An open-shell spin singlet copper-nitrene intermediate binding redoxinnocent metal ions: Influence of the Lewis acidity of the metal ions on spectroscopic and reactivity properties

Article abstract of DOI:10.1002/zaac.201400389

The yield of the Lewis acid adducts of a copper-tosylnitrene intermediate and their decay are dependent on the Lewis acidity of the redox innocent metal ions. Surprisingly, however, their spectroscopic properties and reactivity in nitrene transfer reaction with PPh3 are independent of the nature of the Lewis acid used. The result provides some new insights into the structure and reactivity of the Lewis-acid bound copper nitrene complexes.

Full text of DOI:10.1002/zaac.201400389

Journal of Inorganic and General Chemistry
www.zaac.wiley-vch.de
ARTICLE
Zeitschrift für anorganische und allgemeine Chemie
DOI: 10.1002/zaac.201400389
An Open-Shell Spin Singlet Copper-Nitrene Intermediate Binding Redox-
innocent Metal Ions: Influence of the Lewis Acidity of the Metal Ions on
Spectroscopic and Reactivity Properties
Inés Monte-Pérez,[a] Subrata Kundu,[a] and Kallol Ray*[a]
Dedicated to Professor Lawrence Que Jr. on the Occasion of His 65th Birthday
Keywords: Copper; Nitrene; Nitrene transfer; Lewis acids
Abstract. The yield of the Lewis acid adducts of a copper-tosylnitrene
pendent of the nature of the Lewis acid used. The result provides some
intermediate and their decay are dependent on the Lewis acidity of the new insights into the structure and reactivity of the Lewis-acid bound
redox innocent metal ions. Surprisingly, however, their spectroscopic
properties and reactivity in nitrene transfer reaction with PPh3 are inde-
copper nitrene complexes.
the nitrene nitrogen thereby reducing the strong electron repul-
sion between the electron-rich nitrogen and central copper
atoms. Notably, a similar interaction between a high-valent
manganese-oxo core and Ca2+ ion is proposed to play a vital
role during biological oxygen evolution in photosystem II.[22–
Introduction
Copper-catalyzed alkane amination and alkene aziridation
reactions have been proposed to proceed via copper-nitrene
intermediates with a formal +3 oxidation state of copper.[1–12]
Given the fundamental role these species have, the understand-
ing of their electronic structure and chemical reactivity are of
substantial importance both for obtaining new mechanistic in-
sights into oxidation catalysis and for engineering new cata-
lytic reagents and processes. However, despite considerable
interest, no copper-nitrene intermediates have been isolated to
allow for clear mechanistic studies.[13,14] In the absence of any
isolable intermediate the electronic structure of the active spe-
cies responsible for copper-mediated aziridination or amination
reactions has been controversially discussed in terms of CuI-
nitrene, CuII-imidyl radical, or CuIII-imido models.[1–20]
Very recently, our group reported the spectroscopic charac-
terization of a Lewis acid adduct of a cationic copper-tosylni-
trene species [(AN)CuII-(NTs•)Sc(OTf)3]+ (1-Sc) [Ts = tosyl;
AN = 3,3Ј-iminobis(N,N-dimethylpropylamine)], the elec-
tronic structure of which could be best described in terms of a
CuII ion antiferromagnetically coupled to a nitrene radical
thereby stabilizing an open-shell singlet ground state.[21] In this
study although the binding-site of the scandium ion was not
unambiguously identified, it is believed that the effect of the
Lewis acid in the stabilization of the copper-nitrene intermedi-
ate is related to its interaction with the otherwise transient
[(AN)CuII(NTs•)]+ core, which lowers the electron density at
25]
The wider application of the strategy of utilizing redox-
innocent Lewis acids in trapping the short-lived intermediates
responsible for the oxo- and nitrene-transfer reactions medi-
ated by mid-to-late transition metals has also been demon-
strated in the successful trapping of a low-spin (S = 1/2)[26] and
an intermediate-spin (S = 3/2)[27] cobalt(IV)-oxo intermediates.
