Bioinduced Room-Temperature Methanol Reforming

Article abstract of DOI:10.1002/anie.201503737

Imitating nature′s approach in nucleophile-activated formaldehyde dehydrogenation, air-stable ruthenium complexes proved to be exquisite catalysts for the dehydrogenation of formaldehyde hydrate as well as for the transfer hydrogenation to unsaturated organic substrates at loadings as low as 0.5 mol%. Concatenation of the chemical hydrogen-fixation route with an oxidase-mediated activation of methanol gives an artificial methylotrophic in vitro metabolism providing methanol-derived reduction equivalents for synthetic hydrogenation purposes. Moreover, for the first time methanol reforming at room temperature was achieved on the basis of this bioinduced dehydrogenation path delivering hydrogen gas from aqueous methanol.

Full text of DOI:10.1002/anie.201503737

.
Angewandte
Communications
International Edition: DOI: 10.1002/anie.201503737
German Edition: DOI: 10.1002/ange.201503737
Methanol Economy
Hot Paper
Bioinduced Room-Temperature Methanol Reforming**
Leo E. Heim, Daniel Thiel, Christian Gedig, Jan Deska,* and Martin H. G. Prechtl*
Dedicated to Professor Luis A. Oro on the occasion of his 70th birthday
Abstract: Imitating nature’s approach in nucleophile-activated
formaldehyde dehydrogenation, air-stable ruthenium com-
plexes proved to be exquisite catalysts for the dehydrogenation
of formaldehyde hydrate as well as for the transfer hydro-
genation to unsaturated organic substrates at loadings as low as
0.5 mol%. Concatenation of the chemical hydrogen-fixation
route with an oxidase-mediated activation of methanol gives an
artificial methylotrophic in vitro metabolism providing meth-
anol-derived reduction equivalents for synthetic hydrogenation
purposes. Moreover, for the first time methanol reforming at
room temperature was achieved on the basis of this bioinduced
dehydrogenation path delivering hydrogen gas from aqueous
methanol.
improvement. Herein we present a conceptually unprece-
dented strategy for the hydrogen generation from aqueous
methanol based on a multicatalytic system implementing
enzyme catalysis as new player in the field.[14] The interplay of
methanol-activating biocatalysts with H2-liberating metal
complexes is opening up new opportunities en route to
ambient-temperature hydrogen-production systems or
enzyme-driven hydrogen fuel cells.
In contrast to the recent history of ex vivo approaches
utilizing the C1 feedstock, millions of years of evolution have
provided solutions to the problems of the activation and use
of organic one-carbon entities by creating well-defined low-
temperature pathways for methane- or methanol-feeding
organisms.[15,16] Interestingly, as opposed to the chemical
objective aiming for full dehydrogenation processes (e.g.
aqueous methanol yielding the maximum three equivalents of
H2), in aerobic methylotrophic yeasts and bacteria, formal-
dehyde is found to play a central role in the cellsꢀ metabolism.
Once formed by oxidase/catalase-catalyzed formal oxygen-
ation of methanol, formaldehyde hydrate serves both as
source of reduction equivalents, such as NAD(P)H, as well as
the actual carbon feedstock for the generation of carbohy-
drates.[17] Dehydrogenation generally proceeds by nucleo-
philic activation of formaldehyde through enzyme cofactors,
such as pterins (e.g. tetrahydrofolate) or mercaptanes (e.g.
