Metal-metal distances at the limit: A coordination compound with an ultrashort chromium-chromium bond

Article abstract of DOI:10.1002/anie.200801160

(Chemical Equation Presented) Shorter than ever: Reduction of chromium(II/III) chlorides stabilized by amidopyridinato ligands leads to a dichromium complex that has a Cr-Cr bond length of 1.75 A. This bond is exceptionally short, even for bonding orders

Full text of DOI:10.1002/anie.200801160

Communications
DOI: 10.1002/anie.200801160
Metal–Metal Bonds
Metal–Metal Distances at the Limit: A Coordination Compound with
an Ultrashort Chromium–Chromium Bond**
Awal Noor, Frank R. Wagner,* and Rhett Kempe*
In memory of Franz Hein
The nature of the chemical bond is of fundamental impor-
tance, and has always fascinated scientists.[1] Metal–metal
bonds are of particular interest, as bond orders greater than
four are known[2,3] and are of considerable current interest.[4]
The quest for the shortest metal–metal bond is strongly linked
with the element chromium[2,5] and has very recently been
reinitiated after the first observation of a bond order greater
than four for this metal in a stable compound.[3] Soon
afterwards, the shortest metal–metal bond with a chro-
mium–chromium distance of 1.80 Š was observed in a dimeric
Scheme 1. Shortening of the metal–metal bond by high bond order,
bridging coordination of anionic ligands of type XYZ, and minimizing
the additional metal–ligand interactions by steric shielding.
chromium complex with such a high bond order.[6] Detailed
studies on ArCrCrAr complexes (Ar= aryl) performed at the
same time showed that such small values can be obtained for
this class of compounds as well.[7] Some years ago, we started
working with aminopyridinato complexes of chromium[8] and
herein report the synthesis and the (electronic) structure of a
bimetallic CrI2 complex with a drastically shortened metal–
metal distance. The very short metal–metal bond of only
1.75 Š results from a combination of Powerꢀs concept for the
stabilization of bond orders higher than four,[3,7] Hein–
Cottonꢀs principles on the realization of extremely short
metal–metal bonds with bridging anionic ligands of type
XYZ,[2,9] and a minimization of additional metal–ligand
interactions by optimal steric shielding (Scheme 1).
The deprotonation of 1 with potassiumhydride leads to
potassium [6-(2,4,6-triisopropylphenyl)pyridin-2-yl](2,4,6-tri-
methylphenyl)amide, which readily reacts with [CrCl3(thf)3]
affording complex 2 (Scheme 2). Compound 2 can be isolated
1
as a green crystalline material in good yield. In the H NMR
spectrum, only broad signals can be observed, and magnetic
susceptibility experiments show
a
magnetic moment
meff(300 K) = 3.2 mB. When 1 is deprotonated with BuLi and
allowed to react with CrCl2 in THF, the CrII complex 3 is
2
obtained in good yield as a green crystalline material after
removal of the solvent and subsequent extraction with
[*] Dr. F. R. Wagner
Scheme 2. Synthesis of 2, 3, and 4 (TIP=2,4,6-triisopropylphenyl,
Mes=2,4,6-trimethylphenyl).
Max-Planck-Institut für Chemische Physik fester Stoffe
01187 Dresden (Germany)
Fax: (+49)351-4646-3002
E-mail: Wagner@cpfs.mpg.de
toluene. The molecular structure of
3 is shown in
A. Noor, Prof. Dr. R. Kempe
Lehrstuhl Anorganische Chemie II, Universität Bayreuth
95440 Bayreuth (Germany)
Fax: (+49)921-55-2157
E-mail: Kempe@Uni-Bayreuth.de
Figure 1.[10] Compound 3 is the first CrII complex in which
the deprotonated aminopyridine has a strained bidentate
coordination mode and does not act as a bridging ligand.[11]
The chromium–nitrogen bond lengths clearly distinguish this
compound as an amidopyridine; i.e., the anionic function of
the ligand is localized on the Namido atom (N2).[12] Reduction
of 4 with potassium graphite (KC8) in THF, followed by
[**] We thank Germund Glatz for the single crystal X-ray analyses.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.200801160.