The successful stabilization of a copper-nitrene core allowed
us to carry out an extensive reactivity study on this important
intermediate. For example, [(AN)CuII-(NTs•)Sc(OTf)3]+ was
found to be capable of aminating the sp3 C–H bonds in sub-
strates such as toluene and even cyclohexane in modest yields
(21–35%) by employing a hydrogen atom transfer (HAT)
mechanism.[21] The Lewis acid stabilization of a related copper
nitrene species [(AN)CuII-(NMes•)Sc(OTf)3]+ (2-Sc) (Mes =
mesityl), where the electron withdrawing heteroatom Ts-ni-
trene substituent of 1-Sc is replaced by the electron donating
Mes group in 2-Sc has also been reported in a subsequent
study.[28] A comparative reactivity study between 1-Sc and
2-Sc showed that a two-state reactivity (TSR) model[29,30] ap-
plies for the two compounds, in which a singlet ground state
and a nearby triplet excited-state both contribute to the reactiv-
ity.
In the present study we have synthesized a series of
[CuII-(NTs•)M]n+ (1-M) (Scheme 1) complexes, where the
Lewis acidity of the redox-innocent metal ions has been sys-
tematically varied in the order of their decreasing Lewis-acid-
ity Sc3+ Ͼ Y3+ Ͼ Eu3+ Ͼ Ce3+ Ͼ Zn2+ Ͼ Ca2+ and investi-
gated their spectroscopic properties as well as their reactivity
with triphenylphosphine (PPh3). This study provides some new
* Dr. K. Ray
Fax: +49-3020937387
E-Mail: kallol.ray@chemie.hu-berlin.de
[a] Institut für Chemie
Humboldt-Universität zu Berlin
Brook-Taylor Str. 2
12489 Berlin, Germany
Z. Anorg. Allg. Chem. 2015, 641, (1), 78–82
78
© 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Journal of Inorganic and General Chemistry
www.zaac.wiley-vch.de
ARTICLE
Zeitschrift für anorganische und allgemeine Chemie
insights into the reactivity and structure of the copper-nitrene preventing its spontaneous decay to the copper(II)-amide spe-
core binding redox non-innocent metal ions.
cies 3. Consistent with the above explanation, use of weaker
Lewis acids like Y3+, Eu3+, Ce3+, or Zn2+ instead of Sc3+, re-
sults in a significantly lower yield of 1-M and an increased
formation of 3; for Ca2+ no formation of 1-M was observed.
Scheme 1. Synthesis and the possible tautomeric forms of the
[CuII-(NTs•)M]n+ intermediates.
Results and Discussion
As we previously reported, the reaction of [CuAN](BF4)
with the oxidant [N-(p-toluenesulfonyl)imino](2-tert-butylsulf-
onyl)-phenyliodinane (sPhINTs) in the presence of Sc(OTf)3 at
–90 °C results in the formation of 1-Sc with two intense ab-
sorption bands at λmax (εmax) = 530 nm (3500 L·mol–1·cm–1)
and 750 nm (580 L·mol–1·cm–1). Notably, in the absence of
Sc(OTf)3 the formal copper(III) oxidation state is not access-
ible; the reaction instead yields a copper(II)-tosylamide com-
plex (3) (Scheme 1), which is presumably the product of a
rapid hydrogen-atom abstraction by the transient copper-ni-
trene intermediate; the source of hydrogen atom is not clear at
this point. In contrast to 1-Sc, which is EPR silent, complex 3
shows a slightly rhombic signal in X-band EPR with the g-
values (gx = 2.09, gy = 2.07, and gz = 2.27) and hyperfine
constants (Ax = 54ϫ10–4 cm–1, Ay = 15ϫ10–4 cm–1, Az =
128ϫ10–4 cm–1) consistent with a dx2–y2 ground-state of the
central CuII ions.[21] The methodology of using Sc(OTf)3 as a
Lewis acid for trapping an elusive intermediate 1-Sc can be
extended to other redox-inactive Lewis-acidic metal ions
(Scheme 1). In presence of the triflate salts of other Lewis ac-
ids (e.g. Y3+, Eu3+, Ce3+, Zn2+, and Ca2+), the reaction of
Figure 1. Absorption spectra of 1-M obtained when M = Sc3+ (black
solid trace), Y3+ (dashed trace), Eu3+ (dotted trace), Ce3+ (dash-dot
trace), Zn2+ (dash-dot-dot trace), and Ca2+(gray solid trace). All made
form an initial 0.5 mm [CuAN](BF4) solution in CH2Cl2 at –90 °C.
