Unusual B4N2C2 ligand in a ruthenium pseudo-triple-decker sandwich complex displaying three reversible electron-transfer steps

Article abstract of DOI:10.1002/anie.200703556

(Graph Presented) Open, sesame: The reaction of a heterobicyclic pentalenediyl-like Me₂Ph₄B₄N₂C ₂ dianion with [{(C₅Me₅)RuCl}₄] cleaves the N-N bond of the ligand and affords a pseudo-triple-decker sandwich complex containing a B₄N₂C₂ middle deck (see picture). This eight-membered ring features nearly linear B-N-B moieties and brings the ruthenium centers unusually close. Cyclic voltammetry indicates efficient electron derealization over the framework.

Full text of DOI:10.1002/anie.200703556

Angewandte
Chemie
DOI: 10.1002/anie.200703556
Sandwich Complexes
Unusual B4N2C2 Ligand in a Ruthenium Pseudo-Triple-Decker
Sandwich ComplexDisplaying Three Reversible Electron-Transfer
Steps**
Hanh V. Ly, Heikki M. Tuononen, Masood Parvez, and Roland Roesler*
Dedicated to Professor Richard J. Puddephatt
Polydecker sandwich complexes are highly desirable synthetic
targets because of their potential applications as advanced
materials, owing to the considerable interactions between
their metal centers.[1] Cyclopentadienyl[2] and benzene[3]
ligands were demonstrated early on to coordinate bifacially
to transition metals and thus generate triple-decker sandwich
complexes. Although larger oligomers have been identified in
the gas phase,[4] the energy of such bridging p interactions
involving neutral or singly charged, carbon-based ligands is
low, and oligomers with more than four decks have not been
isolated in the condensed phase. Exceptions are sandwich
compounds of the main-group metals, which form extended
structures in the solid state.[5] To circumvent the low tendency
of organic ligands to bridge metal centers and build stacked
sandwich compounds, annelated cyclopentadienyl derivatives
such as indacene[6] and especially pentalene[6a,7] have been
used as building blocks for staggered multidecker sandwich
compounds. Although promising, these systems have limita-
tions associated with the narrow choice of substitution
patterns available, resulting in low solubility of the higher
oligomers or in the formation of complex mixtures of isomers.
The recent synthesis of hexamethylpentalene is a promising
development in this area.[8] Heterocyclic rings, particularly
those containing boron, have a higher tendency to coordinate
bifacially and generate stacked sandwich compounds.[9] The
largest multidecker sandwich compounds known to date
contain six decks and feature dianionic cyclic ligands with
B2C3 and B3C2 frameworks.[10] The only family of polydecker
complexes reported so far, [{Ni(C3B2Me2RR’R’’)}x] displayed
conducting properties.[11]
Within the scope of our investigation on heterocyclic
cyclopentadienyl analogues, we developed the 1,2-diaza-3,5-
diborolyl derivatives 1. [12] Herein, we report the extension of
this project to the pentalenediyl-like dianion in 3 (Scheme 1),
which is a very promising bridging p ligand for the assembly
of polymeric materials.
Scheme 1. Synthesis of derivatives 2–4. KHMDS=K[N(SiMe3)2],
Cp*=C5Me5.
[*] H. V. Ly, Dr. M. Parvez, Prof. R. Roesler
Department of Chemistry
The precursor 2 was obtained as a mixture of cis and trans
isomers through condensation of hydrazine and 1,1-bis(phe-
nylchloroboryl)ethane[12b] in the presence of triethylamine
(see the Supporting Information). The trans isomer could be
separated by crystallization and was structurally character-
University of Calgary
2500 University Dr. NW, Calgary, AB, T2N 1N4 (Canada)
Fax: (+1)403-289-9488
E-mail: roesler@ucalgary.ca
Homepage:
ized (Figure 1).[13] It features a long N N bond of 1.48 Š,
À
[14]
http://www.chem.ucalgary.ca/research/groups/roesler/
À
slightly longer than the N N bonds in hydrazine (1.45 Š)
and the metal complexes of 1 (1.44–1.47 Š).[12] The endocyclic
Dr. H. M. Tuononen
Department of Chemistry
University of Jyväskylä
À
À
B C and B N bonds (average 1.57 and 1.44 Š) are longer and
slightly shorter, respectively, than the corresponding bonds in
the alkali-metal salts of the monocyclic analogues.[12b] For
P.O. Box 35, 40014 Jyväskylä (Finland)
[**] This work was supported by the Natural Sciences and Engineering
Research Council of Canada, the Canada Foundation for Innovation,
the Academy of Finland and the Alberta Science and Research
Investments Program.
