperse particles which are
composed of coalesced smaller particles. In contrast to the
particles of pTPP and pDPPBe, the particles of pDPPE
are composed of clusters of discrete polydisperse spheres
À1
À1
Ru@pTPP (15050 h ) and Ru@pDPPBe (13650 h ) are
comparable, whereas Ru@pDPPE exhibits a considerably
À1
(
see SI).
Previous studies on pTPP showed that leaching can be
higher TOF of 22900 h . CO concentrations of 67, 13, and
142 ppm were observed for Ru@pTPP, Ru@pDPPE, and
Ru@pDPPBe, respectively. ICP analysis of the filtered
reaction solutions confirmed low levels of Ru (0.2–0.8 ppm,
[19]
minimized by decreasing the metal loading; therefore the
polymers were loaded with 1 wt% [RuCl (p-cymene)]
2
31
(
MeOH, 2 days, RT). ICP analysis of the filtrates confirmed
equaling 0.6–1.4% loss due to leaching). The P NMR
spectrum of the catalyst after the reaction showed no
evidence of oxidation and only 1.5 ppm of phosphorus was
detected in solution which confirms that loss of polymer
through the filter is negligible. The obtained data clearly show
that the polymer-supported catalysts are highly active and
selective. Moreover they outperform the homogeneous
analogue by an order of magnitude.
that approximately 95–99% of the added complex was taken
up by the polymers. The loaded materials (i.e., Ru@pTPP (P/
À1
À1
Ru = 39 molmol ), Ru@pDPPE (P/Ru = 51 molmol ), and
À1
Ru@pDPPBe (P/Ru = 45 molmol )) were tested in the base-
free decomposition of aqueous FA at 1608C (Figure 1,
Table 1). A blank experiment run for 100 min gave only 2%
conversion, confirming that thermal decomposition is negli-
gible. In the presence of Ru/C, conversion reached 24% after
With these promising results in hand, Ru@pDPPE was
chosen to study catalyst recycling in two ways (Figure 2). The
first series of experiments was conducted following a start–
stop procedure where a portion of pure formic acid was added
after each run. The second method entailed a solvent-change
procedure in which the catalyst was filtered and introduced
into a fresh formic acid solution. All reactions were left until
8
0 min and the pressure increased to 15 bar. Analysis of the
gas phase revealed a large impurity of CO (7880 ppm) and
ICP analysis showed that 34% of the introduced metal had
leached into solution. The performance of precatalyst [RuCl2-
(p-cymene)]2 was comparable to that of Ru/C with 23%
conversion after 80 min and the amount of CO was 3270 ppm.
5
598
www.angewandte.org
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2016, 55, 5597 –5601

Angewandte
Communications
Chemie
Table 2: Influence of formic acid concentration on the Ru@pTPP-
[a]
catalyzed FA decomposition.
FA conc.
wt%]
t
FA conv.
[%]
[CO](g)
[ppm]
Ru loss
[%]
[
[min]
3
6
9
0
0
0
126
150
150
99
99
88
249
465
3108
0.7
11.2
31.3
[
a] Conditions: 50 mL Hastelloy autoclave, 3 g (=97.8 mmol) FA, 75 mg
Ru@pTPP (=0.0076 mol% Ru).
Figure 2. Recycling study of the Ru@pDPPE catalyst.
At 30 wt% FA, full conversion was observed after
1
26 min (TON 13160). The CO content of the gas phase
was 249 ppm and 0.7% Ru leached. Compared to the data
above (Table 1), the higher FA concentration led to a slight
increase in CO production, but leaching was not affected. In
contrast, at even higher FA concentrations both CO produc-
tion and leaching increased severely. This suggests that at high
concentrations the carbonic acid displaces the phosphine,
leading to the formation of homogeneous species and possibly
nanoparticles which exhibit significantly lower selectivity.
Nevertheless, the data emphasize that up to 30 wt% FA can
be tolerated without a significant effect on catalyst perfor-
mance.
completion and a final TON of 30760 was achieved after
seven runs for both series.
In the start–stop experiments, the TOF dropped signifi-
cantly after the first run. After that, the rate decreased only
slowly. At the end of the series still 76% of the initial activity
remained, confirming that the supported catalyst is quite
stable. The CO concentration slightly increased to 89 ppm
and in the end 10% of the employed Ru had leached into
solution. TEM images of the spent catalyst did reveal a few
isolated Ru nanoparticles, but overall the catalyst was clean
(
see SI). Thus the data suggest that most of the metal remains
The influence of the presence of LA and H SO was tested
2
4
molecularly dispersed on the polymer. In the solvent-change
series the activity dropped rapidly in the first three runs. After
this, the activity decreased more slowly and the CO concen-
tration increased up to 318 ppm in the last run. Interestingly,
the amount of leached Ru was consistently low (0.4–0.5%)
during the first three runs. Therefore it is proposed that the
steep decrease in activity stems from loss of the catalyst
during the filtration procedure. After the fourth run, leaching
increased (0.6–4.6%) and a cumulative amount of 11% was
lost after the last run.
using Ru@pDDPE (Figure 3). With 0.33 equiv LA (with
respect to FA) in 10 wt% FA solution, FA was rapidly
converted, although the TOF decreased to 14000 h . Leach-
À1
ing and CO formation were not affected. Also with 5 wt% FA
These results clearly demonstrate that the polymer-
supported catalysts exhibit high activity and stability in
clean dilute formic acid solutions. However, the conversion
of cellulose to LA yields a complex mixture containing LA,
FA, and H SO4 as well as tar/chars and soluble humins.
2
Insoluble by-products can be removed, for example, by
filtration. Removal of H SO with CaO yields CaSO , which
Figure 3. Effect of different contaminants on the Ru@pDPPE-catalyzed
decomposition of formic acid.
2
4
4
in turn can be regenerated to CaO and SO . Suitable
3
separation strategies for the other organic components
include extraction or rectification. For example, LA can be
extracted with organic solvents such as MeTHF. Alterna-
tively, fractional distillation can be used to remove water/FA
and 1 equiv LA, the catalyst performed comparably and full
conversion was achieved within 22 min. Interestingly, no GVL
was observed during these reactions. The in situ conversion of
LA was studied in a 24 h reaction and only 6% GVL was
found afterwards indicating that Ru@pDPPE is not suitable
for the hydrogenation of LA under these conditions. Never-
theless, FA could be selectively removed from a feed
containing LA. In the presence of 2 wt% H SO the reaction
(
azeotrope at 22% FA) and LA from the reaction mixture.
Thus depending on the employed process design, different
product streams can arise. Therefore, we chose to test the
capability of the catalyst to work under various reaction
conditions including different FA concentrations and the
presence of sulfuric and levulinic acid, respectively. Finally,
a crude mixture prepared via dehydration of glucose served as
substrate for selective FA decomposition.
The influence of FA concentration was tested using
Ru@pTPP (Table 2). The FA concentrations were set by
varying the amount of water, while keeping the catalyst and
FA amounts constant.
2
4
with 10 wt% FA was considerably slower and CO formation
increased to 5042 ppm. Thus the data show that 2 wt% H SO
2
4
is detrimental for the catalytic activity and should be removed
prior to FA decomposition.
Finally the conversion of FA in a crude LA synthesis
mixture was attempted. For this purpose a glucose solution
containing H SO was subjected to 1908C for 30 min (Fig-
2
4
Angew. Chem. Int. Ed. 2016, 55, 5597 –5601
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5599

Angewandte
Communications
Chemie
Acknowledgements
We acknowledge the Regional Research Cluster Sustainable
Chemical Synthesis (SusChemSys) which is co-financed by
the European Regional Development Fund (ERDF) and the
state North Rhine-Westfalia, Germany, under the Opera-
tional Program Regional Competiveness and Employment
2
007–2013. We kindly thank Karl-Josef Vaeßen for N2-
physisorption, XRD, and TGA experiments, Noah Avraham
for HPLC analyses, Elke Biener for GC analyses, and
Cleopatra Herwartz and Dr. Thomas Weirich for TEM
analysis.
Keywords: formic acid · hydrogen · phosphines ·
Figure 4. Application of Ru@pDPPE for the selective removal of FA
from a crude LA synthesis reaction mixture. a) Reaction scheme;
b) crude reaction mixture arising from LA synthesis after filtration;
porous polymers · ruthenium
How to cite: Angew. Chem. Int. Ed. 2016, 55, 5597–5601
Angew. Chem. 2016, 128, 5687–5691
c) after addition of CaO and removal of CaSO ; d) after FA decom-
4
position and filtration of spent Ru@pDPPE.
[
1] a) “Production of levulinic acid from carbohydrate-containing
materials”: S. W. Fitzpatrick, WO1996040609 A1, 1997; b) A.
Morone, M. Apte, R. A. Pandey, Renewable Sustainable Energy
Rev. 2015, 51, 548 – 565.
ure 4a). The resulting mixture contained a large amount of
char which was removed by filtration (Figure 4b). HPLC
analysis showed that 81% of glucose had been converted to
a complex mixture containing 2.4% LA and 1% FA. After
removal of H SO using CaO (Figure 4c), the mixture was
used in the Ru@pDPPE-catalyzed decomposition of FA. The
reaction was conducted for 24 h at 1608C, during which the
pressure increased to 35 bar. Analysis of the gas phase
revealed a relatively high amount of CO (2385 ppm). An
additional 9% of glucose was converted during the second
reaction and FA was completely removed. The amount of LA
remained unchanged. A dark brown solid was recovered from
the mixture and the filtrate had a light yellow appearance
[
2] M. J. Climent, A. Corma, S. Iborra, Green Chem. 2014, 16, 516 –
5
47.
[3] a) R. Cao, J. Xin, Z. Zhang, Z. Liu, X. Lu, B. Ren, ACS
Sustainable Chem. Eng. 2014, 2, 902 – 909; b) M. G. Al-Shaal,
P. J. C. Hausoul, R. Palkovits, Chem. Commun. 2014, 50, 10206 –
2
4
1
0209.
4] a) W. R. H. Wright, R. Palkovits, ChemSusChem 2012, 5, 1657 –
667; b) W. Luo, M. Sankar, A. M. Beale, Q. He, C. J. Kiely,
P. C. A. Bruijnincx, B. M. Weckhuysen, Nat. Commun. 2015, 6,
540.
[
1
6
[5] a) X.-L. Du, Q.-Y. Bi, Y.-M. Liu, Y. Cao, H.-Y. He, K.-N. Fan,
Green Chem. 2012, 14, 935 – 939; b) M. G. Al-Shaal, A. Dzier-
binski, R. Palkovits, Green Chem. 2014, 16, 1358 – 1364.
[
6] W. Luo, P. C. A. Bruijnincx, B. M. Weckhuysen, J. Catal. 2014,
20, 33 – 41.
(
Figure 4d). This suggests that further polymerization and
3
precipitation of the soluble humins occurs and leads to
a fouling of the catalyst. Nevertheless, these results affirm that
the catalyst still performs in the presence of contaminants.
Hence, it is proposed that the catalyst can be applied in
downstream processing as well as at the end of the pipe.
In conclusion, it could be shown that phosphorus-based
porous polymers are suitable ligands for the aqueous, base-
free Ru-catalyzed decomposition of FA to H and CO . The
[
7] a) S. M. Sen, E. I. Gürbüz, S. G. Wettstein, D. M. Alonso, J. A.
Dumesic, C. T. Maravelias, Green Chem. 2012, 14, 3289 – 3294;
b) D. J. Braden, C. A. Henao, J. Heltzel, C. C. Maravelias, J. A.
Dumesic, Green Chem. 2011, 13, 1755 – 1765.
[8] a) A. DØmolis, N. Essayem, F. Rataboul, ACS Sustainable Chem.
Eng. 2014, 2, 1338 – 1352; b) M. G. Al-Shaal, W. Ciptonugroho,
F. J. Holzhäuser, J. B. Mensah, P. J. C. Hausoul, R. Palkovits,
Catal. Sci. Technol. 2015, 5, 5168 – 5173.
2
2
[
9] J. Zhang, E. Weitz, ACS Catal. 2012, 2, 1211 – 1218.
best results, in terms of TOF and selectivity, were obtained
using a solid molecular analogue of DPPE. In addition it was
shown that the polymeric PPh3 catalyst outperforms its
molecular analogue under the same conditions. Recycling
tests confirm that the activity of the catalysts is stable over
several runs and only minor leaching occurs. Variation of the
FA concentration shows that the catalyst performs efficiently
in solutions with 0–30 wt% FA. Also tests in the presence of
LA and H SO or with a crude product mixture prove that the
catalyst is active and highly selective for FA removal. It is
therefore proposed that Ru-loaded phosphorus-based poly-
mers are promising catalysts for on-site hydrogen production
in future biorefineries.
[
[
[
10] a) H. Mehdi, V. Fµbos, R. Tuba, A. Bodor, L. T. Mika, I. T.
Horvµth, Top. Catal. 2008, 48, 49 – 54; b) H. Heeres, R. Handana,
D. Chunai, C. B. Rasrendra, B. Girisuta, H. J. Heeres, Green
Chem. 2009, 11, 1247 – 1255.
11] a) T. C. Johnson, D. J. Morris, M. Wills, Chem. Soc. Rev. 2010, 39,
8
1 – 88; b) B. Loges, A. Boddien, F. Gärtner, H. Junge, M. Beller,
Top. Catal. 2010, 53, 902 – 914; c) P. G. Jessop, F. Joó, C.-C. Thai,
Coord. Chem. Rev. 2004, 248, 2425 – 2442.
12] a) 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; b) S. Enthaler, J. von Langermann, T. Schmidt, Energy
Environ. Sci. 2010, 3, 1207 – 1217; c) F. Joo, ChemSusChem 2008,
2
4
1, 805 – 808; d) S. Enthaler, ChemSusChem 2008, 1, 801 – 804.
[
[
13] a) M. Grasemann, G. Laurenczy, Energy Environ. Sci. 2012, 5,
8
3
171 – 8181; b) J. Li, Q.-L. Zhu, Q. Xu, Chimia 2015, 69, 348 –
52.
14] Y. Zhao, L. Deng, S.-Y. Tang, D.-M. Lai, B. Liao, Y. Fu, Q.-X.
Guo, Energy Fuels 2011, 25, 3693 – 3697.
5
600 www.angewandte.org
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2016, 55, 5597 –5601

Angewandte
Communications
Chemie
[
[
[
[
15] W. Gan, P. J. Dyson, G. Laurenczy, ChemCatChem 2013, 5,
124 – 3130.
16] P. Stathi, Y. Deligiannakis, G. Avgouropoulos, M. Louloudi,
Appl. Catal. A 2015, 498, 176 – 184.
17] A. V. Bavykina, M. G. Goesten, F. Kapteijn, M. Makkee, J.
Gascon, ChemSusChem 2015, 8, 809 – 812.
18] M. Rose, ChemCatChem 2014, 6, 1166 – 1182.
[19] P. J. C. Hausoul, T. M. Eggenhuisen, D. Nand, M. Baldus, B. M.
3
Weckhuysen, R. J. M. Klein Gebbink, P. C. A. Bruijnincx, Catal.
Sci. Technol. 2013, 3, 2571 – 2579.
Received: November 18, 2015
Revised: February 24, 2016
Published online: April 4, 2016
Angew. Chem. Int. Ed. 2016, 55, 5597 –5601
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5601