Table 1. Yields of 1-M and 3 in presence of different Lewis acids (1.5
equiv.). Yields of 1-M and 3 are independently determined on the basis
of UV/Vis and EPR studies, respectively.[31]
.
M
Yield of 1-M /%
Yield of 3 /%
Sc3+
Y3+
Ͼ90
70
65
58
28
0
5
23
33
45
70
95
Eu3+
Ce3+
Zn2+
Ca2+
s
[CuAN](BF4) with PhINTs at –90 °C yields a mixture of two
The resonance Raman (rRaman) spectrum of 1-Y using
species namely, the formal CuIII purple intermediate 1-M and 514 nm laser excitation in resonance with the 530 nm transi-
the CuII-tosylamide species 3 (Figure 1; Table 1). The forma- tion (Figure 2) displays three bands at 568, 658 and 883 cm–1
tion of 1-M is confirmed on the basis of the characteristic ab- that can be attributed to the CuIII-NTs/CuII-N•Ts core on the
sorption features at 530 and 750 nm, and that of 3 by EPR basis of 15N-isotope labeling. Notably, these bands are only
spectroscopy. Interestingly, the position of the bands at 530 slightly shifted compared to the corresponding bands reported
and 750 nm is found to be independent of the nature of the previously for 1-Sc at 570, 660, and 887 cm–1, respectively.
Lewis acid used. Thus, the Lewis acidic metal ions do not Thus, the rRaman data together with the absorption features
affect the electronic transitions associated with the copper-im- of the 1-M complexes demonstrate that their spectral features
idyl unit in 1-M. However, they play a vital role in the stability predominantly arise from transitions involving the [CuNTs]+
of the copper-imidyl core as evident from the decreasing yield core, and are independent of the Lewis-acidity of the redox-
of the purple species 1-M and the increasing yield of the corre- innocent metal ions employed for the stabilization of 1-M.
sponding decay product 3 with decreasing Lewis acidity of M
Given that the yields/stabilities of the trapped intermediates
in the order Sc3+ Ͼ Y3+ Ͼ Eu3+ Ͼ Ce3+ Ͼ Zn2+ Ͼ Ca2+ are different when using different Lewis acids, it is natural to
(Table 1).[31] The stabilization of 1-M can be attributed to the wonder if the Lewis acidity also affects the reactivities of the
binding of M to the [(AN)CuNTs]+ core, which presumably 1-M complexes. To test it, the oxidative reactivity of the 1-M
lowers the hydrogen atom abstraction ability of 1-M, thereby adducts was evaluated with PPh3. Kinetic studies were per-
Z. Anorg. Allg. Chem. 2015, 78–82
79
© 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Journal of Inorganic and General Chemistry
www.zaac.wiley-vch.de
ARTICLE
Zeitschrift für anorganische und allgemeine Chemie
formed by addition of appropriate amounts of PPh3 to 0.3 mm
solutions of 1-M in CH2Cl2 under completely anaerobic condi-
tions. The decrease of the absorption band at 530 nm, assigned
to the intermediates, was directly monitored by a UV/Vis spec-
trophotometer. Rate constants, kobs, were determined by
pseudo-first-order fitting (in the initial 0–150 s) of the decrease
of this absorption band (for 1-Y see Figure 3A). Although de-
viation from pseudo-first order behavior could be observed in
the latter phases of the reactions, reasonable fits were obtained
for the initial part of the reactions in all cases. Plots of the
kobs values vs. PPh3 concentration afforded second-order rate
constants k2 (Figure 3B). It is noteworthy that the rate constant
values for the 1-M complexes are independent of the nature of
the metal ions (Figure 3B and C); thus despite the differences
in stability/yields of the 1-M complexes, their reactivities are
completely independent of the Lewis-acidity of M. Product
analysis of the reaction solution showed the formation of N-
(p-toluenesulfonyl)iminotriphenylphosphorane (Ph3P = NTs)
product in each case.
Figure 2. rRaman spectra [laser excitation: 514 nm; solvent: CH2Cl2/
acetone (20:1); temperature: –90 °C] of 1-Y-14N (black solid trace),
and 1-Y-15N (dotted trace). Bands originating from the solvents are
marked by asterisks.
