Access to well-defined Ruthenium(I) and Osmium(I) metalloradicals

Article abstract of DOI:10.1002/anie.201001199

Figure Presented Radically complex: Well-defined mononuclear Ru^I and Os^I complexes (see scheme) have metalloradical character, as indicated by EPR spectroscopy and DFT calculations. The Ru^I and Os^I metalloradicals exhibit both one-electron and two-electron redox reactivity. The latter process affords unusual imido complexes with substantial radical character on the {ArN} moiety.

Full text of DOI:10.1002/anie.201001199

Communications
DOI: 10.1002/anie.201001199
Metalloradical Complexes
Access to Well-Defined Ruthenium(I) and Osmium(I)
Metalloradicals**
Ayumi Takaoka, Laura C. H. Gerber, and Jonas C. Peters*
Low-valent metalloradicals of the late second- and third-row
transition metals have garnered recent attention in the
context of their interesting spectroscopic properties and
potential applicability in catalysis.[1] Examples of mononu-
clear RuI and OsI metalloradicals are particularly rare.[2] As
these species are inherently unstable, studies that extend
beyond attempts to rapidly characterize them in situ are not
available. As a consequence, the chemistry of mononuclear
RuI and OsI complexes is essentially unexplored.[3]
Scheme 1. Synthesis of mononuclear RuI and OsI complexes.
Recently, we reported the first mononuclear complexes of
FeI with terminal dinitrogen ligands.[4] The iron centers in
these complexes are chelated by bulky tetradentate tris(phos-
phino)silyl ligands (SiPR3)À ((SiPR3)À = (2-R2PC6H4)3SiÀ, R =
Ph, iPr) that favor mono- rather than dinuclear species. The
steric influence provided by this scaffold and its ability to
accommodate the FeI oxidation state made it a plausible
candidate for exploring access to the unusual oxidation states
RuI and OsI. Herein, we report structural, spectroscopic, and
theoretical studies of well-defined and mononuclear RuI and
OsI complexes, [(SiPiPr3)M(L)] (M = Ru, Os; L = N2, PMe3).
To our knowledge, these complexes are the first such
examples to be isolated and thoroughly characterized,
including X-ray diffraction studies. Moreover, initial reac-
tivity studies with [(SiPiPr3)M(N2)] (M = Ru, Os) complexes
reveal both one- and two-electron reactivity. The latter
reaction affords unusual imido complexes, [(SiPiPr3)M(NAr)]
(M = Ru, Os; Ar= C6H4CF3), that display substantial
“imidyl” radical character. In contrast to the highly unstable
and structurally related FeIII imido derivative, which can only
be observed in a frozen glass,[5] these imidyl radicals are
sufficiently long-lived to be isolated in pure form.
KC8 resulted in green [(SiPiPr3)M(N2)] (M = Ru (3), Os(4)) in
85% and 70% yields, respectively. The 1H NMR spectra of 3
and 4 are similar and show broad features between d = À1–
11 ppm, which is consistent with their expected paramagnet-
ism (S = 1/2). The IR spectra of 3 and 4 show strong vibrations
at 2088 and 2052 cmÀ1, which correspond to the nitrogen
ligands.
Crystals suitable for X-ray diffraction were grown by slow
evaporation of concentrated solutions of 3 and 4 in pentane.
Unlike [(SiPiPr3)Fe(N2)], which is rigorously trigonal bipyr-
amidal (TBP),[5] the solid-state structures of 3 (see Figure 2)
and 4 (see the Supporting Information) feature substantive
distortions from TBP geometries (t = 0.76 (3) and 0.70 (4))[6]
with one of the P-M-P angles notably larger than the others.
À
The N N bond lengths are short (1.097(5) (3) and 1.101(6) ꢀ
(4)) and are consistent with the high nN2 values. The N2 ligands
in 3 and 4 are labile, and addition of one equivalent of PMe3
results in formation of the phosphine adducts, [(SiPiPr3)M-
(PMe3)] (M = Ru (5), Os(6)). Compound 5 has also been
crystallographically characterized (see the Supporting Infor-
mation) and has a geometry similar to 3 (t = 0.86).
Precursors to the MI (M = Ru, Os) complexes,
[(SiPiPr3)MCl] (M = Ru (1), Os (2)), were prepared by heating
a mixture of HSiPiPr3, [(h6-C6H6)M(Cl)(m-Cl)]2, and Et3N in
toluene to produce red 1 and brown 2 in 94% and 95% yields,
respectively (Scheme 1). Chemical reduction of 1 and 2 with
The cyclic voltammogram of 3 shows one oxidation event
at À1.24 V (vs. ferrocene/ferrocenium; Fc/Fc+), and one
reduction event at À2.14 V, which are assigned to the
formal RuII/I and RuI/0 couples, respectively. Chemical oxida-
tion and reduction of 3 with FcBArF (BArF4 = tetrakis(3,5-
[*] A. Takaoka,[+] L. C. H. Gerber, Prof. J. C. Peters[+]
Department of Chemistry, Massachusetts Institute of Technology
Cambridge, MA 02139 (USA)
[+] Current address: Division of Chemistry and Chemical Engineering
California Institute of Technology
Pasadena, CA 91125 (USA)
4
bis(trifluoromethyl)phenyl)borate) and KC8 leads to the
corresponding RuII and Ru0 dinitrogen complexes,
[(SiPiPr3)Ru(N2)][BArF ] (7) and [K(THF)x][(SiPiPr3)Ru(N2)]
4
(8), respectively, which have also been crystallographically
characterized (see the Supporting Information). Complexes 3,
7, and 8 represent a rare series of mononuclear N2 complexes
that span three distinct oxidation states (Scheme 2); the
related (SiPiPr3)Fe system also stabilizes a corresponding N2
series.[7] The N2 ligand in 7 is very labile and appears to be in
equilibrium with an N2-free species,[8] as evidenced by the
shift in the RuII/I couple under argon and its 15N NMR
spectrum, which only shows resonances for the coordinated
Fax: (+1)626-395-6948
E-mail: jpeters@caltech.edu
[**] This work was supported by the NIH (GM070757). Samantha
MacMillan and Dr. Peter Mꢀller are acknowledged for crystallo-
graphic assistance. The NSF is acknowledged for use of instru-
ments at the MIT DCIF (CHE-9808061, DBI-97299592).
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201001199.
4088
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2010, 49, 4088 –4091

Angewandte
Chemie
been assigned as ligand-centered radicals.[13,14] The noticeably
larger Dg values for the Os relative to the Ru complexes are
likely due to a greater spin–orbit coupling constant for the
heavier metal.[15] Although g values alone cannot be used as a
quantitative measure of spin density, the simulated EPR
parameters support our formulations of 3–6 as bona fide
metalloradicals. As a test of our assignment, Mulliken spin
densities (SD) were calculated for 3–6 (Figure 2 and the
Supporting Information). These calculations place 76% (3),
69% (4), 84% (5), and 79% (6) of the SD at the metal center.
In addition, 16% (3), 15% (4), 13% (5), and 13% (6) of the
SD is located at the phosphine units.[16] In each complex, one
of the P atoms possesses a greater value relative to the other
two; this result is consistent with the EPR simulations that
suggest an unpaired spin in the equatorial plane.
Scheme 2. Oxidation and reduction of 3 and reduction of 4.
N2 at low temperature.[9] While the cyclic voltammogram of 4
also displays a reduction event at À1.94 V, an irreversible
oxidation event at À1.17 V is observed. The reduction
product, [K(THF)x][(SiPiPr3)Os(N2)] (9) was accessed in a
similar procedure to that of 8 and its solid-state structure is
isostructural with 8 (see the Supporting Information).
Although complexes 3–6 are formally RuI and OsI
Treatment of 3 with nBu3SnH provided chemical evidence
of its metalloradical character as the reaction cleanly afforded
the hydride complex [(SiPiPr3)Ru(H)(N2)] (10) over 24 hours;
this behavior is similar to the reactivity of other metal-
complexes, the possibility of
a ligand-centered radical
cannot be excluded based on structural studies alone,
especially in light of the growing recognition of redox non-
innocence of many auxiliary ligands.[10] To investigate the
distribution of spin density in 3–6, their EPR spectra were
measured at 77 K in toluene glass (Figure 1 and the Support-
ing Information). Each spectrum exhibits rhombic features
with large hyperfine coupling to one phosphorus atom,
consistent with unpaired spin density localized in an orbital
of the equatorial plane of the TBP.
centered radicals towards nBu3SnH.[17a] In addition, 3 reacts
II
À
cleanly with I2 and PhS SPh to afford the corresponding Ru
iodide and thiolate complexes, [(SiPiPr3)RuI] and
[(SiPiPr3)RuSPh] (see the Supporting Information).
The reactivity of late-second- and third-row metallorad-
icals often follows one-electron processes.[1,17] Having
In assessing metal radical character, the anisotropy of the
g values (Dg = gmaxÀgmin) is particularly noteworthy, since a
large Dg value has been noted as a crude indication of
metalloradical character for S = 1/2 systems.[11,12] Overall the
Dg values for 3–6, which are 0.135, 0.257, 0.166, and 0.318
respectively, are significantly larger than complexes that have
Figure 2. Solid-state structures (thermal ellipsoids set at 50% proba-
bility) and spin density plots (0.0004 isocontours) for 3 (top) and
11 (bottom). Selected bond lengths [ꢁ] and angles [o] for 3: Ru–N1
2.049(3), Ru–P1 2.3221(9), Ru–P2 2.3743(9), Ru–P3 2.3253(9), Ru–Si
3.2187(9), N1–N2 1.097(5); Si-Ru-N1 177.0(1), P1-Ru-P2 109.53(3),
P2-Ru-P3 131.32(3), P1-Ru-P3 111.85(3). For 11, Ru–N 1.869(2), Ru–P1
2.2968(7), Ru–P2 2.4242(6), Ru–P3 2.3756(7), Ru–Si 2.3949(6);
Si-Ru-N 162.4(1), Ru-N-C(Ar) 172.0(2), P1-Ru-P2 106.30(2), P2-Ru-P3
130.24(2), P3-Ru-P1 109.71(2).
Figure 1. a) EPR spectrum of 3 (77 K). (gx, gy, gz)=(2.130, 2.076,
1.995). b) EPR spectrum of 4 (77 K). (gx, gy, gz)=(2.239, 2.133, 1.982)
c) EPR spectrum of 5 (77 K). (gx, gy, gz)=(2.175, 2.075, 2.009). d) EPR
spectrum of 11 (RT). giso =2.020. The lower curves are simulations.
See the Supporting Information for other parameters.
Angew. Chem. Int. Ed. 2010, 49, 4088 –4091
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 4089

Communications
observed one-electron reactivity in 3 we sought, in turn, to
imido complex that was suggested to possess considerable
ligand radical character.[25] Further supporting the largely
ligand-centered radical character of 11 and 12, the EPR
spectra at 77 K reveal much smaller g anisotropies (Dg =
0.072 (11) and 0.128 (12)) relative to their corresponding
Ru(I) and Os(I) metalloradicals 3–6. DFT calculations are
consistent with the EPR parameters and show that 54% (11)
and 54% (12) of the SD is distributed throughout the NAr
moiety, of which 27% (11) and 24% (12) is on the nitrogen
atom and 40% (11) and 39% (12) is located at the metal
center (see the Supporting Information). While delocalization
investigate whether two-electron processes might also be
feasible. To this end, complex 3 was treated with organoazides
to see if metal imido or nitrene species would be formed
through loss of N2, akin to the recently observed reactivity of
related FeI complexes.[18] Treatment of 3 with para-CF3-
substituted phenylazide led to formation of the formally RuIII
imido species, [(SiPiPr3)Ru(NAr)] (Ar= C6H4CF3; 11;
Scheme 3). The solid-state structure of 11 (Figure 2) reveals
a geometry midway between TBP (t = 0.54) and square
À
pyramidal with a Ru N bond length of 1.869(2) ꢀ. Whilst this
À
À
bond length is significantly shorter than Ru N bonds between
of the spin density along the M NAr moiety is evident, both
typical ruthenium anilides (Ru–N > 1.95 ꢀ),[19] it is appreci-
ably longer than prototypical ruthenium imido complexes
(Ru–N < 1.80 ꢀ).[20] Treatment of 4 with para-CF3 substituted
phenylazide also leads to the corresponding OsIII imido
species [(SiPiPr3)Os(NAr)] (Ar= C6H4CF3; 12). Crystallo-
graphic characterization established that 12 is isostructural
to its ruthenium analogue 11 (see the Supporting Informa-
tion).
EPR and DFT data suggest that perhaps 11 and 12 are best
considered MII complexes with a ligand-localized radical
(Scheme 3). This ligand radical is a one-electron oxidized
Complexes 11 and 12 represent interesting examples of
five-coordinate, formally d5 imido complexes. Qualitative
molecular orbital diagrams predict low bond orders (less than
or equal to 1.5) because of the occupation of p* orbitals.[21] It
is emphasized that TBP complexes with metal–ligand multi-
ple bonds and d-electron configurations greater than 1 are
virtually unknown. Que and co-workers have reported a
noteworthy recent exception.[22] The stability of 11 and 12 is,
therefore, surprising and distinct from its chemically related
and highly unstable iron derivative [(SiPiPr3)Fe(N-p-tolyl)],
which has a calculated geometry[23] close to 11 and 12.
[(SiPiPr3)Fe(N-p-tolyl)] is only observable by EPR when
generated photolytically in a frozen glass, and decomposes
rapidly by presumed bimolecular nitrene coupling to yield
azobenzenes.[5] While complexes 11 and 12 decompose in
solution at room temperature over several days, they are
stable at À358C as solids for extended periods.
Scheme 3. Synthesis of formal MIII imido complexes with significant
radical character on the NAr unit.
imido ligand (·NAr)À and exhibits properties of a rare imidyl
radical that has only very recently been described in
coordination chemistry.[25,26] The electronic configurations of
11 and 12 distinguish themselves from [(SiPiPr3)Fe(N-p-
tolyl)],[5] whose DFT-predicted ground state (S = 1= ) is
2
calculated to consist of a largely metal-centered radical.
In conclusion, we have introduced several well-defined
examples of mononuclear RuI and OsI complexes. These
unusual complexes have been shown, through EPR measure-
ments and DFT calculations, to consist of predominantly
metal-centered radical character with a minority of the spin
density delocalized onto the chelated phosphines. The
reactivity of the dinitrogen adduct derivatives 3 and 4 were
shown to exhibit formal MI/III group-transfer reactivity.
Detailed analysis of the imido and nitrene products suggests
that they possess substantial imidyl radical character.
The
difference
in
stability/reactivity
between
[(SiPiPr3)Fe(N-p-tolyl)] and complexes 11 and 12 could
potentially be attributed to differences in electronic config-
uration. Though they are formally MIII imido complexes, close
examination of their EPR spectra indicate that they possess
significant nitrogen-centered radical character. Unlike the
spectra of 3–6, which show broad features at room temper-
ature, the spectra of 11 and 12 (Figure 1 and the Supporting
Information) show relatively sharp four-line patterns with
isotropic g values of 2.02 and 2.01, respectively, which are
much closer in value to that of a free electron (ge = 2.0023)
compared to the corresponding metalloradicals 3–6. Ruthe-
nium and osmium hyperfine coupling are also observed
(ARu = 48 MHz (11), AOs = 150 MHz (12)) and the spectra are
best simulated by assigning hyperfine coupling to one nitro-
gen atom (AN = 98 MHz (11), AN = 93 MHz (12)) and smaller
coupling to one phosphorus atom (AP = 64 MHz (11), AP =
58 MHz (12)). These isotropic AN values are surprisingly
large. For comparison, the similarly sp-hybridized NO radical
Received: February 26, 2010
Published online: May 7, 2010
Keywords: group transfer · imido ligands · imidyl ligands ·
.
metalloradicals · nitrogen complexes
[1] B. de Bruin, D. G. H. Hetterscheid, A. J. J. Koekkoek, H.
Grꢁtzmacher, Prog. Inorg. Chem. 2007, 55, 247.
[2] a) For RuI, see: C. Bianchini, F. Laschi, M. Peruzzini, P. Zanello,
Gazz. Chim. Ital. 1994, 124, 271; b) R. J. Angelici, B. Zhu, S.
Fedi, F. Laschi, P. Zanello, Inorg. Chem. 2007, 46, 10901; c) Q. G.
Mulazzani, S. Emmi, P. G. Fuochi, M. Z. Hoffman, M. Venturi, J.
Am. Chem. Soc. 1978, 100, 981; d) S. D. Robinson, M. F. Uttley,
J. Chem. Soc. Dalton Trans. 1972, 1; e) G. Zotti, G. Pilloni, M.
Bressan, M. Martelli, J. Electroanal. Chem. 1977, 75, 607; for OsI
see: f) C. Bianchini, M. Peruzzini, A. Ceccanti, F. Laschi, P.
Zanello, Inorg. Chim. Acta 1997, 259, 61; g) N. V. Vugman, A. M.
Rossi, J. Danon, J. Chem. Phys. 1978, 68, 3152; h) G. Zotti, G.
has
a
nitrogen hyperfine coupling constant of AN =
77 MHz.[24] In addition, the Ru hyperfine coupling constant
ARu in 11 is smaller than a spectroscopically detected RuIII
4090 www.angewandte.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2010, 49, 4088 –4091

Angewandte
Chemie
Pilloni, M. Bressan, M. Martelli, Inorg. Chim. Acta 1978, 30,
L311.
[12] Some complexes with large Dg values have been found to be
mainly ligand-based radicals. See: Y. Miyazato, T. Wada, J. T.
Muckerman, E. Fujita, K. Tanaka, Angew. Chem. 2007, 119,
5830; Angew. Chem. Int. Ed. 2007, 46, 5728.
[3] a) M. Schrꢂder, T. A. Stephenson in Comprehensive Coordina-
tion Chemistry, Vol. 4 (Ed.: G. Wilkinson), Permagon, Oxford,
1987, pp. 277 – 519; b) W. P. Griffith in Comprehensive Coordi-
nation Chemistry, Vol. 4 (Ed.: G. Wilkinson), Permagon, Oxford,
1987, pp. 519 – 635; c) C. E. Housecroft, in Comprehensive
Coordination Chemistry II, Vol. 5 (Eds.: J. A. McCleverty, T. J.
Meyer), Elsevier, Oxford, 2004, pp. 555 – 733.
[13] a) T. Bꢁttner, J. Geier, G. Frison, J. Harmer, C. Calle, A.
Schweiger, H. Schꢂnberg, H. Grꢁtzmacher, Science 2005, 307,
235; b) L. Cataldo, S. Choua, T. Berclaz, M. Geoffroy, N.
Mꢃzailles, N. Avarvari, F. Mathey, P. Le Floch, J. Phys. Chem. A
2002, 106, 3017.
[4] a) N. P. Mankad, M. T. Whited, J. C. Peters, Angew. Chem. 2007,
119, 5870; Angew. Chem. Int. Ed. 2007, 46, 5768; b) M. T.
Whited, N. P. Mankad, Y. Lee, P. F. Oblad, J. C. Peters, Inorg.
Chem. 2009, 48, 2507.
[14] Two mixed-valent dinuclear Ru complexes, each featuring a
formal RuI center, exhibit Dg values of 0.20 and 0.26. See B.
Sarkar, W. Kaim, J. Fiedler, C. Duboc, J. Am. Chem. Soc. 2004,
126, 14706.
[5] N. P. Mankad, J. C. Peters, J. Am. Chem. Soc. 2010, 132, 4083.
[6] t = (bÀa)/60 where b and a represent the two largest angles. t =
1 for TBP, 0 for square pyramid. See: A. W. Addison, T. N. Rao,
J. J. Van Rijn, G. C. Veschoor, J. Chem. Soc. Dalton Trans. 1984,
1349. CCDC 764569 (3), 764570 (4), 764571 (5), 764572 (10),
764573 (11), 765970 (12), 765971 (7), 765972 (8), and 765973 (9)
contain the supplementary crystallographic data for this paper.
These data can be obtained free of charge from The Cambridge
Crystallographic Data Centre via www.ccdc.cam.ac.uk/data_
request/cif.
[15] E. M. Kober, T. J. Meyer, Inorg. Chem. 1984, 23, 3877.
[16] Calculated from DFT-optimized structures. See the Supporting
Information for details.
[17] a) M. C. Baird, Chem. Rev. 1988, 88, 1217; b) R. S. Paonessa,
N. C. Thomas, J. Halpern, J. Am. Chem. Soc. 1985, 107, 4333.
[18] a) S. D. Brown, T. A. Betley, J. C. Peters, J. Am. Chem. Soc. 2003,
125, 322; b) I. Nieto, F. Ding, R. P. Bontchev, H. Wang, J. M.
Smith, J. Am. Chem. Soc. 2008, 130, 2716; c) R. E. Cowley, N. A.
Eckert, J. Lehaꢄk, P. L. Holland, Chem. Commun. 2009, 1760.
[19] For selected examples, see: a) K. N. Jayaprakash, T. B. Gunnoe,
P. D. Boyle, Inorg. Chem. 2001, 40, 6481; b) J. L. Liang, J. S.
Huang, Z. Y. Zhou, K. K. Cheung, C. M. Che, Chem. Eur. J.
2001, 7, 2306.
[7] A related series of N2 complexes has been characterized for the
corresponding iron system. See: Y. Lee, N. P. Mankad, J. C.
Peters, Nat. Chem. 2010, in press.
[8] A crystal grown from a THF/pentane mixture at room temper-
ature was shown to have a solid-state structure with neither N2
nor THF bound trans to the silyl anchor. This result contrasts
that of crystals grown from a CH2Cl2/pentane mixture at À358C
that afforded the N2 adduct structure 7. Both the N2-bound and
N2-free structures show an unusually close contact distance to
one methyl group from the SiPiPr3 ligand, and is indicative of an
agostic interaction. See the Supporting Information for struc-
tural details.
[20] For selected examples, see: a) A. K. Burell, A. J. Steedman, J.
Chem. Soc. Chem. Commun. 1995, 2109; b) A. A. Danopoulos,
G. Wilkinson, B. Hussain-Bates, M. B. Hursthouse, Polyhedron
1992, 11, 2961.
[21] T. A. Betley, Q. Wu, T. Van Voorhis, D. G. Nocera, Inorg. Chem.
2008, 47, 1849.
[22] For a d4 multiply bonded species with TBP geometry see: J.
England, M. Martinho, E. R. Farquhar, J. R. Frisch, E. L.
Bominaar, E. Mꢁnck, L. Que Jr. , Angew. Chem. 2009, 121,
3676; Angew. Chem. Int. Ed. 2009, 48, 3622.
[9] This sample was prepared from an independent route involving
salt metathesis of complex 1 with NaBArF , thus ruling out any
[23] The calculated structure has iPr groups replaced with Me groups
and Ar= Ph instead of PhMe. t = 0.49 for this compound. See
Ref. [5].
4
exchange process with impurities of 3 that would broaden the
resonances of the spectrum.
[10] a) C. C. Lu, E. Bill, T. Weyhermꢁller, E. Bothe, K. Wieghardt, J.
Am. Chem. Soc. 2008, 130, 3181; b) D. Adhikari, S. Mossin, F.
Basuli, J. C. Huffman, R. K. Szilagyi, K. Meyer, D. J. Mindiola, J.
Am. Chem. Soc. 2008, 130, 3676; c) S. B. Harkins, N. P. Mankad,
A. J. Miller, R. K. Szilagyi, J. C. Peters, J. Am. Chem. Soc. 2008,
130, 3478; d) M. R. Haneline, A. F. Heyduk, J. Am. Chem. Soc.
2006, 128, 8410.
[24] N. A. Ashford, F. H. Jarke, I. J. Solomon, J. Chem. Phys. 1972, 57,
3867.
[25] A. N. Walstrom, B. C. Fullmer, H. Fan, M. Pink, D. T. Busch-
horn, K. G. Caulton, Inorg. Chem. 2008, 47, 9002.
[26] C. C. Lu, S. D. George, T. Weyhermꢁller, E. Bill, E. Bothe, K.
Wieghardt, Angew. Chem. 2008, 120, 6484; Angew. Chem. Int.
Ed. 2008, 47, 6384.
[11] B. de Bruin, D. Hetterscheid, Eur. J. Inorg. Chem. 2007, 211.
Angew. Chem. Int. Ed. 2010, 49, 4088 –4091
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 4091