glutathione (GSH)).[18] In particular, the mechanism of GSH-
dependent formaldehyde dehydrogenation is found as highly
conserved motif even in higher eukaryotes playing a key role
in the detoxification systems.[19,20] In this case, GSH adds
reversibly to formaldehyde to generate S-(hydroxymethyl)-
glutathione.[21,22] Tight binding of the hemithioacetal GSH-
I
n light of anthropogenic global warming and the depletion
of ecologically and economically reasonable sources for fossil
fuels, in recent years, one-carbon molecules have emerged as
highly attractive renewable hydrogen carriers.[1–11] Owing to
difficulties in the selective activation of methane at low
temperatures and the rather poor energy balance of formic
acid, methanol is currently discussed as most promising C1
unit.[12] Combining a high hydrogen content (12.5 wt%) with
the convenience of a liquid fuel, catalytic methods for the
dehydrogenation of methanol as well as its production from
CO2 and H2 have attracted major interest from the scientific
community over the past decade to establish a carbon-neutral
methanol economy.[13] However, despite a number of suc-
cessful examples on the methanol-to-H2 conversion, in
particular this process is still in great need for further
[*] M.Sc. L. E. Heim, Dipl.-Chem. D. Thiel, C. Gedig,
Priv.-Doz. Dr. J. Deska,[+] Priv.-Doz. Dr. M. H. G. Prechtl
Department Chemie, Universität zu Kçln
Greinstrasse 4-6
50939 Cologne (Germany)
E-mail: martin.prechtl@uni-koeln.de
Homepage: http://www.catalysislab.de
[+] Current address: Department of Chemistry, Aalto University
Kemistintie 1, 00076 Espoo (Finland)
E-mail: jan.deska@aalto.fi
[**] We gratefully acknowledge financial support provided by the
Ministerium für Innovation, Wissenschaft und Forschung (NRW-
returnee award 2009 to M.H.G.P.), the Fonds der Chemischen
Industrie (Liebig fellowship to J.D.), the Deutsche Forschungsge-
meinschaft (DE1599/4-1), the Dr.-Otto-Rçhm-Gedächtnisstiftung
and the Ernst-Haage-Stiftung. A patent has been filed under
102013011379.2 at the German Patent Office DMPA and extended
to PCT in 2014. A special thank goes to Svetlana Botov and Hagga
Schmalz for the access to the GC-TCD analysis.
Scheme 1. Nucleophile-activated formaldehyde dehydrogenation:
a) biotic methylotrophic NADH formation and b) artificial ruthenium-
catalyzed H2-generation from tetrahedral formaldehyde conjugates.[25]
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201503737.
10308
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2015, 54, 10308 –10312

Angewandte
Chemie
conjugate to the zinc-dependent dehydrogenases results in
Meerwein–Ponndorf–Verley-type transfer hydrogenation
forming NADH and S-formylglutathione (Scheme 1a), with
the S-formylglutathione being subsequently hydrolyzed lib-
erating formic acid to serve as the second reduction equiv-
alent.[23,24] While biogenic formaldehyde transfer hydrogena-
tion requires complex cofactors to facilitate the reduction of
NAD(P)+, much more simplistic settings are easily conceiv-
able in an abiotic environment allowing for nucleophiles as
simple as water to be involved in the formaldehyde activation.
In aqueous solution, formaldehyde is fully hydrated as
methanediol with an equilibrium constant of K = 2 ” 103.[25]
With the development of air- and water-stable transition-
metal complexes acting on formaldehyde hydrate, this highly
energetic, yet widely unheeded, small molecule ought to
become an important player within the family of low-
molecular-weight organic hydrogen equivalents. Most
recently, we demonstrated that methanediol (formalin) can
indeed be considered a high-potential H2 source as dehydro-
genation in presence of [{Ru(p-cymene)Cl2}2] (1) as the
precatalyst provides two equivalents of pure hydrogen gas
along with one equivalent of CO2 (Scheme 1b).[25] In this
bioinspired approach, apparently formaldehyde enables
water splitting via methanediol formation delivering H2
which originated both from H2O and HCHO. A major
advantage of formalin in comparison to formic acid or
isopropanol is particularly its superior hydrogen content
accounting for 8.4 wt% H2 as opposed to 4.4 wt% (HCO2H)
and 3.3 wt% (iPrOH), respectively, a result of the feasibility
to deliver two equivalents of hydrogen from a one-carbon
entity.