7246
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 7246 –7249

Angewandte
Chemie
In complex 4, the Cr–Namido bond lengths are very short
(1.998(4) Š) and clearly lie below the average value for this
bond (2.050 Š, dimeric chromium complexes with deproton-
ated aminopyridines as bridging ligands[11]), and are shorter
than the shortest reported bond of this kind (2.019 Š).[11a]
A
similar situation is observed for the Cr–Npyridine bond lengths
(2.028(4) Š for 4, average: 2.062 Š,[11] minimum: 2.023Š [11e]).
These values strongly indicate a stable metal–ligand bond. A
weak coordination of the amido ligand, which would then
cause a maximal approach of the central atoms, cannot be an
explanation for the exceptionally short metal–metal distance
in 4. However, a possible explanation could be the spatial
proximity of both N-donor functions in the ligands of type
XYZ (Scheme 1).
The electronic structure of 4 was studied in position space
at the DFT level[16] by means of the topological analysis of the
calculated electron density and the electron localizability
indicator (ELI-D),[17] and by calculation of the delocalization
index.[18] The calculations were performed on different
structural models[19] of 4. Below we explicitly discuss only
the results for model 4’a (structure with terminal H atoms:
see Figure 3).
Figure 1. ORTEP of 3. Ellipsoids for non-carbon atoms are set at 50%
probability; hydrogen atoms and two toluene molecules per complex
omitted for clarity. Selected bond lengths [Š] and angles [8]: N1–Cr1
2.1041(15), N2–Cr1 2.0411(14), Cl1–Cr1 2.3773(5), Cl1’–Cr1 2.6219(5),
Cr1–O1 2.0869(13); N2-Cr1-N1 64.77(6), O1-Cr1-N1 98.74(5).
removal of the solvent and extraction with toluene, affords 5
as a red crystalline material in 21% yield (Scheme 2). The
reduction of 3 with potassium graphite also leads to 4 (15%
1
yield). In the H NMR spectrum of the diamagnetic com-
In analogy to the model calculations for the two already
reported binuclear Cr2 complexes with a formal quintuple
bond,[3,6] in the present case, a (sg)2(pu)4(dg)4 configuration of
the chromium-based MOs is also obtained. These MOs can be
found in all the models within the seven highest occupied
pound 4, only one set of signals is observed at room
temperature.
The X-ray crystal structure analysis of 4 shows it to be a
bridged bimetallic complex, with an exceptionally short
metal–metal distance of 1.749(2) Š (Figure 2).[13] The hitherto
Figure 2. ORTEP of 4. Ellipsoids for non-carbon atoms are set at 50%
probability; hydrogen atoms omitted for clarity. Selected bond lengths
[Š] and angles [8]: Cr1–Cr1’ 1.749(2), Cr1–N2 1.998(4), Cr1–N1
2.028(4); Cr1’-Cr1-N2 98.55(13), Cr1’-Cr1-N1 96.78(13), N2-Cr1-N1
164.64(16).
shortest metal–metal bond known (1.8028(9) Š) also belongs
to a bimetallic chromium complex stabilized by an N-ligand.
The electronic structure of that complex was calculated, and it
has an effective bond order of 4.3.[6] For nearly 30 years an
aryl chromium(II) compound, structurally characterized by
Cotton and Koch[14] and first prepared by F. Hein and Tille
more than 40 years ago,[9a] claimed to have the shortest
experimentally obtained metal–metal distance of 1.830(4) Š.
The chromium–chromium bond length of low-temperature
laser-evaporated Cr2 in the gas phase is about 1.68 Š.[15]
Figure 3. a) Isosurface diagram of the pELI-D contributions of sd
(purple), pd (yellow and orange), and dd orbitals (green und dark
blue); The section planes show the sum of the five pELI-D contribu-
tions. b) Section plane showing (total) ELI-D; the dark blue isosurface
of ELI-D (value: 1.44) shows the structuring of the third chromium
shell; the semitransparent surface shows the QTAIM basin of one of
the chromium atoms.
