Heterolytic splitting of hydrogen with rhodium(I) amides

Article abstract of DOI:10.1002/anie.200500773

A split decision: Efficient heterolytic hydrogen splitting is achieved with novel tetracoordinate rhodium(I) amides. This reaction is reversible, and DFT calculations imply that the cleavage is a one-step process with a relatively low activation barrier. Ketones and imines are hydrogenated by the amide complexes without the need for further additives. (Chemical Equation Presented)

Full text of DOI:10.1002/anie.200500773

Communications
Hydrogenation Catalysts
DOI: 10.1002/anie.200500773
Heterolytic Splitting of Hydrogen with
Rhodium(i) Amides**
Pascal Maire, Torsten Büttner, Frank Breher,
Pascal Le Floch, and Hansjörg Grützmacher*
Dedicated to Professor Gottfried Huttner
on the occasion of his 68th birthday
The classical mechanisms for the catalytic hydrogenation of
=
C C double bonds with rhodium(i) or iridium(i) complexes
consist of six steps: 1) ligand dissociation from the catalyst
precursor, 2) oxidative addition of H2, 3) olefin coordination,
=
ꢀ
4) insertion of the coordinated C C bond in a Rh H bond,
5) isomerization, and 6) reductive elimination of the prod-
uct.[1] In the Halpern mechanism,[2] the olefin insertion and in
the Brown mechanism the reductive elimination of the alkane
is rate-determining.[3] In both mechanisms, a T-shaped 14-
electron [ML2X] complex is the key intermediate which adds
H2 oxidatively in an almost barrierless exothermic reaction.
The heterolytic addition of hydrogen across a metal–nitrogen
bond was first investigated systematically by Fryzuk and co-
workers.[4] In the meantime, this reaction has been recognized
as a key step in the very efficient catalytic hydrogenation of
1
[5–7]
=
unsaturated substrates RR C X,
especially of ketones
(X = O), which became known as metal–ligand bifunctional
catalysis through the work of Noyori and Morris. This
mechanism is different and involves: 1) heterolytic addition
of H2 across the metal–amide bond as the rate-determining
1
=
step, 2) binding of RR C X in the second coordination
sphere of the MHdꢀ–NHd+ unit, 3) “concerted” transfer of
d+
Hdꢀ from the metal atom to the C X carbon atom and H
=
from the nitrogen atom to the X center, and 4) product
release. In this mechanism no change of the formal oxidation
state and no major structural changes in the first coordination
sphere of the metal center occur. Combined with the
possibility to isolate the species that are directly involved in
[*] Dr. P. Maire, Dr. T. Büttner, Dr. F. Breher, Prof. Dr. P. Le Floch,[+]
Prof. Dr. H. Grützmacher
Department of Chemistry and Applied Biosciences
ETH-Hönggerberg, 8093 Zürich (Switzerland)
Fax: (+41)1-633-1032
E-mail: gruetzmacher@inorg.chem.ethz.ch
[+] Permanent address:
Department of Chemistry
Laboratory “HØtØroØlØments et Coordination”
UMR CNRS 7653, Ecole Polytechnique
91128 Palaiseau cedex (France)
[**] This work was supported by the LANXESS AG and the Swiss
National Science Foundation. We thank our colleague A. Togni for
fruitful discussions and the reviewers for valuable comments.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
6318
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2005, 44, 6318 –6323

