Catalytic activation of N-N multiple bonds: A homogeneous palladium catalyst for mechanistically unprecedented reduction of azo compounds

Article abstract of DOI:10.1002/anie.200503875

Just two oxidation states are involved in the homogeneous reduction of azo compounds at a single palladium catalyst site. The reaction proceeds through an intermediate state in which the catalyst displays a novel palladadiaziridine structure, and ethanol or related alcohols can be employed as the terminal reducing agent (see scheme). (Chemical Equation Presented).

Full text of DOI:10.1002/anie.200503875

Angewandte
Chemie
transition-metal complexes for dinitrogen reduction under
nonbiological conditions.[3,4]
The coordination of dinitrogen to transition metals
generates the conceptually simplest complexes with activated
À
N N multiple bonds. Despite all the progress in nitrogen
fixation and in the design of suitable precursor transition-
metal complexes,[5] almost all of these systems are limited to
stoichiometric reactions. A noteworthy exception is the
monomeric molybdenum complex bearing a tetradentate
trisanilineamide ligand recently reported by Schrock[6] which
was shown to be capable of coordinating dinitrogen and
catalyzing consecutive cycles of ammonia formation. The
mechanism of this process involves end-on binding of the N2
molecule through preferential lone-pair coordination and
follows the classical Chatt model.[7]
Electronic and structural constraints usually prevent the
alternative, h2-dinitrogen coordination mode A. However,
these constraints do not necessarily apply to related nitrogen–
nitrogen multiple-bond systems. In these
Homogeneous Catalysis
DOI: 10.1002/anie.200503875
cases, the side-on coordination in
B
should allow equal activation of the two
nitrogen atoms involved and thus lead to
simultaneous reduction. Moreover, the
À
Catalytic Activation of N N Multiple Bonds:
A Homogeneous Palladium Catalyst for
Mechanistically Unprecedented Reduction of
Azo Compounds**
À
conversion of N N multiple bonds other than dinitrogen is
a promising source of alternative nitrogen-containing prod-
ucts for subsequent organic transformations.[8] In view of the
manifold catalytic utility of palladium,[9] the lack of catalytic
Kilian Muæiz* and Martin Nieger
À
activation of N N multiple bonds by this element is
notable.[10] Thus, the synthesis and study of palladium com-
À
Ready access to nitrogen compounds is of crucial importance,
as these molecules are key substances for research in the life
sciences. The catalytic reductive transformation of nitrogen–
nitrogen multiple bonds at a metal complex constitutes a
fundamental endeavor to this end. However, even a century
after Haberꢀs pioneering work on the heterogeneously
catalyzed synthesis of ammonia from its elements,[1] only
pounds resulting from the complexation of N N multiple
bonds is an interesting task.[11]
Herein, we present the synthesis of structurally defined
palladium complexes of azo compounds, which originate from
an initial h2-coordination mode (B), and report the first
À
catalytic reduction of an N N double bond at a single
palladium center under homogeneous conditions. Our inves-
tigation of the potential catalytic reactivity began with the
synthesis of palladium complexes with azo ligands, which can
be considered to be formal surrogates for the products of the
first step in the reduction of dinitrogen. The known (bath-
ocuproine)palladium(0) complex 1[12] (bathocuproine = 2,9-
dimethyl-4,7-diphenyl-1,10-phenanthroline) was treated with
an equimolar amount of the azodiesters 2a–c to form in an
irreversible reaction the desired monomeric complexes 3a–c
as air- and moisture-stable yellow-to-orange solids
(Scheme 1). The coordinated azo ligands in complexes 3a–c
do not dissociate on the NMR time scale, and the products
À
limited progress on the activation of N N multiple bonds has
been achieved. For the purpose of dinitrogen activation,
nature employs nitrogenase metalloenzymes, which under
physiological conditions transform dinitrogen into ammo-
nia.[2] This fascinating process has stimulated interest in
[*] Prof. Dr. K. Muæiz
FacultØ de Chimie, Institut Le Bel
UniversitØ Louis Pasteur
4, rue Blaise Pascal, 67070 Strasbourg Cedex (France)
Fax: (+33)3-9024-1526
E-mail: muniz@chimie.u-strasbg.fr
Dr. M. Nieger
X-Ray Analysis
Institut für Anorganische Chemie
Universität Bonn
Gerhard-Domagk-Strasse 1, 53121 Bonn (Germany)
[**] This work was supported by the UniversitØ Louis Pasteur and by the
Fonds der Chemischen Industrie. K.M. thanks Prof. J.-M. Lehn for
providing temporary laboratory space at the Institut de Science et
d’IngØnierie SupramolØculaires (ISIS).