Figure 3. A: Changes in the absorption spectra associated with the reaction of 1-Y at –90 °C in CH2Cl2 with 50 equiv. of PPh3. Inset: Time
trace of the pseudo-first order decay of the absorption band at 530 nm as a function of time in presence of 25, 50, 75, and 100 equiv. of PPh3.
B-C: Linear dependence of kobs on the concentration of PPh3 for each M used.
Z. Anorg. Allg. Chem. 2015, 78–82
80
© 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Journal of Inorganic and General Chemistry
www.zaac.wiley-vch.de
ARTICLE
Zeitschrift für anorganische und allgemeine Chemie
The formation of the purple intermediate was monitored by UV/Vis
spectroscopy by following the increasing absorbance at 530 nm.
Conclusions
We have found that in the stabilization of the copper-tosylni-
rRaman: rRaman spectra were measured at –92 °C (Bruker cryostat)
with 514 nm excitation, by using a Horiba Jobin-Yvon LabRAM
HR800 confocal Raman spectrometer. An 8 mm solution of
[CuAN](BF4) was used to generate the 1-Y intermediate both with
sPhINTs and with 15N-sPhINTs.
trene intermediate via its interaction with an external redox
non-innocent metal ion, the nature of the metal ions employed
affects the yield of the formal CuIII species; a higher yield is
obtained when a stronger Lewis acid is used. This behavior is
consistent with the proposed model in which the binding of
the external Lewis acids helps to reduce the hydrogen atom
abstraction ability of the transient copper-nitrene intermediate,
thereby preventing its spontaneous decay to the copper(II)-
amide species 3. Interestingly, however, the electronic struc-
tures of the 1-M complexes are found to be independent on
the nature of the Lewis-acidic metal ions. Similarly, the metal
ions proved to have no influence on the reactivity of the 1-M
complexes since near-identical rates were obtained for PPh3
when different Lewis acids were employed. This may indicate
the presence of a common reactive intermediate in the
reactions of 1-M complexes with PPh3. A plausible candidate
for the reactive intermediate can be the metal free
[(AN)Cu(NTs)]+ species, which may be generated in solution
in presence of excess PPh3 owing to the strong binding of
PPh3 to M (Scheme 2). Alternatively, the metal-independent
spectroscopic properties of the Lewis-acid bound copper-ni-
trene complexes may point to a different binding mode of M
than originally proposed[21] in our earlier communication; a
tautomeric structure B of 1-M (Scheme 1) can be visualized,
where M is bound to an amidate nitrogen of the ancillary li-
gand instead of the previously proposed binding to the nitrene
nitrogen (structure A in Scheme 1). Detailed spectroscopic, re-
activity, and theoretical studies are ongoing in our laboratory
in order to obtain further insights into the structure and reactiv-
ity of the Lewis-acid bound copper-nitrene intermediates.
EPR: X-band EPR derivative spectra were recorded with a Bruker
ELEXSYS E500 spectrometer equipped with the Bruker dual-mode
cavity (ER4116DM) and a Helium flow cryostat (Oxford Instrument
ESR 910). Microwave frequencies were calibrated with a Hewlett-
Packard frequency counter (HP5352B), and field control was cal-
ibrated with a Bruker NMR field probe (ER035M).
Quantification of the Formal CuIII present in Solution: 1-M inter-
mediates were generated in situ from 2 mL of a 0.35 mm solution of
[CuAN](BF4) at –90 °C in CH2Cl2. When the 530 nm band reached
its maximum, 0.10 mL of a solution containing 40 equiv. of ferrocene
were added. The resulting amount of ferrocenium produced was deter-
mined using the intensity of the band at 620 nm and its extinction
coefficient. The ferrocenium concentrations were then converted into
the concentrations of CuIII by using the stoichiometric coefficients of
the reaction.