À
comparison, the B N bonds in borazines measure 1.42–
1.44 Š.[15]
Double deprotonation of 2 with K[N(SiMe3)2] could be
conducted directly or in two successive steps, yielding the
orange dipotassium salt 3. As expected, both isomers
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
Angew. Chem. Int. Ed. 2008, 47, 361 –364
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
361

Communications
Figure 1. Molecular structure of trans-2. Thermal ellipsoids are set at
50 % probability. Selected bond lengths [Š] and angles [8]: N–N
1.480(2), B–N 1.436(2), 1.437(2), B–C(1) 1.571(2), 1.574(2); B-N-B
140.88(14), N-B-C(1) 108.89(13), 109.02(14), B-C(1)-B 101.86(13),
B-N-N 109.54(15), 109.58(14).
Figure 2. Molecular structure of 3(tmeda)2. Hydrogen atoms have
been omitted for clarity. Thermal ellipsoids are set at 50 % probability.
Selected bond lengths [Š] and angles [8]: N–N 1.442(4), B–N 1.474(4),
1.483(4), B–C(1) 1.494(4), 1.502(4), K–N 2.820(2), 2.889(2), K–B
3.268(3)–3.377(3); B-N-B 143.3(2), N-B-C(1) 108.5(2), 109.6(2), B-C-B
104.7(2), B-N-N 107.9(2), 108.8(2).
produced the same mono- and dianions. The loss of the
methine protons was noticeable in the H NMR spectrum of
3, and the deprotonation resulted in a d = 14 ppm upfield shift
of the 11B NMR signal. Despite its very similar geometry, the
8-p-electron heterocyclic ligand in 3 is not a direct analogue
of the 10-p-electron pentalenediyl dianion.[7,8]
1
The crystal structure of 3(tmeda)2 (tmeda = N,N,N’,N’-
tetramethylethylenediamine) was determined, revealing that
the planar B4N2C2 ligand is coordinated by two potassium
ions, each supporting a tmeda molecule, in a centrosymmetric
arrangement (Figure 2).[13] The transition from 2 to 3 involves
À
À
a shortening of the intraannular N N and B C bonds and a
À
slight lengthening of the B N bonds, resembling the tran-
sition from cyclopentadiene to the cyclopentadienyl anion.
The separation between the two potassium ions measures
5.52 Š and is comparable to the K···K separation observed in
polymeric cyclopentadienyl salts (5.52–5.85 Š)[12b] and the
dimeric potassium pentalenediyl derivative (5.43, 5.48 Š).[16]
Reaction of 3 with [{Cp*RuCl}4] yielded the pseudo-
triple-decker sandwich 4. The Cp* groups, as well as the boron
centers and the phenyl and methyl substituents of the bridging
p ligand are equivalent on the time scale of NMR spectros-
copy, and the corresponding chemical shifts are unremark-
able. The compound generated an intense signal correspond-
ing to the molecular ion in the mass spectrum, and its identity
was confirmed by a high-resolution mass spectrum as well as
by elemental analysis. An X-ray diffraction study on a single
crystal of 4 validated the pseudo-triple-decker sandwich
Figure 3. Molecular structure of 4. For clarity, only the ipso carbon
atoms of the phenyl substituents are represented, and all hydrogen
atoms have been omitted. Thermal ellipsoids are set at 50 % proba-
bility. Selected bond lengths [Š] and angles [8]: N(1)···N(2) 2.676(11),
B(1)–N(1) 1.437(9), B(3)–N(1) 1.440(9), B(2)–N(2) 1.400(9), B(4)–
N(2) 1.462(9), B(1)–C(1) 1.542(10), B(2)–C(1) 1.565(9), B(3)–C(2)
1.582(9), B(4)–C(2) 1.558(9), Ru(1)···Ru(2) 3.240(6), Ru–N, 2.118(5)–
32(6), Ru–B 2.351(7)–2.546(7), Ru–C 2.257(6)–2.548(6), B-N-B
166.4(6), 172.1(6), N-B-C 110.7(6)–114.2(5), B-C-B 123.8(6), 130.1(5).