Scheme 1. Synthesis of palladadiaziridines 3a–c. The yields of 3a–c
(given in parentheses) refer to isolated material. dba=trans,trans-
dibenzylideneacetone.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
Angew. Chem. Int. Ed. 2006, 45, 2305 –2308
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2305

Communications
3a–c are inert towards subsequent
exchange reactions with alkenes and other
azo compounds.[13] For example, treatment
of 3a with an excess of 2b led to no
interconversion of the coordinated azo-
diester.
The product 3a was crystallized from a
solution in dichloromethane, and the struc-
ture obtained from the X-ray diffraction
study of 3a·CH2Cl2 is depicted in Figure 1.
The expected C2 symmetry of the complex
is distorted as a result of unsymmetrical
hydrogen bonding of the cocrystallized
solvent molecule to the metal-coordinated
À
azo group. Importantly, rather short Pd N
bond lengths in the range of 2.014–2.043 Š
Scheme 2. Protonolysis of palladadiaziridines 3a–c. R: see Scheme 1.
basis of its NMR data by comparison with an authentic
sample.[14] No reversible reaction between the formed palla-
dium(ii) complex and the hydrazine was observed, which
suggests that hydrazine formation from protonation of
complexes 3 represents the desired irreversible process. In
an analogous reaction sequence, treatment of 3a with formic
acid led to quantitative production of the corresponding
hydrazine 5a and an unstable palladium formiate complex 6,
which decomposed over a period of two hours at room
temperature to palladium black.[15]
The same reactivity of the preformed complexes 3a–c was
observed for reactions on a preparative (1-mmol) scale
(Table 1). In this case, protonolysis with acetic acid
Figure 1. Solid-state structure of 3a. a) ORTEP plot (thermal ellipsoids
are drawn at 50% probability level). The cocrystallized dichlorome-
thane molecule and all hydrogen atoms have been omitted for clarity.
Selected bond lengths [Š] and angles [8]: N1A-N1B 1.404(5), Pd-N1A
2.043(3), Pd-N1B 2.014(4); N1B-Pd1-N1A 40.49(14), N1-Pd1-N10
78.93(14), N1B-N1A-Pd1 68.62(19), N1A-N1B-Pd1 70.9(2). b) Space-
filling model of 3a. Pd cyan, C gray, H white, N blue, O red.
Table 1: Protonolysis of palladadiaziridines
3 under homogeneous
conditions.[a]
Entry
Compound
Coordinated
R2N2
Product
R2(NH)2
Yield [%][b]
1
3a
3a
R=CO2Et
R=CO2Et
R=CO2tBu
R=CO2iPr
R=Ph
5a
5a
5b
5c
5d
92
93
94
88
90
2[c]
3
À
are observed together with an N N bond length of 1.404 Š,
3b
3c
À
which establishes the single-bond character of the N N bond
4
5
3d[d]
and leads to the conclusion that a Pd-N-N three-membered
ring is the dominating bonding situation. The overall struc-
tural composition is therefore best described as a palladadia-
ziridine. Thus, complexes 3a–c represent a hitherto-unknown
bonding situation within Pd complexes. In other words, the
coordinative interaction of the azo compound with the initial
Pd0 center involves a formal redox process between the metal
center and the incoming ligand and already entails the desired
[a] Reaction conditions: 1 mmol 3 (c=0.25m solution in CH2Cl2, T=
258C), addition of 2.2 mmol acetic acid. [b] Yield of product after
purification. [c] With formic acid. [d] Complex generated in situ.
(2.2 mmol) furnished the corresponding hydrazines 5a–c in
very good yields (88–94%). Also, protonolysis of 3d, formed
in situ by complexation of azobenzene, gave N,N’-diphenyl-
hydrazine (5d) in 90% yield. This result confirms that this
reaction is equally efficient for azo ligands with either
electron-withdrawing or electron-donating substituents.