Reactivity Evaluation: To 2 mL of the in situ generated intermediate
in CH2Cl2 at –90 °C, different amounts of PPh3 in CH2Cl2 were added,
always in a large excess to assure pseudo-first order conditions. The
decrease of the absorbance at 530 nm was monitored by taking one
UV/Vis spectrum every 0.5 s. The obtained trace was fitted to a
pseudo-first order curve to obtain a series of kobs that were found to
be linearly dependent on PPh3 concentration. The slope of the plot kobs
vs. PPh3 concentration provided the second-order rate constant k2. This
was obtained for each of the 1-M complexes. Ph3P=NTs product was
identified by 1H NMR spectroscopy; quantification was made by using
2 μL of nitromethane as an internal standard.
Acknowledgements
We gratefully acknowledge financial support of this work from the
Cluster of Excellence “Unifying Concepts in Catalysis” (EXC 314/1),
Berlin. IMP thanks BIG-NSE for a scholarship. Dr. U. Kuhlmann and
Prof. Dr. P. Hildebrandt are acknowledged for help with the rRaman
measurements.
Scheme 2. Proposed generation of the metal free [Cu(NTs)]+ as the
active intermediate in the reactions of 1-M with excess PPh3.
References
[1] H. Kwart, A. A. Kahn, J. Am. Chem. Soc. 1967, 89, 1950–1951.
[2] P. Müller, C. Fruit, Chem. Rev. 2003, 103, 2905–2920.
[3] Y. M. Badiei, A. Dinescu, X. Dai, R. M. Palomino, F. W. Heine-
mann, T. R. Cundari, T. H. Warren, Angew. Chem. Int. Ed. 2008,
47, 9961–9964.
[4] M. R. Fructos, S. Trofimenko, M. M. Díaz-Requejo, P. J. Pérez,
J. Am. Chem. Soc. 2006, 128, 11784–11791.
[5] D. N. Barman, P. Liu, K. N. Houk, K. M. Nicholas, Organometal-
lics 2010, 29, 3404–3412.
[6] T. R. Cundari, A. Dinescu, A. B. Kazi, Inorg. Chem. 2008, 47,
Experimental Section
All the chemicals used were purchased and used directly.
s
[CuAN](BF4), PhINTs and 15N-sPhINTs were synthesized following
previous reports.[33,34] Preparation and handling of air sensitive sam-
ples were performed in a nitrogen glove box OMNI-Lab 2 (VAC) with
O2 and moisture concentrations less than 1 ppm. UV/Vis spectra were
recorded with an Agilent 8453 diode array spectrometer connected
with a cryostat from Unisoku Scientific Instruments, Japan.
10067–10072.
[7] S. M. Tekarli, T. G. Williams, T. R. Cundari, J. Chem. Theory
Comput. 2009, 5, 2959–2966.
[8] M. J. B. Aguila, Y. M. Badiei, T. H. Warren, J. Am. Chem. Soc.
2013, 135, 9399–9406.
[9] Y. M. Badiei, A. Krishnaswamy, M. M. Melzer, T. H. Warren, J.
Am. Chem. Soc. 2006, 128, 15056–15057.
Generation of 1-M: 1-M intermediates were generated in situ in an
inert atmosphere at –90 °C from a CH2Cl2 solution of [CuAN](BF4)
after the addition of 1.5 equiv. of Lewis acid [Sc(OTf)3, Y(OTf)3,
Eu(OTf)3, Ce(OTF)3 Zn(OTf)2] and 1.5 equiv. of the oxidant sPhINTs.
Z. Anorg. Allg. Chem. 2015, 78–82
81
© 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Journal of Inorganic and General Chemistry
www.zaac.wiley-vch.de
ARTICLE
Zeitschrift für anorganische und allgemeine Chemie
[10] L. Maestre, W. M. C. Sameera, M. M. Díaz-Requejo, F. Maseras, [25] H. Dau, C. Limberg, T. Reier, M. Risch, S. Roggan, P. Strasser,
ChemCatChem. 2010, 2, 724–761.
P. J. Pérez, J. Am. Chem. Soc. 2013, 135, 1338–1348.
[11] P. Brandt, M. J. Södergren, P. G. Andersson, P.-O. Norrby, J. Am.
Chem. Soc. 2000, 122, 8013–8020.
[12] P. Comba, C. Lang, C. L. de Laorden, A. Muruganantham, G. Ra-
jaraman, H. Wadepohl, M. Zajaczkowski, Chem. Eur. J. 2008, 14,
5313–5328.
[13] R. T. Gephart, T. H. Warren, Organometallics 2012, 31, 7728–
7752.