Figure 1. Gas-phase analysis: a) decoupled kinetics of hydrogen forma-
In analogy to the natural model in which formaldehyde-
derived NAD(P)H serves as general reducing agent within
the metabolism, we envisioned that in the presence of suitable
organic acceptor molecules direct or indirect transfer of
hydrogen from formalin should occur. In our previous
mechanistic investigations about the composition of the
liquid and gaseous phase in ruthenium-mediated formalin
reforming, we found that [{Ru(p-cymene)Cl2}2] (1) is readily
converted into [(Ru(p-cymene))2m-H(m-HCO2)m-Cl]Cl (2-Cl)
in presence of formalin or formic acid. Both complexes are
active for the dehydrogenative decomposition of methanediol
into H2 and CO2, and most likely complex 2-Cl plays a key
role in the catalytic cycle for dehydrogenation and hydro-
genation processes.[25] In our studies about the decomposition
of methanediol using the ruthenium dimers 1 and 2-X (X = Cl
or BF4), we found that this approach is also rather generally
applicable for transfer hydrogenation reactions. In an adap-
tation of our hydrogen generation approach using methane-
diol, formalin-derived hydrogen can be smoothly transferred
to, for example, acetophenone by the same catalyst, and most
notably, 1-phenylethanol is formed even at room temperature
in concentrated as well as in diluted aqueous formalin
(Supporting Information, Table S1). To elucidate the under-
lying processes, kinetic studies on the dehydrogenation and
hydrogenation steps were conducted in a decoupled fashion in
separate reaction vessels (Figure 1a). Under the assumption
that methanediol decomposition proceeds initially at normal
pressure, H2-formation was measured in the absence of the
tion and consumption using precatalyst 1 and b) pressure profiles of
transfer hydrogenations of acetophenone using 1 or 2.
ketone acceptor in a flow-setup providing rates of dehydro-
genation with an initial turnover frequency of 31.5 minÀ1.
Additionally, rates of hydrogen consumption were obtained
by the hydrogenation of acetophenone in the absence of
formalin under 12 bar H2 atmosphere resembling the partial
hydrogen pressure observed in coupled transfer hydrogena-
tions. The lower turnover frequency of 2.6 minÀ1 of this
process is also reflected in the pressure profile of the transfer
hydrogenations (Figure 1b). Both with the neutral complex
1 as well as the cationic species 2-BF4, a sharp pressure
increase was observed cause by the fast decomposition of
formalin to H2 and CO2 towards a maximum pressure of
20 bar. From here, with the lower rate of acetophenone
reduction, the pressure decreases in a shallow slope reaching
a plateau at 17 bar after 18 h. These results for the transfer
hydrogenation indicate that dehydrogenation and hydroge-
nation are operating as two independent processes rather
than a Noyori-type direct hydrogen-transfer reaction.[26]
Strikingly, the constant increase in pressure even at room
temperature caused by formation of H2 and CO2 underlines
the high potential of this simple catalytic arrangement
particularly in dehydrogenative processes making it a very
promising system in the context of C1-to-H2-reforming.
Angew. Chem. Int. Ed. 2015, 54, 10308 –10312
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 10309

.
Angewandte
Communications
Although homogeneously catalyzed direct methanol
reforming is no longer beyond reach, current processes are
still hampered by operational handicaps, such as excessive
base-loading or high-temperature requirements.[6–9] On the
other hand, nature does not only teach us how to circumvent
such obstacles and perform C1-dehydrogenation under phys-
iological conditions. With the enormous progress in the field
of biotechnological protein production, natural tools are
becoming broadly available and enzyme-catalyzed transfor-
mations gradually enter the chemist’s toolbox.[27,28] We
envisioned that the lack of reactivity of 1 and 2 towards
methanol at ambient temperature could be compensated in
a chemoenzymatic approach combining biocatalytic metha-
nol activation with the use of the thus formed methanediol in
subsequent metal-catalyzed dehydrogenative processes. In
this case, alcohol oxidases from methylotrophic organisms
would provide the activity for the room-temperature dehy-
drogenation. In combination with a catalase, methanol is
directly converted into formaldehyde hydrate under net
consumption of one atom of aerial oxygen. After successful
methanediol accumulation, four hydrogen atoms would be
available a) for transfer hydrogenation to organic substrates
by precatalyst 1 resulting in an artificial methanol metabolism
or b) for acceptorless dehydrogenation resembling formal
methanol reforming on the basis of an enzyme–metal-coupled
catalytic cascade (Scheme 2).[29]
Figure 2. Activity screening of commercial oxidases under substrate
stress conditions: Initial rates of methanol conversion at high concen-
tration (2.0m MeOH) were recorded photospectrometrically using
horseradish peroxidase/ABTS as a reporter system.
a final methanediol concentration of 0.53m (Æ 0.04m) after
24 h while formalin production using P. pastoris oxidase
already terminated at 0.08m concentration.