structure and revealed
a
few surprising features
(Figure 3).[13] The N N bond cleaves upon the formation of
4, and the 8-p-electron bicyclic ligand transforms into a
B4N2C2 ring. Its geometry can be best described as a severely
elongated hexagon, and is, to our knowledge, unprecedented
for eight-membered rings. The structure is asymmetric, with
Ru(1) situated within bonding distance from all the B atoms
(2.488(7)–2.546(7) Š), while Ru(2) is clearly h5-coordinated
by N(1), B(3), C(2), B(4), and N(2). The coordination of
Ru(1) can be described as h7 or h8, depending on the
À
À
À
vs. Ru(1) C(1) 2.402(5) Š and Ru(2) C(2) 2.257(6) Š). The
metal atoms are situated close to the B4N2C2 ligand (1.58,
À
1.62 Š vs. Ru Cp* 1.82, 1.86 Š) and consequently very close
to each other (3.24 Š). For comparison, the separation
between the metals measures 3.90 Š in [(Cp*Ru)2B8H14].[17b]
Unfortunately, no complete structural data is available for the
À
involvement of C(2) in the bonding (Ru(1) C(2) 2.548(6) Š
362
www.angewandte.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 361 –364

Angewandte
Chemie
analogues containing all-carbon ligands, [(Cp*Ru)2C8H6][6a]
and transoid-[(CpRu)2C8H8] (Cp = C5H5).[17a]
with [{Cp*RuCl}4] prompted the cleavage of the N N bond
À
and yielded the pseudo-triple-decker sandwich complex 4,
featuring a highly unusual monocyclic bridging ligand with an
elongated hexagonal B4N2C2 skeleton. A computational
analysis showed that this ligand could be considered a
distorted form of the bicyclic ligand, brought about by
coordination to the transition metal. The cyclic voltammo-
gram of 4 displayed three reversible, well-separated one-
electron transfer steps indicating efficient electron delocali-
zation over the framework. Its ligand properties render the
highly unusual bridging p ligand a very promising building
block for the design of larger sandwich-like architectures.
A computational analysis at the DFT level (see the
Supporting Information) performed on the model structure
4a, which features a hydrogen-substituted B4N2C2H6 ligand
and Cp groups, gave a molecular geometry in reasonable
agreement with the X-ray diffraction data. The frontier
molecular orbital (MO) analysis of 4a shows that substantial
mixing of the ligand orbitals with the MOs of the {CpRu}+
fragments takes place, thus resulting in a formal insertion of
À
the metal atoms into the N N bond. The ruthenium d orbitals
interact mainly with the two nitrogen p orbitals that are
orientated parallel and perpendicular to the B4N2C2 plane.
The similar oxidative addition of hydrazines to transition
metals has recently been reported.[18] Energy decomposition
analysis conducted for 4a confirms that the primary, thermo-
dynamic driving force for its formation is the cleavage of the
Received: August 6, 2007
Revised: September 27, 2007
Published online: November 16, 2007
Keywords: boron · cyclic voltammetry · nitrogen · ruthenium ·
À
À
N N bond and the subsequent formation of the Ru N
interactions. The total orbital interaction between the for-
mally cationic and anionic fragments of the complex is strong,
with a calculated value of around À1900 kJmolÀ1 (total
calculated bond energy is À2400 kJmolÀ1). Mayer bond
orders obtained for 4a support the description of Ru(2) as
h5-coordinated, whereas coordination of Ru(1) can be best
described as h8; the largest bond orders (around 0.45) are
.
sandwich complexes
[1] a) R. N. Grimes, Appl. Organomet. Chem. 1996, 10, 209; b) N. S.
Hosmane, J. A. Maguire, Organometallics 2005, 24, 1356 – 1389.
[2] a) H. Werner, A. Salzer, Synth. Inorg. Met.-Org. Chem. 1972, 2,
239; b) A. Salzer, H. Werner, Angew. Chem. 1972, 84, 949 – 950;
Angew. Chem. Int. Ed. Engl. 1972, 11, 930 – 932; c) E. Dubler, M.
Textor, H.-R. Oswald, A. Salzer, Angew. Chem. 1974, 86, 125;
Angew. Chem. Int. Ed. Engl. 1974, 13, 135 – 136; d) H. Werner,
Angew. Chem. 1977, 89, 1 – 10; Angew. Chem. Int. Ed. Engl.
1977, 16, 1 – 9.
À
calculated for Ru N interactions. As expected, the bonding
À
analysis shows no sign of N N interactions. Hence, the
computational data indicate that the atypical linear B-N-B
moieties in 4 can be described with formally sp-hybridized
nitrogen atoms whose unhybridized p orbitals have an
important contribution to the ligand–metal bonding.
[3] A. W. Duff, K. Jonas, R. Goddard, H.-J. Kraus, C. Krüger, J. Am.
Chem. Soc. 1983, 105, 5479 – 5480.