Angewandte
Chemie
the catalytic cycle,[7b,8] new possibilities for rational catalyst
design emerge.
fully deprotonated by addition of one equivalent of KOtBu to
4a,b in THF; this is accompanied by an immediate color
change of the reaction solution from orange-red to intense
green. Deep green, highly air-sensitive crystals of 5a,b grew
from a 1:1 mixture of THF/toluene which was layered with n-
hexane. The result of a structure analysis on a single crystal of
5b is shown in Figure 1.[11]
Recently, we isolated rhodium(i) amide olefin complexes
with a rigid tetrahedrally distorted square-planar structure.[9]
In accord with a calculated high activation barrier (>
29 kcalmolꢀ1), these compounds do not split H2 heterolyti-
cally. Here we report the syntheses of rhodium amides with a
novel structure. These complexes are easily prepared in a few
steps, can be isolated, split H2 heterolytically, and are directly
active hydrogenation catalysts for ketones and imines without
the need for any additives.
The reaction of bis(5-H-dibenzo[a,d]cyclohepten-5-yl)-
amine (1, bis(tropylidenyl)amine, trop2NH)[10a] with [Rh2(m2-
Cl)2(cod)2] (2) gives the dinuclear complex [Rh2(m2-Cl)2-
(trop2NH)2] (3, Scheme 1).
Figure 1. Structure of 5b. Thermal ellipsoids are drawn at 30% prob-
ability; hydrogen atoms are omitted for clarity. Selected bond lengths
ꢀ
ꢀ
ꢀ
[Š] and angles [8]: Rh N1 2.007(1), Rh P1 2.316(1), Rh C5 2.165(2),
ꢀ
ꢀ
ꢀ
ꢀ
Rh C4 2.190(2), Rh C19 2.174(2), Rh C20 2.199(2), Rh ct1 2.058(2),
ꢀ
ꢀ
ꢀ
Rh ct2 2.070(2), C4 C5 1.423(3), C19 C20 1.407(3); N1-Rh-P1
166.18(5), ct1-Rh-ct2 135.81(7), C16-N1-C1 109.5(1), C16-N1-Rh
118.5(1), C1-N1-Rh 119.0(1).
The structure of the neutral amide 5b is unique among
tetracoordinate d8 rhodium complexes. However, a close
relationship exists to the zero-valent 16 valence-electron
(VE) [Ru(CO)2(PR3)2] complexes studied by Caulton, Eisen-
stein, et al.[12] and to the highly reactive transient carbonyls
[M(CO)4] (M = Fe, Ru, Os).[13] Indeed, like these species, 5b
adopts a “sawhorse” structure with a N-Rh-P angle of
166.18(5)8 and a ct1-Rh-ct2 angle of 135.81(7)8 (ct = centroid
=
of the coordinated C C bond). Comparable angles in
[Ru(CO)2(PtBu2Me)2] are: P-Ru-P 165.56(8)8, and C-Ru-C
133.3(4)8.
Scheme 1. Synthesis of rhodium bis(trop)amine complexes 4a,b, 5a,b,
6a,b, and 7a,b. cod=cycloocta-1,5-diene.
The amide nitrogen atom N1 has a pyramidal coordina-
tion sphere (ꢀ = 3478). At temperatures below 220 K, sharp
and distinct 1H NMR resonances for the inequivalent protons
at the olefinic carbon atoms C4/C20 and C5/C19, respectively,
are observed, which demonstrates that the sawhorse structure
corresponds to the ground-state structure. At room temper-
ature, these NMR resonances collapse to give one broadened
singlet, indicating inversion at the nitrogen and rhodium
centers, probably via a planar transition state.
The rhodium amides 5a,b react rapidly and quantitatively
with H2 (1 atm) even at ꢀ788C to give the yellow rhodium
hydride complexes [RhH(trop2NH)(PPh2R)] (6a,b). This
reaction is very likely reversible as 6b reacts with D2 to give
Subsequent reaction with a phosphane leads to the
mononuclear complexes [RhCl(trop2NH)(PPh2R)] 4a (R =
Ph) and 4b (R = 4-MeC6H4 = Tol) in which the phosphane
ligand is in the equatorial position and the chloro substituent
is in the apical position (these stereochemical assignments are
based on the NMR spectroscopic data). The complexes 3 (red
crystals) and 4a,b (yellow crystals) are obtained quantita-
tively and can be stored in air.
The NH function of the rhodium(i)-coordinated trop2NH
ligands is sufficiently acidic (pKa 15–20 in DMSO)[9,10] to be
Angew. Chem. Int. Ed. 2005, 44, 6318 –6323
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
6319