To our delight, the biocatalytically produced formalin
could also be used as reducing agent in the Ru-catalyzed
transfer hydrogenation of acetophenone. However, direct use
of the untreated protein solution resulted in low conversion
(18% after 48 h), owing to catalyst deactivation in presence of
protein at elevated temperature (958C). While the develop-
ment of more robust (de)hydrogenation catalysts will be an
important requirement on the longer run, simple protein
removal by means of membrane-based techniques would
allow for the generation of more catalyst-friendly biogenic
methanediol. Incubation of aqueous methanol employing the
previously identified biocatalytic cocktail in a membrane
reactor followed by ultrafiltration (MWCO = 10 kDa) gave
rise to an easily processed formalin-enriched solution. Sub-
sequent hydrogen transfer using ruthenium dimer 1 (5 mol%)
at 958C was not only successful for acetophenone (72%
conv.) but even also in the reduction of cyclooctene where full
conversion was observed (Supporting Information,
Scheme S1).
Currently, the prime limitations of a bioinduced transfer
hydrogenation lies in the relatively poor hydrogenation
abilities of the dehydrogenation precatalyst 1, which result
in the need for high catalyst loading and elevated temper-
atures. We speculated that the acceptorless dehydrogenation
of the enzymatically derived methanediol should be more
feasible even with a lower energy input, thus providing
a room-temperature pathway for the catalytic methanol
reforming. In a first experiment, untreated aqueous methanol
was added to complex 1 (5 mol%) and the solution was
stirred at 258C in a sealed autoclave for 24 h, but no change in
gas composition or internal pressure was detected (Fig-
ure 3a). In contrast, using the same solution after preincuba-
tion with the oxidase/catalase-system, gas formation com-
menced instantaneously as indicated by a steady increase in
pressure. Headspace GC-TCD analysis of the gas phase
confirmed the generation of hydrogen in substantial amounts
(Figure 3b). To our delight, not only the decomposition of the
membrane-filtered formalin generated a hydrogen-enriched
Scheme 2. Chemoenzymatic interpretation of the methanol metabo-
lism found in methanol-feeding microorganisms.
To identify potential biocatalysts for the activation of
methanol, four commercial hydroxy oxidases were evaluated
with regard to their activity in the methanediol formation.
Aiming for a practical low-dilution overall process, a high
methanol concentration (2.0m) was chosen as the critical
parameter to validate the biocatalystsꢀ performance
(Figure 2). The two oxidases tested, which act on sugars as
their native substrate, failed, galactose oxidase (Dactylium
dentroides) exhibited only marginal activity with vmax
=
2.3 nmolminÀ1 mgÀ1 while glucose oxidase (Aspergillus
niger) did not provide any measurable turnover. Short-chain
alcohol oxidases on the other hand converted methanol at
rates that were higher by several orders of magnitude.[30] The
good activity of oxidase from Candida boidinii (vmax
=
1.7 mmolminÀ1 mgÀ1 was impressively surpassed by Pichia
)
pastoris oxidase with an initial rate of 29.1 mmolminÀ1 mgÀ1.
However, on a preparative scale, in combination with catalase
from Corynebacterium glutamicum inducing disproportiona-
tion of the liberated hydrogen peroxide, oxidase from
C. boidinii proved to be more robust and reliable reaching
10310 www.angewandte.org
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2015, 54, 10308 –10312

Angewandte
Chemie
fashion. In a closed aerial atmosphere, time-resolved mon-
itoring of the gaseous constituents showed fast consumption
of oxygen as expected for the oxidase-catalyzed methanol
oxygenation. At lower rates of dehydrogenation, metal-
induced hydrogen liberation is recorded proving a catalyst
lifetime of more than 24 h (Figure 3c), although inhibitory
effects still prevent quantitative conversions (see also Sup-
porting Information, Figure S16). Regarding compatibility
issues, modern immobilization and protein-hybrid techniques
will be exploited in future studies to further improve the
efficiency of this enzyme–metal-coupled system.[31,32]
Bringing together enzymatic and metal catalysis, we were
able to construct the first effective system to perform
methanol reforming at room temperature in aqueous media.