Cyclic voltammetry of 4 in THF revealed two reversible
reduction steps at À1.30 and À1.99 V vs. standard calomel
electrode (SCE) and one reversible oxidation at 0.42 V. The
expansion of the electrochemical window using dimethoxy-
ethane (DME) confirmed this behavior and allowed for the
observation of an additional, irreversible oxidation at 1.10 V,
while the electrochemical range of CH2Cl2 enabled the
observation of only one reduction and one oxidation step at
À1.45 and 0.18 V, respectively. All one-electron transfer steps
are well-separated, indicating significant electron delocaliza-
tion over the framework. The reversible oxidation likely
involves the Ru2+/Ru3+ process, while the irreversible electron
transfer could be attributed to the oxidation of the Cp* ligand,
as observed for [Cp*2Ru] (0.55 and 1.25 V in CH2Cl2),[19]
[(Cp*Ru)2C8H6] (0.11 and 0.40 V in THF, À0.02 and 0.48 V
in CH2Cl2),[6a] and [(Cp*Ru)2B8H14] (0.17 and 0.93 V in 8.5:1.5
CH2Cl2/toluene; V versus SCE in all cases).[17b] The reduction
steps have no parallel in the chemistry of related species
containing all-carbon and all-boron ligands and are attributed
to processes centered mostly on the [B4N2C2]2À ligand. A
theoretical analysis supports this view, as the optimized
geometry of [4a]2À is C2h-symmetric and contains a planar
[4] a) S. Nagao, A. Kato, A. Nakajima, K. Kaya, J. Am. Chem. Soc.
2000, 122, 4221 – 4222; b) K. Miyajima, S. Yabushita, M. B.
Knickelbein, A. Nakajima, J. Am. Chem. Soc. 2007, 129, 8473 –
8480.
[5] a) S. Harder, Coord. Chem. Rev. 1998, 176, 17 – 66, and erratum,
Coord. Chem. Rev. 2000, 199, 331 – 334; b) P. Jutzi, N. Burford,
Chem. Rev. 1999, 99, 969 – 990; c) P. Jutzi, G. Reumann, J. Chem.
Soc. Dalton Trans. 2000, 2237 – 2244.
[6] a) J. M. Manríquez, M. D. Ward, W. M. Reiff, J. C. Calabrese,
N. L. Jones, P. J. Carroll, E. E. Bunel, J. S. Miller, J. Am. Chem.
Soc. 1995, 117, 6182 – 6193; b) D. R. Cary, C. G. Webster, M. J.
Drewitt, S. Barlow, J. C. Green, D. OꢀHare, Chem. Commun.
1997, 953 – 954; c) P. Roussel, M. J. Drewitt, D. R. Cary, C. G.
Webster, D. OꢀHare, Chem. Commun. 1998, 2205 – 2206.
[7] a) B. Oelckers, I. Chµvez, J. M. Manríquez, E. Romµn, Organo-
metallics 1993, 12, 3396 – 3397; b) O. T. Summerscales, F. G. N.
Cloke, Coord. Chem. Rev. 2006, 250, 1122 – 1140.
[8] A. E. Ashley, A. R. Cowley, D. OꢀHare, Chem. Commun. 2007,
1512 – 1514.
[9] M. C. Böhm, Z. Naturforsch. A 1984, 39, 223 – 246.
[10] a) T. Kuhlmann, W. Siebert, Z. Naturforsch. B 1985, 40, 167 –
170; b) X. Wang, M. Sabat, R. N. Grimes, J. Am. Chem. Soc.
1994, 116, 2687 – 2688; c) X. Wang, M. Sabat, R. N. Grimes, J.
Am. Chem. Soc. 1995, 117, 12218 – 12226; d) X. Wang, M. Sabat,
R. N. Grimes, J. Am. Chem. Soc. 1995, 117, 12227 – 12234.
[11] a) T. Kuhlmann, S. Roth, J. Rozi›re, W. Siebert, Angew. Chem.
1986, 98, 87 – 88; Angew. Chem. Int. Ed. Engl. 1986, 25, 105 – 107;
b) T. Kuhlmann, S. Roth, J. Rozi›re, W. Siebert, U. Zenneck,
Synth. Met. 1987, 19, 757 – 762; c) M. Y. Lavrentiev, H. Köppel,
M. C. Böhm, Chem. Phys. 1993, 169, 85 – 102; d) U. Fenner, T.
Kuhlmann, G. Brodt, J. Rozi›re, W. Siebert, Z. Naturforsch. B
1993, 48, 1512 – 1518.