In a subsequent experiment, the addition of ethanol to
[(bathocuproine)Pd(OAc)2] (4) led to facile reduction and
regeneration of a formal palladium(0) complex. The reaction
was complete within minutes at room temperature and
produced a black precipitate. An NMR experiment in
[D6]benzene revealed that the product of the ethanol
oxidation is acetaldehyde, which is in agreement with the
earlier reports on bathocuproine-based palladium catalysts
À
reduction of the N N double bond. The space-filling model of
3a (Figure 1b) illustrates that external electrophiles should
be readily able to initiate reactions at the coordinated
nitrogen atoms.
These conclusions were supported further by the obser-
vation that complexes 3a–c readily undergo displacement of
hydrazines upon protonolysis. For example, the addition of
acetic acid to NMR sample solutions of 3a and 3b in CD2Cl2
led within seconds to clean formation of [(bathocuproine)Pd-
(OAc)2] (4), as well as the N,N’-di(ethyloxycarbonyl)- and
N,N’-di(tert-butyloxycarbonyl)-protected hydrazines 5a and
5b, respectively (Scheme 2). Complex 4 was identified on the
2306 www.angewandte.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 2305 –2308

Angewandte
Chemie
for efficient alcohol oxidation.[16] More importantly, when the
same reaction was carried out in the presence of an excess of
diethyl azodicarboxylate (dead, 2a), quantitative formation
of the initial azo complex 3a was observed and the deposition
of palladium black complex was thus prevented (Scheme 3).
89% yield, respectively). In one case, the catalyst loading of
isolated 3a could be further reduced to 0.5 mol% (substrate/
catalyst ratio S/C = 200), and hydrazine 5a was isolated after a
reaction time of 36 h in 90% yield, which is equivalent to an
average catalyst turnover number of 180. From a mechanistic
point of view, the transition-metal catalysis cycles between the
two Pd0 and PdII oxidation states and involves only three
defined catalyst states. Hence, any potential catalyst deacti-
vation is largely avoided.
The observed process is universal. For example, the same
reactivity at an S/C ratio of 50 was obtained for the reduction
of azobenzene (2d) in ethanol (93% yield of hydrazine 5d)
and for 2c in 2-propanol (87% yield of hydrazine 5c). The
latter catalysis required 3 days to reach completion because
palladium reduction with 2-propanol to yield acetone is
comparatively slow.[17]
Scheme 3. Reductive regeneration of palladadiaziridine 3a from diac-
etate 4.
On the basis of these three individual stoichiometric
reactions, a catalytic cycle could be devised for the catalytic
reduction of azo compound 2a to give hydrazine 5a. A
solution (5m) of 2a in ethanol in the presence of palladium
catalyst (2 mol%) and acetic acid (5 mol%) was reduced
cleanly at 508C to yield hydrazine 5a in excellent yield.
Scheme 4 depicts the catalytic cycle. The palladadiaziridine is
The overall cycle from Scheme 4 is reminiscent of that
discovered by EniChem for palladium-catalyzed production
of hydrogen peroxide through dioxygen activation.[18] The
corresponding (peroxo)palladium catalyst from this reaction
was isolated by Stahl.[12] However, the azo activation reported
herein has the advantage that it constitutes an irreversible
process in which interaction between hydrazines and the
palladium catalyst are absent, whereas the production of
hydrogen peroxide requires biphasic conditions to prevent
disproportionation.
In summary, we have discovered the first homogeneous
transition-metal-catalyzed activation of azo compounds,
À
which enables efficient reductive transformation of N N
multiple bonds at a single palladium catalyst site. The catalysis
involves an intermediate state displaying a novel azo coordi-
nation mode to palladium(0), and the structure was estab-
lished by independent isolation of a palladadiaziridine.[19] The
present process constitutes a novel catalytic reaction and is
À
important for the reduction of N N multiple bonds.
Experimental Section
3a: A solution of [(bathocuproine)Pd(dba)] (802 mg, 1.14 mmol) in
freshly distilled dichloromethane (2 mL) was treated with dead (2a,
200 mL, 1.27 mmol), and the resulting mixture was stirred at room
temperature for 45 min. The solvent was removed under reduced
pressure to a volume of approximately 0.5 mL, and the crude product
was precipitated by addition of absolute diethyl ether (20 mL). The
yellowish crude product was washed with further portions of diethyl
ether, dried under reduced pressure, dissolved in dichloromethane
(10 mL), and then filtered through a pad of celite. The remaining
solution was evaporated to dryness under reduced pressure to leave
the product as a yellow-to-orange solid (592 mg, 0.92 mmol, 81%).
Crystals suitable for X-ray analysis were grown from a solution in
dichloromethane at À208C.
Scheme 4. Catalytic cycle for reduction of azo compounds to hydra-
zines with a bathocuproine–Pd catalyst. The bathocuproine ligand is
represented schematically. S=solvent molecule.
protonated and releases the free hydrazine to form a
palladium diacetate complex. Oxidation of ethanol to acet-
aldehyde concomitantly forms a palladium(0) complex and
regenerates the catalytic amount of acetic acid. Coordination
of the azo substrate to palladium(0) regenerates the pallada-
diaziridine and completes the catalytic cycle. This proposed
overall catalytic cycle is supported by the fact that both the
palladadiaziridine 3a and the palladium diacetate 4 can be
employed as the catalyst source with equal efficiency (91 and
Crystal structure of 3a: C32H30N4O4Pd · CH2Cl2: yellow crystals,
crystal dimensions 0.25 ” 0.15 ” 0.10 mm3; M = 725.93; monoclinic,
space group P21/c (No. 14), a = 12.3133(4), b = 15.1348(6), c =
16.8231(7) Š, b = 97.226(2)8, V= 3110.2(2) Š3, Z = 4, m(MoKa) =
0.813 mmÀ1, T= 123(2) K, F(000) = 1480. 19374 reflections up to
2qmax = 508 were measured on a Nonius KappaCCD diffractometer
with MoKa radiation, 5469 of which were independent and used for all
calculations. The structure was solved by direct methods and refined
to F2 anisotropically; the positions of the H atoms were refined with a
riding model. The final quality coefficient wR2(F2) for all data was
Angew. Chem. Int. Ed. 2006, 45, 2305 –2308
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
2307

Communications
0.0871, with a conventional R(F) = 0.0425 for 399 parameters. An
empirical absorption correction was applied.
[9] J. Tsuji, Palladium Reagents and Catalysts, Wiley, New York,
2004.
CCDC-288043 contains 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.
[10] For the use of palladium catalysts in conventional hydrogenation
reactions of azo compounds, see: a) F. Figueras, B. Coq, J. Mol.
Catal. A 2001, 173, 223; b) K. Abiraj, G. R. Srinivasa, D. C.
Gowda, Can. J. Chem. 2005, 83, 517.
[11] For example, detection of palladium dinitrogen complexes has so
far been limited to low-temperature matrix-isolation techniques:
a) H. Huber, E. P. Kꢁndig, M. Moskovits, G. A. Ozin, J. Am.
Chem. Soc. 1973, 95, 332; b) for an early direct synthetic attempt,
see: C. A. Tolman, D. H. Gerlach, J. P. Jesson, R. A. Schunn, J.
Organomet. Chem. 1974, 65, C23.
Received: November 2, 2005
Published online: March 3, 2006
Keywords: azo compounds · homogeneous catalysis ·
.
hydrazines · palladium · small rings
[12] S. S. Stahl, J. L. Thorman, R. C. Nelson, M. A. Kozee, J. Am.
Chem. Soc. 2001, 123, 7188.
[13] For reversible exchange of alkenes on bathocuproine-Pd-alkene
complexes, see: S. S. Stahl, J. L. Thorman, N. de Silva, I. A.
Guzei, R. W. Clark, J. Am. Chem. Soc. 2003, 125, 12.
[14] Treatment of bathocuproine with palladium(ii) acetate in
dichloromethane gives 4 as a yellow-to-ochre solid (96% yield
after precipitation from diethyl ether). We find this procedure
more convenient than those reported in the literature: R. Santi,
A. M. Romano, R. Garrone, R. Millini, J. Organomet. Chem.
1998, 566, 37.
[1] a) F. Haber, G. Z. van Oordt, Z. Anorg. Chem. 1905, 47, 42; b) F.
Haber, Naturwissenschaften 1922, 10, 133.
[2] a) O. Einsle, F. A. Tezcan, S. L. A. Andrade, B. Schmidt, M.
Yoshida, J. B. Howard, D. C. Reed, Science 2002, 297, 1696;
b) P. C. Dos Santos, R. Y. Igarashi, H.-I. Lee, B. M. Hoffman,
L. C. Seefeldt, D. R. Dean, Acc. Chem. Res. 2005, 38, 208;
c) B. K. Burgess, D. J. Lowe, Chem. Rev. 1996, 96, 2983; d) J. B.
Howard, D. C. Rees, Chem. Rev. 1996, 96, 2965.
[3] a) C. E. Laplaza, C. C. Cummins, Science 1995, 268, 861;
b) M. D. Fryzuk, J. B. Love, S. J. Rettig, V. G. Young, Science
1997, 275, 1445; c) Y. Nishibayashi, S. Iwai, M. Hidai, Science
1998, 279, 540; d) J. A. Pool, E. Lobkovsky, P. J. Cirik, Nature
2004, 427, 527; e) J. M. Smith, R. J. Lachicotte, K. A. Pittard,
T. R. Cundari, G. Lukat-Rodgers, K. R. Rodgers, P. L. Holland,
J. Am. Chem. Soc. 2001, 123, 9222.
[4] For recent reviews, see: a) B. A. MacKay, M. D. Fryzuk, Chem.
Rev. 2004, 104, 385, and references therein; b) M. P. Shaver,
M. D. Fryzuk, Adv. Synth. Catal. 2003, 345, 1061; c) see also: M.
Reiher, B. Kirchner, J. Hutter, D. Sellmann, B. A. Hess, Chem.
Eur. J. 2004, 10, 4443, and references therein.
[5] T. A. Bazhenova, A. E. Shilov, Coord. Chem. Rev. 1995, 144, 69.
[6] a) D. V. Yandulov, R. R. Schrock, Science 2003, 301, 76; b) R. R.
Schrock, Chem. Commun. 2003, 2389; c) R. R. Schrock, Acc.
Chem. Res. 2005, 39, 955; for theoretical treatments, see: d) Z.
Cao, Z. Zhou, H. Wan, Q. Zhang, Int. J. Quantum Chem. 2005,
103, 344; e) F. Studt, F. Tuczek, Angew. Chem. 2005, 117, 5783;
Angew. Chem. Int. Ed. 2005, 44, 5639; f) B. Le Guennic, B.
Kirchner, M. Reiher, Chem. Eur. J. 2005, 11, 7448; g) M. Reiher,
B. Le Guennic, B. Kirchner, Inorg. Chem. 2005, 44, 9640.
[7] a) J. Chatt, A. J. Pearman, R. L. Richards, Nature 1975, 253, 39;
b) J. Chatt, J. R. Dilworth, R. L. Richards, Chem. Rev. 1978, 78,
589.
[15] Once the black palladium(0) species formed, it could not be
treated with excess azo compound to regenerate complexes 3.
We thus rule out the involvement of heterogeneous reduction
pathways.
[16] a) G.-J. ten Brink, I. W. C. E. Arends, R. A. Sheldon, Science
2000, 287, 1636; b) G.-J. ten Brink, I. W. C. E. Arends, M.
Hoogenraad, G. Verspui, R. A. Sheldon, Adv. Synth. Catal.
2003, 345, 497.
[17] Monitoring the corresponding stoichiometric reduction by
NMR spectroscopy revealed only incomplete reaction after 5 h
at room temperature. This result suggests that palladium
reduction is the rate-limiting step in the related catalytic
reduction of 3c to hydrazine 5c, which is in agreement with
the general behavior of 4 in alcohol oxidation reactions.[16]
[18] a) D. Bianchi, R. Bortolo, R. DꢀAloisio, M. Ricci, Angew. Chem.
1999, 111, 734; Angew. Chem. Int. Ed. 1999, 38, 706; b) D.
Bianchi, R. Bortolo, R. DꢀAloisio, M. Ricci, J. Mol. Catal. A
1999, 150, 87; c) R. Bortolo, D. Bianchi, R. DꢀAloisio, C. Querci,
M. Ricci, J. Mol. Catal. A 2000, 153, 25.
[19] The catalytic behavior is not restricted to the bathocuproine
ligand. Other 1,10-phenanthroline ligands induce comparable
reactivity. Studies of the structural requirements of the ligand
constitution are currently under way.
[8] M. Hidai, Y. Mizobe, Can. J. Chem. 2005, 83, 358.
2308 www.angewandte.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 2305 –2308