[14] A copper bis nitrene complex has been structurally characterized
but no reactivity study is reported: F. Dielmann, D. M. Andrada,
G. Frenking, G. Bertrand, J. Am. Chem. Soc. 2014, 136, 3800–
3802.
[15] M. M. Díaz-Requejo, P. J. Pérez, M. Brookhart, J. L. Templeton,
Organometallics 1997, 16, 4399–4402.
[16] P. Brandt, M. J. Södergren, P. G. Andersson, P.-O. Norrby, J. Am.
Chem. Soc. 2000, 122, 8013–8020.
[17] P. Müller, C. Baud, Y. Jacquier, Can. J. Chem. 1998, 76, 738–
750.
[18] K. M. Gillespie, E. J. Crust, R. J. Deeth, P. Scott, Chem. Commun.
[26] S. Hong, F. F. Pfaff, E. Kwon, Y. Wang, M.-S. Seo, E. Bill, K.
Ray, W. Nam, Angew. Chem. Int. Ed. 2014, DOI:10.1002/
anie.201405874.
[27] F. F. Pfaff, S. Kundu, M. Risch, S. Pandian, F. Heims, I. Pryjom-
ska-Ray, P. Haack, R. Metzinger, E. Bill, H. Dau, P. Comba, K.
Ray, Angew. Chem. Int. Ed. Engl. 2011, 50, 1711–1715.
[28] S.-L. Abram, I. Monte-Pérez, F. F. Pfaff, E. R. Farquhar, K. Ray,
Chem. Commun. 2014, 50, 9852–9854.
[29] D. Usharani, D. Janardanan, C. Li, S. Shaik, Acc. Chem. Res.
2012, 46, 471–482.
[30] D. Schröder, M. C. Holthausen, H. Schwarz, J. Phys. Chem. B
2004, 108, 14407–14416.
[31] The yield of 1-M was determined on the basis of their reaction
with ferrocene to generate a CuI species and two ferrocenium
cations (Fc+) with
a signature-UV/Vis absorption (ε =
505 m–1·cm–1 at 620 nm in CH2Cl2), which allows the quantifica-
tion of the amount of formal CuIII present in solution. The yields
of 3 were obtained by spin quantification of the CuII EPR signal
corresponding to 3. For spin quantification, an EPR spectrum was
integrated twice to obtain the area under the signal. A strong pitch
(0.11% pitch in KCl) external standard was used to quantify the
S = 1/2 signals.
2001, 785–786.
[19] P. Comba, M. Kerscher, W. Schiek, Prog. Inorg. Chem. 2007, 55,
613.
[20] P. Comba, C. Lopez de Laorden, H. Pritzkow, Helv. Chim. Acta
[32] S. Fukuzumi, K. Ohkubo, J. Am. Chem. Soc. 2002, 124, 10270–
2005, 88, 647–664.
10271.
[21] S. Kundu, E. Miceli, E. Farquhar, F. F. Pfaff, U. Kuhlmann, P.
Hildebrandt, B. Braun, C. Greco, K. Ray, J. Am. Chem. Soc. 2012,
134, 14710–14713.
[22] M. Yagi, M. Kaneko, Chem. Rev. 2000, 101, 21–36.
[23] J. P. McEvoy, G. W. Brudvig, Chem. Rev. 2006, 106, 4455–4483.
[24] J. Yano, J. Kern, K. Sauer, M. J. Latimer, Y. Pushkar, J. Biesiadka,
B. Loll, W. Saenger, J. Messinger, A. Zouni, V. K. Yachandra,
Science 2006, 314, 821–825.
[33] H.-C. Liang, C. X. Zhang, M. J. Henson, R. D. Sommer, K. R.
Hatwell, S. Kaderli, A. D. Zuberbühler, A. L. Rheingold, E. I.
Solomon, K. D. Karlin, J. Am. Chem. Soc. 2002, 124, 4170–4171.
[34] D. Macikenas, E. Skrzypczak-Jankun, J. D. Protasiewicz, J. Am.
Chem. Soc. 1999, 121, 7164–7165.
Received: August 12, 2014
Published Online: November 11, 2014
Z. Anorg. Allg. Chem. 2015, 78–82
82
© 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim