Palladium-Catalyzed Oxidative Synthesis of α-Acetoxylated Enones from Alkynes

Article abstract of DOI:10.1002/anie.201600696

We report a palladium-catalyzed oxidative functionalization of alkynes to generate α-acetoxylated enones in one step. A range of functional groups are well-tolerated in this reaction. Mechanistic studies, including the use of ¹⁸O-labeled DMSO,

Full text of DOI:10.1002/anie.201600696

Angewandte
Communications
Chemie
International Edition: DOI: 10.1002/anie.201600696
German Edition: DOI: 10.1002/ange.201600696
Synthetic Methods
Palladium-Catalyzed Oxidative Synthesis of a-Acetoxylated Enones
from Alkynes
Tuo Jiang, Xu Quan, Can Zhu, Pher G. Andersson, and Jan-E. Bäckvall*
Abstract: We report a palladium-catalyzed oxidative function-
alization of alkynes to generate a-acetoxylated enones in one
step. A range of functional groups are well-tolerated in this
reaction. Mechanistic studies, including the use of 18O-labeled
DMSO, revealed that the ketone oxygen atom in the product
originates from DMSO.
À
R
eactions involving C H functionalization have been
widely recognized as highly demanding methodologies for
the direct introduction of new functional groups into mole-
cules. For several decades, these reactions have been com-
prehensively studied in the realm of late-transition-metal
catalysis, most recently with the assistance of some directing
groups.[1,2] In the absence of directing groups, the reaction
outcome is mainly determined by its intrinsic reactivity and/or
3
À
selectivity. In the palladium-catalyzed allylic C(sp ) H acet-
oxylation reaction, early studies on cyclic alkenes by the
Škermark[3] and our group[4] showed that in acetic acid, the
reaction proceeds via a p-allylpalladium(II) intermediate.
White and co-workers later found that in the reaction of
acyclic terminal olefins, the addition of sulfoxide ligands
À
significantly promoted the allylic C H oxidation and also
allowed the amount of acetic acid to be lowered to
4 equivalents (Scheme 1a).[5]
Following our long-term interest in palladium-catalyzed
oxidation chemistry, we have actively engaged in the develop-
ment of various oxidative carbocyclization reactions.[6,7] In
our recent studies of the cyclization of allenynes, we observed
À
an unconventional propargylic C H activation pathway, and
the vinylallene products could be formed selectively (Sche-
À
Scheme 1. Palladium-catalyzed allylic and propargylic C H functionali-
zation. DMSO=dimethyl sulfoxide.
me 1b).[8] An interesting question is whether a similar prop-
À
argylic C H functionalization can also take place with simple
alkynes, via either an allenyl- or a propargylpalladium
intermediate, which could then undergo a suitable function-
alization step, for example, acetoxylation (Scheme 1c).[9,10] In
our evaluation of various acetate sources in the presence of
sulfoxides and 1,4-benzoquinone (BQ), no desired allenyl/
propargyl acetates were observed. However, we found that
the treatment of alkynes with catalytic amounts of Pd(OAc)2
and (diacetoxyiodo)benzene (PIDA) in DMSO afforded (Z)-
a-acetoxylated enones with high selectivity (Scheme 1c).[11]
To the best of our knowledge, a-acetoxylated enones were
previously only accessible from prefunctionalized starting
materials, for example, through the isomerization of prop-
argylic acetates[12] or acylation of 1,2-diketone compounds.[13]
The use of simple alkynes to directly generate such function-
ality-rich structures is unprecedented.[14] Herein, we report
a palladium-catalyzed one-step oxidative protocol for the
conversion of alkynes into a-acetoxylated enones.
We initiated the investigation by using 1-phenyl-1-butyne
(1a) as the model substrate. When alkyne 1a was treated with
Pd(OAc)2 (5 mol%) and PIDA (1.5 equiv) in [D6]DMSO at
508C for 18 h, the a-acetoxylated enone 2a was formed in
46% yield, and the vicinal diketone 3a was also observed as
a side product (7% yield; Table 1, entry 1).[15] No reaction
[*] Dr. T. Jiang, Dr. X. Quan, Dr. C. Zhu, Prof. Dr. P. G. Andersson,
Prof. Dr. J.-E. Bäckvall
Department of Organic Chemistry, Arrhenius Laboratory
Stockholm University, 10691 Stockholm (Sweden)
E-mail: jeb@organ.su.se
Supporting information for this article can be found under:
http://dx.doi.org/10.1002/anie.201600696.
ꢀ 2016 The Authors. Published by Wiley-VCH Verlag GmbH & Co.
KGaA. This is an open access article under the terms of the Creative
Commons Attribution Non-Commercial NoDerivs License, which
permits use and distribution in any medium, provided the original
work is properly cited, the use is non-commercial, and no
modifications or adaptations are made.
5824
ꢀ 2016 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2016, 55, 5824 –5828

Angewandte
Communications
Chemie
Table 1: Optimization of the reaction conditions.[a]
Entry
PIDA
(x equiv)
“BQ”
Yield of
Yield of
2a [%][b]
3a [%][b]
[c]
1
2
3
4
5
6
7
8
9
1.5
2.5
3
3
3
3
3
3
3
–
46
62
64
69
66
64
65
65
62
64
7
9
[c]
–
[c]
–
11
11
13
12
12
15
16
14
BQ
2,6-Me2-BQ
2,6-tBu2-BQ
2,6-(MeO)2-BQ
2-Cl-BQ
F4-BQ
maleic anhydride
BQ
10
3
3
11[d]
67 (65)[e]
13[e]
[a] Reaction conditions: 1a (0.1 mmol), Pd(OAc)2 (5 mol%), PIDA
(x equiv), “BQ” (10 mol%), [D6]DMSO (0.5 mL), 508C, 18 h. [b] The yield
was determined by 1H NMR spectroscopy of the crude mixture with
anisole as the internal standard. [c] No quinone or quinone analogue was
added. [d] The reaction was carried out on a 0.3 mmol scale in DMSO.
[e] Yield of the crude product after workup. The yield of isolated 2a is
given in parenthesis.
occurred in the absence of either Pd(OAc)2 or PIDA.[16]
Acetonitrile, chloroform, and toluene could be used as
cosolvents, but no increase in yield was observed under
these conditions.[17] Increased amounts of PIDA improved the
reaction outcome, and with 3 equivalents of PIDA, 2a was
generated in 64% yield (Table 1, entry 3). The addition of
ligands, such as 2,2’-bipyridine and 1,10-phenanthroline
(5 mol%), completely shut down the reaction. Among all
the tested additives,[18] we found that BQ enhanced the
reaction: With 10 mol% of BQ, the yield of 2a was elevated
to 69% (Table 1, entry 4). The evaluation of quinone
analogues, including maleic anhydride, showed that non-
substituted BQ was optimal (Table 1, entries 5–10). When the
reaction was carried out on a 0.3 mmol scale in DMSO, the a-
acetoxylated enone 2a was obtained in 65% yield (Table 1,
entry 11).
Scheme 2. Acetoxylation of substrates with different propargylic sub-
stituents. Reactions were carried out on a 0.3 mmol scale. Yields are
for the isolated product. [a] Reaction time: 24 h. [b] For the detailed X-
ray crystal structure of 2j, see the Supporting Information. Bn=benzyl,
NPhth=phthalimido, TBS=tert-butyldimethylsilyl.
Having optimized the reaction conditions, we continued
to explore the scope of the reaction, first by examining
different substituents at the propargylic position of the
substrate (Scheme 2). Variation of the length of the alkyl
chain at the propargylic position had a minor influence on the
yield of the corresponding a-acetoxylated enone 2b–e.
Alkyne 1 f containing a benzyl substituent underwent the
reaction smoothly to deliver 2 f in 68% yield. In sharp
contrast, substrate 1g bearing an allyl group on the alkyne
was converted into the desired product 2g’ in only 14% yield,
which implies poor compatibility of the reaction with olefins.
We were delighted to find that the propargyl ether 1h was
well-tolerated, as well as a number of other functional groups,
including alkyl halides (substrate 1i) and protected hetero-
atoms, such as phthalimides (substrate 1j) and silyl ethers
(substrate 1k). Notably, the products 2a and 2c–k were
detected as only the Z isomers.[19] As for propargylic disub-
stituted alkynes, the isopropyl alkyne 1l gave the tetrasub-
stituted olefin 2l in a diminished yield of 46%. More
importantly, we noticed that substrates incorporating a cyclo-
alkyl group differed significantly in their reactivity. The
cyclopentyl- and cyclobutyl-substituted alkynes 1m and 1n
were converted into the corresponding cycloalkylidene prod-
ucts 2m and 2n in good yields. However, in the case of the
cyclopropyl alkyne 1o, the expected cyclopropylidene prod-
uct was not formed. The reaction was dominated by a ring-
opening pathway and provided the a-acetoxylated dienone
2o’ in 42% yield (Z/E = 10:1).
Following these encouraging results for different prop-
argylic substituted alkynes, we investigated the scope of the
reaction with respect to the arene unit (Scheme 3). Both
Angew. Chem. Int. Ed. 2016, 55, 5824 –5828
ꢀ 2016 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5825

Angewandte
Communications
Chemie
Scheme 4. Reactions of 1a with PhI(OPiv)2 or [18O]DMSO.
Scheme 3. Acetoxylation of substrates with different aryl substituents.
Reactions were carried out on a 0.3 mmol scale. Yields are for the
isolated product. [a] Reaction time: 40 h.
electron-donating and electron-withdrawing groups, such as
alkyl groups, alkoxyl groups, and halides, were compatible
with the reaction conditions. Moreover, the ketone- and ester-
containing substrates 1t and 1u reacted smoothly, and these
carbonyl groups are valuable handles for subsequent trans-
formations. Alkyne 1v with an olefin moiety underwent the
reaction poorly to give the desired a-acetoxylated enone 2v
only in 25% yield. For synthetically attractive heteroarenes,
we were delighted to find that an electron-rich thiophene
group (substrate 1w) was well-tolerated, and no side products
Scheme 5. Isotope-labeling experiments.
irreversible step of the catalytic cycle. When alkyne 1b and
[D3]1b reacted separately, a KIE value (kH/kD from the initial
rates) of 1.04 was observed (Scheme 5b).[21] Such a negligible
À
arising from potential arene C H oxidation were detected.
Similarly, indole moieties could also be incorporated; the a-
acetoxylated enone 2x could be utilized as a precursor in the
synthesis of indole-type natural products.
À
isotope effect clearly indicates that propargylic C H bond
cleavage is not involved in the rate-determining step.[22]
On the basis of the experimental results, we propose
a plausible mechanism for this oxidative alkyne functional-
ization reaction (Scheme 6). Initially, a Wacker-type nucleo-
philic attack[23] by DMSO on the palladium(II)-activated
alkyne provides a vinylpalladium(II) intermediate int-1.[24–26]
This palladium(II) species is oxidized by PIDA to
We carried out additional experiments to gain a deeper
understanding of the reaction. When PhI(OAc)2 was replaced
with PhI(OPiv)2, a pivalate group could be installed at the
a position of enone 4 in good yield, thus revealing that the
carboxylate functionality can be varied by the use of different
iodine(III) reagents (Scheme 4a). When the reaction of
alkyne 1a was carried out with [18O]DMSO (86% 18O), we
observed the incorporation of 18O into the product [18O]2a
(62% 18O), as determined by ESI-MS analysis (Scheme 4b;
see also the Supporting Information).[20] This result shows that
the ketone oxygen atom originates from the sulfoxide.
a
vinylpalladium(IV) intermediate int-2,[27] which can
undergo b-hydride elimination (path A) or reductive elimi-
nation (path B)[28] to generate int-3 or int-4, respectively.
Nucleophilic attack by acetate on int-3 (path A) would give
the desired product 2; alternatively, the cleavage of an allylic
À
À
As the propargylic C H bond is broken in the reaction,
C H bond, followed by double-bond migration (path B),
we performed some kinetic isotope effect (KIE) studies with
the deuterium-labeled substrate [D3]1b. In an intermolecular
competition experiment, an equimolar amount of 1b and
[D3]1b gave 2b and [D2]2b in a nearly 1:1 ratio (kH/kD = 1.05;
Scheme 5a). This observation shows that the cleavage of
would also deliver product 2. Dimethyl sulfide is expelled as
the by-product and was detected by GC–MS analysis of the
crude mixture.[29,30] Although it is known that BQ can
promote different catalytic steps in palladium-catalyzed
oxidation reactions,[31] the exact role of BQ in this reaction
remains unclear.
À
À
propargylic C H and C D bonds occurs after the first
5826
www.angewandte.org
ꢀ 2016 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2016, 55, 5824 –5828

Angewandte
Communications
Chemie
Keywords: alkynes · enones · oxidation · palladium · sulfoxides
How to cite: Angew. Chem. Int. Ed. 2016, 55, 5824–5828
Angew. Chem. 2016, 128, 5918–5922
À
[1] For recent reviews of late-transition-metal-catalyzed C H
functionalization reactions, see: a) L. Yang, H. Huang, Chem.
Rev. 2015, 115, 3468; b) N. Dastbaravardeh, M. Christakakou, M.
Haider, M. Schnürch, Synthesis 2014, 1421; c) L. Ackermann,
Acc. Chem. Res. 2014, 47, 281; d) G. Song, F. Wang, X. Li, Chem.
Soc. Rev. 2012, 41, 3651; e) P. B. Arockiam, C. Bruneau, P. H.
Dixneuf, Chem. Rev. 2012, 112, 5879; f) J. Wencel-Delord, T.
Drçge, F. Liu, F. Glorius, Chem. Soc. Rev. 2011, 40, 4740.
À
[2] For C H functionalization reactions in total synthesis, see:
a) W. R. Gutekunst, P. S. Baran, Chem. Soc. Rev. 2011, 40, 1976;
b) L. McMurray, F. OꢀHara, M. J. Gaunt, Chem. Soc. Rev. 2011,
40, 1885.
[3] a) S. Hansson, A. Heumann, T. Rein, B. Škermark, J. Org.
Chem. 1990, 55, 975; b) A. Heumann, B. Škermark, Angew.
Chem. Int. Ed. Engl. 1984, 23, 453; Angew. Chem. 1984, 96, 443.
[4] a) H. Grennberg, V. Simon, J.-E. Bäckvall, J. Chem. Soc. Chem.
Commun. 1994, 265; b) H. Grennberg, J.-E. Bäckvall, Chem.
Eur. J. 1998, 4, 1083.
[5] a) M. S. Chen, N. Prabagaran, N. A. Labenz, M. C. White, J. Am.
Chem. Soc. 2005, 127, 6970; b) M. S. Chen, M. C. White, J. Am.
Chem. Soc. 2004, 126, 1346; c) K. J. Fraunhoffer, N. Prabagaran,
L. E. Sirois, M. C. White, J. Am. Chem. Soc. 2006, 128, 9032.
[6] a) J. Mazuela, D. Banerjee, J.-E. Bäckvall, J. Am. Chem. Soc.
2015, 137, 9559; b) C. Zhu, B. Yang, T. Jiang, J.-E. Bäckvall,
Angew. Chem. Int. Ed. 2015, 54, 9066; Angew. Chem. 2015, 127,
9194; c) A. K. Š. Persson, T. Jiang, M. T. Johnson, J.-E. Bäck-
vall, Angew. Chem. Int. Ed. 2011, 50, 6155; Angew. Chem. 2011,
123, 6279; d) J. Piera, A. Persson, X. Caldentey, J.-E. Bäckvall, J.
Am. Chem. Soc. 2007, 129, 14120; e) J. FranzØn, J.-E. Bäckvall, J.
Am. Chem. Soc. 2003, 125, 6056.
Scheme 6. Proposed reaction mechanism.
To demonstrate the synthetic utility of the products, we
performed an iridium-catalyzed asymmetric hydrogenation
reaction of the a-acetoxylated enone 2b by using a chiral
N,P ligand.[32] Upon selective reduction of the olefin moiety,
the enantiomerically enriched a-acetoxylated ketone 5,
a valuable building block, was obtained in 81% yield with
e.r. 80:20 (Scheme 7).[33]
[7] For an enantioselective study, see: T. Jiang, T. Bartholomeyzik, J.
Mazuela, J. Willersinn, J.-E. Bäckvall, Angew. Chem. Int. Ed.
2015, 54, 6024; Angew. Chem. 2015, 127, 6122.
[8] a) Y. Deng, J.-E. Bäckvall, Angew. Chem. Int. Ed. 2013, 52, 3217;
Angew. Chem. 2013, 125, 3299; b) Y. Deng, T. Bartholomeyzik,
J.-E. Bäckvall, Angew. Chem. Int. Ed. 2013, 52, 6283; Angew.
Chem. 2013, 125, 6403; c) T. Bartholomeyzik, J. Mazuela, R.
Pendrill, Y. Deng, J.-E. Bäckvall, Angew. Chem. Int. Ed. 2014,
53, 8696; Angew. Chem. 2014, 126, 8840.
Scheme 7. Asymmetric hydrogenation of a-acetoxylated enones.
tAmOH=2-methyl-2-butanol, cod=1,5-cyclooctadiene.
In conclusion, we have developed a palladium-catalyzed
oxidative protocol for the synthesis of a-acetoxylated enones
from readily available alkynes. This reaction shows good
tolerance of a range of functional groups. Mechanistic studies
revealed that DMSO acts as an oxygen nucleophile and
[9] For a leading review on palladium – alkyne chemistry, see: R.
Chinchilla, C. Nµjera, Chem. Rev. 2014, 114, 1783.
À
[10] For direct propargylic C H functionalization reactions, see:
a) S. M. Paradine, J. R. Griffin, J. Zhao, A. L. Petronico, S. M.
Miller, M. C. White, Nat. Chem. 2015, 7, 987; b) H. Lu, C. Li, H.
Jiang, C. L. Lizardi, X. P. Zhang, Angew. Chem. Int. Ed. 2014, 53,
7028; Angew. Chem. 2014, 126, 7148; c) T. Wang, W. Zhou, H.
Yin, J. A. Ma, N. Jiao, Angew. Chem. Int. Ed. 2012, 51, 10823;
Angew. Chem. 2012, 124, 10981; d) R. D. Grigg, J. W. Rigoli,
S. D. Pearce, J. M. Schomaker, Org. Lett. 2012, 14, 280; e) D.
Cheng, W. Bao, J. Org. Chem. 2008, 73, 6881; f) E. C. Mclaughlin,
M. P. Doyle, J. Org. Chem. 2008, 73, 4317.
À
propargylic C H cleavage occurs at a later stage of the
reaction, as evidenced by KIE studies. We are currently
À
pursuing further studies on propargylic C H oxidative
functionalization.
Acknowledgments
[11] For a review of sulfoxide ligands in transition-metal catalysis,
see: G. Sipos, E. E. Drinkel, R. Dorta, Chem. Soc. Rev. 2015, 44,
3834.
[12] a) R. Mahrwald, H. Schick, Angew. Chem. Int. Ed. Engl. 1991,
30, 593; Angew. Chem. 1991, 103, 577; b) A. Bartels, R.
Mahrwald, K. Müller, Adv. Synth. Catal. 2004, 346, 483; c) Y.
Peng, L. Cui, G. Zhang, L. Zhang, J. Am. Chem. Soc. 2009, 131,
5062; d) K. Ji, J. Nelson, L. Zhang, Beilstein J. Org. Chem. 2013,
9, 1925; e) B. G. Pujanauski, B. A. Bhanu Prasad, R. Sarpong, J.
Am. Chem. Soc. 2006, 128, 6786.
Financial support from the European Research Council
(ERC AdG 247014), the Swedish Research Council (621-
2013-4653), and the Berzelii Center EXSELENT is gratefully
acknowledged. We also thank Cristiana Margarita for LC–MS
analysis and Prof. Lars Eriksson (Department of Materials
and Environmental Chemistry, Stockholm University) for X-
ray crystallographic analysis.
[13] T. Sun, X. Zhang, Adv. Synth. Catal. 2012, 354, 3211.
Angew. Chem. Int. Ed. 2016, 55, 5824 –5828
ꢀ 2016 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5827

Angewandte
Communications
Chemie
[14] For the synthesis of enones from alkynes, see: B. Lu, C. Li, L.
Zhang, J. Am. Chem. Soc. 2010, 132, 14070.
[15] Control experiments showed that 2a was not converted into 3a
under the reaction conditions and vice versa.
[16] The use of [bis(trifluoroacetoxy)iodo]benzene instead of PIDA
led to a complex crude product mixture.
[17] For detailed results of the optimization study, see the Supporting
Information.
[18] Acidic additives such as HOAc and Ac2O diminished the yield of
2a. Basic additives, including various sodium and potassium
salts, DIPEA, TMP and 2,6-lutidine, hindered the reaction.
[19] The configuration of olefins was assigned according to Ref. [12d]
and by analogy with the X-ray crystal structure of 2j.
[20] The loss of 18O from 86% in DMSO to 62% in the product is
most likely due to exchange of the ketone with traces of water.
[21] For details of KIE studies, see the Supporting Information.
[22] E. M. Simmons, J. F. Hartwig, Angew. Chem. Int. Ed. 2012, 51,
3066; Angew. Chem. 2012, 124, 3120.
[28] For a related reductive elimination from PdIV, see: A. R. Dick,
K. L. Hull, M. S. Sanford, J. Am. Chem. Soc. 2004, 126, 2300.
[29] The direct formation of int-3 from int-1 through b-hydride
elimination cannot be completely ruled out: a) I. T. Crouch, T.
Dreier, D. E. Frantz, Angew. Chem. Int. Ed. 2011, 50, 6128;
Angew. Chem. 2011, 123, 6252; b) I. T. Crouch, R. K. Neff, D. E.
Frantz, J. Am. Chem. Soc. 2013, 135, 4970; c) N. Nella, E. Parker,
J. Hitce, P. Larini, R. Jazzar, O. Baudoin, Chem. Eur. J. 2014, 20,
13272.
[30] A mechanism involving the formation of int-3 from 1 through
À
direct propargylic C H activation can be ruled out on the basis
of the negligible KIE in the competition experiment (Sche-
me 5a). Such a pathway would have given a large kH/kD value in
the intermolecular competition experiment.
[31] A. Vasseur, J. Muzart, J. Le Bras, Eur. J. Org. Chem. 2015, 4053.
[32] a) J. Mazuela, J. J. Verendel, M. Coll, B. Schaffner, A. Bçrner,
P. G. Andersson, O. Pamies, M. Dieguez, J. Am. Chem. Soc. 2009,
131, 12344; b) P. Kaukoranta, M. Engman, C. Hedberg, J.
Bergquist, P. G. Andersson, Adv. Synth. Catal. 2008, 350, 1168;
c) P. Cheruku, J. Diesen, P. G. Andersson, J. Am. Chem. Soc.
2008, 130, 5595; d) P. Cheruku, S. Gohil, P. G. Andersson, Org.
Lett. 2007, 9, 1659.
[33] A minor amount of side-product formation through possible
ketone reduction followed by intramolecular 1,2-acyl migration
was observed. For results with different N,P ligands, see the
Supporting Information. For a rhodium-catalyzed protocol with
P, P ligands, see Ref. [13].
ˇ
´
[23] P. Kocovsky, J.-E. Bäckvall, Chem. Eur. J. 2015, 21, 36.
[24] a) S. Mori, M. Takubo, T. Yanase, T. Maegawa, Y. Monguchi, H.
Sajiki, Adv. Synth. Catal. 2010, 352, 1630; b) K.-W. Chi, M. S.
Yusubov, V. D. Filimonov, Synth. Commun. 1994, 24, 2119.
[25] Diketone 3 may be formed by a similar pathway to that
described by Sajiki and co-workers (see Ref. [24a]).
[26] The aryl group ensures the reactivity and selectivity of the
reaction, as the reaction of 3-hexyne gave the desired product in
21% yield.
[27] a) P. Sehnal, R. J. K. Taylor, I. J. S. Fairlamb, Chem. Rev. 2010,
110, 824; b) L.-M. Xu, B.-J. Li, Z. Yang, Z.-J. Shi, Chem. Soc.
Rev. 2010, 39, 712; c) K. MuÇiz, Angew. Chem. Int. Ed. 2009, 48,
9412; Angew. Chem. 2009, 121, 9576.
Received: January 21, 2016
Published online: April 6, 2016
5828
www.angewandte.org
ꢀ 2016 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2016, 55, 5824 –5828