Angew. Chem. Int. Ed. 2008, 47, 7246 –7249ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
7247

Communications
orbitals. The two dg MOs are always the HOMO and
chromium–chromium bond corresponds to the delocalization
index d(WCr1,WCr2) between the touching density basins
(QTAIM method)[23] WCr1 and WCr2 of the chromium atoms.
The density basin of one chromium atom is depicted in
Figure 3b. It includes the complete third shell and cuts the
two ELI-D bonding basins in the middle between the two
HOMOÀ1, and the two pu MOs HOMOÀ5 and HOMOÀ6.
Energetically, the sg MO and two ligand-centered MOs with a
strong Namido contribution can be found between these two
groups. The HOMO–LUMO gap lies between 1.5 eV and
1.7 eV, depending on the corresponding model.
For further analysis, the electron density and ELI-D were
calculated in position space.[20a] As recently shown, ELI-D can
be split in a physically transparent way into additive positive
orbital contributions (pELI-D contributions),[17e] where the
corresponding orbital density is simply multiplied by a
position-dependent weighting function (the so-called pair–
volume function). An illustration of the pELI-D contribu-
tions for the five chromium-centered MOs is given in
Figure 3a. The isosurfaces encompass the regions where the
corresponding MOs have the highest localizability contribu-
tions (pELI-D contributions) to the total ELI-D distribution.
From Figure 3a, it can be seen that the sg MO, the two pu
MOs, and the one dg MO have pELI-D maxima in the region
between the two chromium atoms. The remaining dg MO
(HOMO; 2d in Figure 3a) has a pELI-D topology with four
maxima at each atom (according to the shape of the dd
orbital), and not in the interatomic region, which closely
resembles a situation with two separated chromium atoms.
Interestingly, this behavior is observed for only one of the two
dg MOs. The other MO shows a strong mixing of Cr(4s)
contributions, which largely eliminates the dd contributions in
the direction of the ligand and reinforces those in the
perpendicular direction. This results in the formation of a
pELI-D maximum perpendicular to the molecular plane and
relatively far from the bond axis. The sum of these five pELI-
D contributions yields the pELI-D distribution shown in
Figure 3a. It shows the topological points for the chromium–
chromium bonding situation displayed by total (all-electron)
ELI-D (Figure 3b). These are the two ELI-D maxima which
are perpendicular to the molecular plane, resulting from the
sum of a pu and dg pELI-D orbital contribution (so-called
banana bonds), and two axially situated maxima, which are
not found for the less-simplified model 4a, on the bond-
opposed side of the chromium atoms. Furthermore, a
significant structuring of the chromium third atomic shell
signals the participation of the d orbitals in the bond
formation. The electronic population of the two bonding
chromium atoms. In the present case, a value for d(WCr1,WCr2)
of 4.2 is found (2.4 of which results solely from the
delocalization between the two chromium third atomic
shells, see above), which significantly differs from the
formal bond order of 5.0. However, this finding is consistent
with the weakly bonding dd MO discussed above. Interest-
ingly, a very similar bond order of 4.3was obtained for a
similar compound using natural resonance theory analysis in
Hilbert space.[6] As the d bonds in the Cr2 model have only a
small contribution to the bond formation, and the 4s–4s bond
is energetically repulsive at the equilibrium distance,[4b] the
most important effect of these electrons for the short bond
distance in Cr2 (1.68 Š)[15] could be the avoidance of a positive
charge at the metal centers. This would then be the decisive
factor which has to be overcome to realize similar short
distances with formally fivefold-bonded metal atoms.
In future investigations, we are interested in minimizing
the metal–metal bond through variation of the ligand
environment and to explore the reactivity of the such
metal–metal bonds, and to describe in detail the bonding
situation for 4 in position space at an explicitly correlated
level of theory.
Received: March 10, 2008
Revised: May 13, 2008
Published online: August 13, 2008
Keywords: chemical bonds · chromium · electronic structure ·
.
multiple bonds · N ligands
[1] a) L. Pauling, Die Natur der chemischen Bindung, 3rd ed., VCH,
Weinheim, 1973; b) G. N. Lewis, J. Am. Chem. Soc. 1916, 38,
762 – 785; c) “90 Years of Chemical Bonding”: G. Frenking, S.
Shaik, J. Comput. Chem. 2007, 28, 1 – 3 .
[2] F. A. Cotton, L. A. Murillo, R. A. Walton, Multiple Bonds
Between Metal Atoms, 3rd ed., Springer, Berlin, 2005.
[3] a) T. Nguyen, A. D. Sutton, M. Brynda, J. C. Fettinger, G. J.
Long, P. P. Power, Science 2005, 310, 844 – 847; b) M. Brynda, L.
Gagliardi, P.-O. Widmark, P. P. Power, B. O. Roos, Angew. Chem.
2006, 118, 3888 – 3891; Angew. Chem. Int. Ed. 2006, 45, 3804 –
3807; c) G. Frenking, Science 2005, 310, 796 – 797.
[4] a) L. Gagliardi, B. O. Ross, Nature 2005, 433, 848 – 851; b) B. O.
Roos, A. C. Borin, L. Gagliargi, Angew. Chem. 2007, 119, 1491 –
1494; Angew. Chem. Int. Ed. 2007, 46, 1469 – 1472; c) G.
Frenking, R. Tonner, Nature 2007, 446, 276 – 277; d) G. Merino,
K. J. Donald, J. S. DꢀAcchioli, R. Hoffmann, J. Am. Chem. Soc.
2007, 129, 15295 – 15302.
basins in the valence region amounts to 1.8 eÀ (banana bond)
[20b]
in total, and for the two bond-opposed basins only 0.3e À
.
The electronic population of the chromium third atomic shell,
having a total of 11.8 eÀ, exceeds the value corresponding to a
3s2p6 configuration by 3.8 eÀ. Therefore, the electrons for the
chromium–chromium bonding interaction are not only local-
ized in the valence region, but can also largely be found in the
spatial region of the third shell of the chromium atoms, where
they contribute to the above-mentioned structuring of ELI-D.
The former statement can be verified by calculation of the
delocalization index[18] d(A,B) between the third shells of
both chromium atoms. A relatively high value of 2.4 for the
corresponding delocalization index is found,[21] which repre-
sents an indirect contribution in the calculation of the bond
order in position space according to ngyµn, Loos, and
Mayer.[22] Thus, the bond order in the case of a symmetrical
[5] U. Radius, F. Breher, Angew. Chem. 2006, 118, 3072 – 3077;
Angew. Chem. Int. Ed. 2006, 45, 3006 – 3010.
[6] K. A. Kreisel, G. P. A. Yap, O. Dmitrenko, C. R. Landis, K. H.
Theopold, J. Am. Chem. Soc. 2007, 129, 14162 – 14163.
[7] R. Wolf, Ch. Ni, T. Nguyen, M. Brynda, G. J. Long, A. D. Sutton,
R. C. Fischer, J. C. Fettinger, M. Hellman, L. Pu, P. P. Power,
Inorg. Chem. 2007, 46, 11277 – 11290.
7248 www.angewandte.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 7246 –7249

Angewandte
Chemie
[8] A. Spannenberg, A. Tillack, P. Arndt, R. Kirmse, R. Kempe,
Polyhedron 1998, 17, 845 – 850.
s) electron pair:
gD(r) = 1s(r) ꢁ VD(r)
s
˜
[9] a) F. Hein, D. Tille, Z. Anorg. Allg. Chem. 1964, 329, 72 – 82;
b) L. H. Gade, Koordinationschemie, Wiley-VCH, Weinheim,
1998.
The values for ELI-D are limited to the range of positive
numbers. Typically, ELI-D values up to 2.5 (except for H atoms)
can be found in the chemically relevant valence region for
molecules. ELI-D has been defined at explicitly correlated and
at uncorrelated quantum chemical level. At the Hartree–Fock
level, the ELI-D formula simplifies and strongly resembles the
inverse kernel of the electron localization function (ELF)
defined by Becke and Edgecombe: b) A. D. Becke, K. E.
Edgecombe, J. Chem. Phys. 1990, 92, 5397 – 5403. Nevertheless,
ELI-D should not be regarded as a generalization of the ELF.
Although it is in a certain sense related to ELF, the ELI-D
represents a separate quantity, which reflects one possible
interpretation of the Becke ELF kernel at the correlated level
of theory: c) M. Kohout, Int. J. Quantum Chem. 2004, 97, 651 –
658; d) M. Kohout, K. Pernal, F. R. Wagner, Yu. Grin, Theor.
Chem. Acc. 2004, 112, 453– 459; e) M. Kohout, F. R. Wagner,
Yu. Grin, Int. J. Quantum Chem. 2006, 106, 1499 – 1507; f) M.
Kohout, Faraday Discuss. 2007, 135, 43– 54; g) F. R. Wagner, V.
Bezugly, M. Kohout, Yu. Grin, Chem. Eur. J. 2007, 13, 5724 –
5741; h) M. Kohout, F. R. Wagner, Yu. Grin, Theor. Chem. Acc.
2008, 119, 413– 420.
¯
[10] P1, a = 12.3740(7), b = 12.9060(8), c = 15.1170(10) Š; a =
71.269(5), b = 75.700(5), g = 68.868(5)8 and R1 = 0.0375 (I >
2s(I)); wR2 = 0.0900 (all data).
[11] Examples of dimeric CrII complexes in which deprotonated
2
aminopyridines act as bridging ligands: a) F. A. Cotton, R. H.
Niswander, J. C. Sekutowski, Inorg. Chem. 1978, 17, 3541 – 3545;
b) J. J. H. Edema, S. Gambarotta, A. Meetsma, A. L. Spek,
W. J. J. Smeets, M. Y. Chiang, J. Chem. Soc. Dalton Trans. 1993,
789 – 797; c) F. A. Cotton, L. M. Daniels, C. A. Murillo, I.
Pascual, Inorg. Chem. Commun. 1998, 1, 1 – 3; d) F. A. Cotton,
L. M. Daniels, C. A. Murillo, I. Pascual, H.-C. Zhou, J. Am.
Chem. Soc. 1999, 121, 6856 – 6861; e) F. A. Cotton, L. M.
Daniels, P. Lei, C. A. Murillo, X. Wang, Inorg. Chem. 2001, 40,
2778 – 2784; f) R. Clerac, F. A. Cotton, S. P. Jeffery, C. A.
Murillo, X. Wang, Dalton Trans. 2003, 3022 – 3027.
[12] S. Deeken, G. Motz, R. Kempe, Z. Anorg. Allg. Chem. 2007, 633,
320 – 325.
¯
[13] P1, a = 9.2100(12), b = 12.2290(14), c = 12.6890(15) Š; a =
94.314(9), b = 105.714(10), g = 108.185(10)8 and R1 = 0.0570
[18] The delocalization index d(A,B) is a quantitative measure for
the sharing of electrons between two non-overlapping regions A
and B in position space: a) X. Fradera, M. A. Austen, R. F. W.
Bader, J. Phys. Chem. A 1999, 103, 304 – 314; b) R. F. W. Bader,
M. E. Stephans, J. Am. Chem. Soc. 1975, 97, 7391 – 7399.
[19] Model 4a was obtained by the exclusive relaxation of the H-
atom positions based on the structural data of 4. On the basis of
4a in model 4b the Cr–Cr distance was relaxed, which led to a
shortening of the bond to 1.69 Š. A radical simplification of the
model by substituting all substituents on the deprotonated
aminopyridine with H leads to model 4’a, for which only the
newly added H-atom positions were relaxed. Therefore, the Cr,
N, and C atom positions of molecule 4’a shown in Figure 3are
identical with those for model 4a and of the experimental
structure 4. Starting from 4’a, the complex was completely
relaxed (model 4’b), which led to a planar molecule (C2h
symmetry) with a Cr–Cr distance of 1.68 Š. The consistent
shortening of the Cr–Cr distance obtained by the present
DFT(BLYP) calculations in complex 4b with the original
ligand and in complex 4’b with the model ligand is caused by
an incomplete treatment of the electron correlation in these
calculations. It shows that the experimentally found short Cr–Cr
distance is not caused by packing effects. For the coordinates of
the models, see the Supporting Information.
(I > 2s(I)); wR2 = 0.1226 (all data).
[14] F. A. Cotton, S. A. Koch, Inorg. Chem. 1978, 17, 2021 – 2024.
[15] V. E. Bondybey, J. H. English, Chem. Phys. Lett. 1983, 94, 443–
447.
[16] The calculations of the electronic structure were carried out at
the DFT level with the program ADF: a) G. te Velde, F. M.
Bickelhaupt, S. J. A. van Gisbergen, C. Fonseca Guerra, E. J.
Baerends, J. G. Snijders, T. Ziegler, J. Comput. Chem. 2001, 22,
931 – 967; b) C. Fonseca Guerra, J. G. Snijders, G. te Velde, E. J.
Baerends, Theor. Chem. Acc. 1998, 99, 391 – 403;
c) ADF2006.01, SCM, Theoretical Chemistry, Vrije Universiteit,
Amsterdam, The Netherlands, http://www.scm.com. The BLYP
functional was used: d) A. D. Becke, Phys. Rev. A 1988, 38,
3098 – 3100; e) C. Lee, W. Yang, R. G. Parr, Phys. Rev. B 1988,
37, 785 – 789. In the case of single-point energy calculations, all
electrons were explicitly included using a TZ2P basis set (triple
z + 2 sets of polarization functions) for each type of atom. The
structure optimizations[19] were performed within the frozen-
core approximation (C,N: 1s2 core; Cr: 1s22s2p6 core) using the
corresponding TZP or TZ2P basis set. Calculation of the
delocalization index was performed on the basis of a DFT
(BLYP functional) single-point calculation with the Gaussian03
program system: f) M. J. Frisch et al. (entire citation: see
Supporting Information), Gaussian 03, Revision C.02, Gaussian,
Inc., Wallingford, CT, 2004. Herein, the corresponding TZVP
basis set was used: g) A. Schaefer, H. Horn, R. Ahlrichs, J.
Chem. Phys. 1992, 97, 2571 – 2577; h) A. Schaefer, H. Horn, R.
Ahlrichs, J. Chem. Phys. 1994, 100, 5829 – 5835.
[20] a) M. Kohout, program DGrid, version 4.3, Dresden 2008. b) M.
Kohout, program Basin, version 4.2, Dresden, 2007.
[21] The delocalization index was calculated with the ToPMoD
program on the basis of a single-point calculation performed
with Gaussian 03:[16f] S. Noury, X. Krokidis, F. Fuster, B. Silvi,
Program ToPMoD, Universite Pierre et Marie Curie, Paris, 2008.
[22] a) J. G. ngyµn, M. Loos, I. Mayer, J. Phys. Chem. 1994, 98,
5244 – 5248; b) X. Fradera, J. Poater, S. Simon, M. Duran, M.
Solꢂ, Theor. Chem. Acc. 2002, 108, 214 – 224.
[17] a) The ELI-D is a scalar field which is defined in position space
and momentum space. Roughly, ELI-D (gsD) in position space
can be regarded as a quasi-continuous weighted electronic
charge distribution 1s (the s-spin electron density), where the
˜
weighting factor VD (the so-called pair–volume function) is a
[23] R. F. W. Bader, Atoms in Molecules: AQuantum Theory, Oxford
University Press, Oxford, 1994.
local measure for the volume needed to build a same-spin (spin
Angew. Chem. Int. Ed. 2008, 47, 7246 –7249ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 7249