À
B4N2C2 framework with no apparent N N interaction. A two-
electron oxidation was reported for the related transoid-
À
[(CpRu)2C8H8] complex, resulting in cleavage of a C C bond
in the bridging ligand and formation of a flyover dication.[17a]
In conclusion, a pentalenediyl-like dianion with B4N2C2
framework was isolated as a dipotassium salt, 3. Its reaction
Angew. Chem. Int. Ed. 2008, 47, 361 –364
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
363

Communications
[12] a) H. V. Ly, T. D. Forster, D. Maley, M. Parvez, R. Roesler,
cally. Hydrogen atoms were included at geometrically calculated
positions and refined using the riding model. CCDC-655685,
-655686, and -655687 (2, 3, and 4, respectively) contain the
supplementary crystallographic data for this paper. These data
can be obtained free of charge from The Cambridge Crystallo-
graphic Data Centre via www.ccdc.cam.ac.uk/data_request/cif.
[14] K. Kohata, T. Fukuyama, K. Kuchitsu, J. Phys. Chem. 1982, 86,
602 – 606.
Chem. Commun. 2005, 4468 – 4470; b) H. V. Ly, T. D. Forster,
A. M. Corrente, D. J. Eisler, J. Konu, M. Parvez, R. Roesler,
Organometallics 2007, 26, 1750 – 1756; c) H. V. Ly, T. D. Forster,
M. Parvez, R. McDonald, R. Roesler, Organometallics 2007, 26,
3516 – 3523.
[13] Crystallographic data: trans-2: C28H28B4N2, 173 K, 0.20 ” 0.20 ”
0.08 mm, monoclinic, space group P21/n, a = 14.1990(10), b =
6.0800(2),
c = 14.6470(11) Š,
b = 103.012(3)8,
V=
[15] a) R. Boese, A. H. Maulitz, P. Stellberg, Chem. Ber. 1994, 127,
1887 – 1889; b) C. A. Jaska, K. Temple, A. J. Lough, I. Manners,
J. Am. Chem. Soc. 2003, 125, 9424 – 9434; c) T. Jꢁschke, M.
Jansen, Z. Anorg. Allg. Chem. 2004, 630, 239 – 243.
[16] F. G. N. Cloke, M. C. Kuchta, R. M. Harker, P. B. Hitchcock, J. S.
Parry, Organometallics 2000, 19, 5795 – 5798.
[17] a) W. E. Geiger, A. Salzer, J. Edwin, W. von Philipsborn, U.
Piantini, A. L. Rheingold, J. Am. Chem. Soc. 1990, 112, 7113 –
7121; b) S. Ghosh, B. C. Noll, T. P. Fehlner, Angew. Chem. 2005,
117, 6726 – 6729; Angew. Chem. Int. Ed. 2005, 44, 6568 – 6571.
[18] J. M. Hoover, A. DiPasquale, J. M. Mayer, F. E. Michael,
Organometallics 2007, 26, 3297 – 3305.
[19] a) U. Koelle, A. Salzer, J. Organomet. Chem. 1983, 243, C27 –
C30; b) U. Kölle, J. Grub, J. Organomet. Chem. 1985, 289, 133 –
139.
1232.01(13) Š3, Z = 2, 1calcd = 1.175 gcmÀ3
,
3.64 ꢀ q ꢀ 27.428,
R1 = 0.0502 (I > 2s(I)), wR2 = 0.1296 (all data); residual electron
À3
density 0.203/À0.172 eŠ
. 3(tmeda)2: C20H29B2KN3, 173 K,
¯
0.16 ” 0.08 ” 0.06 mm, triclinic, space group P1, a = 7.4770(5),
b = 11.9330(9), c = 12.3420(10) Š, a = 94.030(3), g = 97.951(4),
b = 101.459(5)8, V= 1063.43(14) Š3, Z = 2, 1calcd = 1.162 gcmÀ3
,
2.50 ꢀ q ꢀ 24.98, R1 = 0.051 (I > 2s(I)), wR2 = 0.138 (all data);
À3
residual electron density 0.22/À0.26 eŠ
.
4·1.5C6H14:
C57H77B4N2Ru2, 173 K, 0.20 ” 0.12 ” 0.05 mm, monoclinic, space
group C2, a = 24.955(8), b = 19.629(5), c = 10.626(3) Š, b =
114.964(15)8, V= 4719(2) Š3, Z = 4, 1calcd = 1.458 gcmÀ3, 3.24 ꢀ
q ꢀ 27.488, R1 = 0.0387 (I > 2s(I)), wR2 = 0.1125 (all data);
residual electron density 0.968/À0.695 eŠÀ3. MoKa radiation
(l = 0.71073 Š). Non-hydrogen atoms were refined anisotropi-
364
www.angewandte.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 361 –364