Identification of suitable biocatalysts for the activation of
methanol in combination with a novel bioinspired ruthenium-
catalyzed dehydrogenation method employing the in situ
formed methanediol as the active hydrogen species proved to
be the key to overcome the high-temperature constraint of
purely metal-based approaches. As a proof-of-principle study,
this conceptual work will serve as a foundation for the
development of more efficient chemoenzymatic H2-generat-
ing systems exploiting other modes to couple bio- and metal
catalysts to cope with the current formaldehyde-related
concentration limitations.
Keywords: biocatalysis · bioinspired reactions ·
chemoenzymatic reactions · hydrogen · ruthenium
How to cite: Angew. Chem. Int. Ed. 2015, 54, 10308–10312
Angew. Chem. 2015, 127, 10447–10451
[1] C. Fellay, P. J. Dyson, G. Laurenczy, Angew. Chem. Int. Ed. 2008,
47, 3966 – 3968; Angew. Chem. 2008, 120, 4030 – 4032.
[2] B. Loges, A. Boddien, H. Junge, M. Beller, Angew. Chem. Int.
Ed. 2008, 47, 3962 – 3965; Angew. Chem. 2008, 120, 4026 – 4029.
[3] R. Tanaka, M. Yamashita, K. Nozaki, J. Am. Chem. Soc. 2009,
131, 14168 – 14169.
[4] A. Boddien, B. Loges, F. Gartner, C. Torborg, K. Fumino, H.
Junge, R. Ludwig, M. Beller, J. Am. Chem. Soc. 2010, 132, 8924 –
8934.
[5] J. D. Scholten, M. H. G. Prechtl, J. Dupont, ChemCatChem 2010,
2, 1265 – 1270.
[6] M. Nielsen, E. Alberico, W. Baumann, H. J. Drexler, H. Junge, S.
Gladiali, M. Beller, Nature 2013, 495, 85 – 89.
[7] R. E. Rodríguez-Lugo, M. Trincado, M. Vogt, F. Tewes, G.
Santiso-Quinones, H. Grützmacher, Nat. Chem. 2013, 5, 342 –
347.
[8] E. Alberico, P. Sponholz, C. Cordes, M. Nielsen, H. J. Drexler,
W. Baumann, H. Junge, M. Beller, Angew. Chem. Int. Ed. 2013,
52, 14162 – 14166; Angew. Chem. 2013, 125, 14412 – 14416.
[9] P. Hu, Y. Diskin-Posner, Y. Ben-David, D. Milstein, ACS Catal.
2014, 4, 2649 – 2652.
Figure 3. Bioinduced methanol dehydrogenation: a) pressure profile of
the enzyme-driven metal-catalyzed methanol decomposition. b) Chro-
matographic traces of headspace GC-TCD for the quantification of the
generated hydrogen gas from preformed formalin (filtered and unfil-
tered) and one-pot in situ formed formaldehyde solution. c) Time-
resolved monitoring of the hydrogen and oxygen content in the gas
phase in the one-pot chemoenzymatic methanol reforming.
gas phase (15.3 mmolmLÀ1 H2), the experiments further
revealed that also treatment of enzyme-containing aqueous
methanediol with precatalyst 1 led to effective dehydrogen-
ation (8.7 mmolmLÀ1 H2) (Figure 3c). Most remarkably, even
co-incubation of oxidase, catalase, and [{Ru(p-cymene)Cl2}2]
in aqueous methanol at room temperature resulted in
detectable amounts of hydrogen in the headspace
(3.2 mmolmLÀ1 H2), unequivocally demonstrating that che-
moenzymatic methanol reforming is feasible in a one-pot
[10] S. Wesselbaum, T. vom Stein, J. Klankermayer, W. Leitner,
Angew. Chem. Int. Ed. 2012, 51, 7499 – 7502; Angew. Chem.
2012, 124, 7617 – 7620.
[11] S. Savourey, G. Lefevre, J. C. Berthet, P. Thuery, C. Genre, T.
Cantat, Angew. Chem. Int. Ed. 2014, 53, 10466 – 10470; Angew.
Chem. 2014, 126, 10634 – 10638.
[12] G. A. Olah, Angew. Chem. Int. Ed. 2013, 52, 104 – 107; Angew.
Chem. 2013, 125, 112 – 116.
[13] G. A. Olah, Angew. Chem. Int. Ed. 2005, 44, 2636 – 2639; Angew.
Chem. 2005, 117, 2692 – 2696.
Angew. Chem. Int. Ed. 2015, 54, 10308 –10312
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 10311

.
Angewandte
Communications
[14] For a purely biocatalytic methanol dehydrogenation see: S.
Kara, J. H. Schrittwieser, S. Gargiulo, Y. Ni, H. Yanase, D. J.
Opperman, W. J. H. van Berkel, F. Hollmann, Adv. Synth. Catal.
2015, 357, 1687 – 1691.
[15] C. Anthony, The Biochemistry of Methylotrophs, 1st ed., Aca-
demic Press, London, 1982.
[16] L. Chistoserdova, M. G. Kalyuzhnaya, M. E. Lidstrom, Annu.
Rev. Microbiol. 2009, 63, 477 – 499.
[17] J. L. Cereghino, J. M. Cregg, FEMS Microbiol. Rev. 2000, 24, 45 –
66.
[18] J. A. Vorholt, Arch. Microbiol. 2002, 178, 239 – 249.
[19] L. M. Liu, A. Hausladen, M. Zeng, L. Que, J. Heitman, J. S.
Stamler, Nature 2001, 410, 490 – 494.
[20] D. P. Dixon, I. Cummins, D. J. Cole, R. Edwards, Curr. Opin.
Plant Biol. 1998, 1, 258 – 266.
[26] J. S. M. Samec, J. E. Bäckvall, P. G. Andersson, P. Brandt, Chem.
Soc. Rev. 2006, 35, 237 – 248.
[27] U. T. Bornscheuer, G. W. Huisman, R. J. Kazlauskas, S. Lutz,
J. C. Moore, K. Robins, Nature 2012, 485, 185 – 194.
[28] P. Goswami, S. S. R. Chinnadayyala, M. Chakraborty, A. K.
Kumar, A. Kakoti, Appl. Microbiol. Biotechnol. 2013, 97, 4259 –
4275.
[29] While methanol reforming in the narrow sense comprises only
the transformation of CH3OH and H2O to 3H2 and CO2,
practical variants, such as autothermal methanol reforming,
extend to other processes including the aerobic dehydrogenation
as described in this work. See also: K. Geissler, E. Newson, F.
Vogel, T.-B. Truong, P. Hottinger, A. Wokaun, Phys. Chem.
Chem. Phys. 2001, 3, 289 – 293.
´
[30] D. Thiel, D. Doknic, J. Deska, Nat. Commun. 2014, 5, 5278.
[21] R. P. Mason, J. K. M. Sanders, A. Crawford, B. K. Hunter,
Biochemistry 1986, 25, 4504 – 4507.
[22] P. C. Sanghani, C. L. Stone, B. D. Ray, E. V. Pindel, T. D. Hurley,
W. F. Bosron, Biochemistry 2000, 39, 10720 – 10729.
[23] Z. N. Yang, W. F. Bosron, T. D. Hurley, J. Mol. Biol. 1997, 265,
330 – 343.
[31] V. Kçhler, Y. M. Wilson, M. Dürrenberger, D. Ghislieri, E.
Churakova, L. Knçrr, D. Häussinger, F. Hollmann, N. J. Turner,
T. R. Ward, Nat. Chem. 2013, 5, 93 – 99.
[32] M. Basauri-Molina, C. F. Riemersma, M. A. Würdemann, H.
Kleijn, R. J. M. Klein Gebbink, Chem. Commun. 2015, 51, 6792 –
6795.
[24] P. C. Sanghani, H. Robinson, W. F. Bosron, T. D. Hurley,
Biochemistry 2002, 41, 10778 – 10786.
[25] L. E. Heim, N. E. Schlçrer, J. H. Choi, M. H. G. Prechtl, Nat.
Commun. 2014, 5, 3621.
Received: April 23, 2015
Published online: July 14, 2015
10312 www.angewandte.org
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2015, 54, 10308 –10312