Communications
[D2]-6b exclusively and, vice versa, [D2]-6b reacts with H2 to
trienyl, Figure 3).[15] The formation of the hydride II is
exothermic (DRH = ꢀ15.8 kcalmolꢀ1) and proceeds in one
step[16] via the transition state TS-I. The calculated NBO
charges show that the H2 molecule is significantly polarized
give 6b. The latter reaction was monitored by 1H NMR
spectroscopy and shows that the intensities of the signals for
the NH and RhH protons build up simultaneously. We have
no evidence that deuterium labeling occurs at any other
position in the molecule. The structure of 6b was determined
by X-ray diffraction (Figure 2).[11]
Figure 3. DFT[15] computed energies for the heterolytic (shown in
black) and oxidative addition (shown in gray) of H2 to the model com-
plex I. NBO charges are given for the Rh, N, and the adding H2 mole-
cule in I, TS-I, and II.
(DqNBO = 0.33 e) in the transition state (compare with
I
DqNBO = 0.57 e in the product II). The polarity of the Rh
N
=
ꢀ
bond varies little throughout this addition process (DqNBO
0.94 e (I), 0.91 e (TS-I), 0.83e ( II)). The “classical” oxidative
addition of H2 leading to the rhodium(iii) dihydride III is an
unfavorable endothermic process (DRH = 17.9 kcalmolꢀ1)
and has to encompass the transition state TS-II which is
higher in energy than TS-I by 3.4 kcalmolꢀ1.
Figure 2. Structure of 6b. Thermal ellipsoids are drawn at 30% prob-
ability; hydrogen atoms apart from H1 and H2 and one THF solvent
molecule are omitted for clarity. One position of the disordered para-
methyl substituent within the PPh2Tol ligand is shown. Selected bond
The isolated crystalline amide 5b and the hydride 6b were
used in catalytic hydrogenations without any further addi-
tives. The conditions and results are listed in Table 1. Under
these (not yet optimized) conditions,[17] alkyl- (8) and aryl-
substituted ketones (9) and the imine 10 are completely
converted to the corresponding hydrogenated products cyclo-
hexanol, 1-phenylethanol, and N-phenylbenzylamine, respec-
tively. The amide 5b and the hydride 6b are equally active
(see Table 1, entries 2 and 3). Addition of an excess of
phosphine does not significantly slow down the catalytic
activity. However, at the end of each catalytic run, the axial
ꢀ
ꢀ
ꢀ
lengths [Š] and angles [8]: Rh N1 2.178(1), Rh P1 2.230(1), Rh H2
ꢀ
ꢀ
ꢀ
ꢀ
1.59(3), Rh C5 2.159(1), Rh C4 2.199(1), Rh C19 2.158(1), Rh C20
ꢀ
ꢀ
ꢀ
ꢀ
2.203(1), Rh ct1 2.057(1), Rh ct2 2.059(1), C4 C5 1.437(2), C19 C20
1.436(2), N1···O1 3.02; N1-Rh-P1 169.95(3), ct1-Rh-ct2 132.54(5), C16-
N1-C1 110.7(1), C16-N1-Rh 116.15(7), C1-N1-Rh 117.36(7).
The structural differences between 5b and 6b are very
ꢀ
ꢀ
small. In 6b the Rh N bond is 9% longer and the Rh P bond
is 4% shorter than in 5b; the ct1-Rh-ct2 angle is 2% more
acute, whereas the N-Rh-P angle is 2% more open. Also the
coordination sphere at N1 is not influenced very much (ꢀ(C-
N-C, 2”C-N-Rh) = 344.28). ATHF molecule is coordinated to
the NH function (N1···O1 3.02 Š) and indicates, as does the
Table 1: Hydrogenation of the ketones 8 and 9 and the imine 10 with 5b
or 6b as catalysts.[a]
1
high-frequency H NMR shift (d: 5.56 ppm (6a), 5.09 ppm
(6b)), its acidic character. The hydride ligand H2 in the
equatorial position of the trigonal-bipyramidal structure of 6b
causes the typical[14] shift of the olefinic 13C resonances to low
frequency (by about 20 ppm) and an elongation of the
Entry
Substrate
Cat.
S/C[b]
t [h]
Conversion [%]
=
coordinated C C bonds.
Solutions of the recrystallized hydrides 6a,b in THF are
stable for at least 24 h. However, impurities provoke the
quantitative isomerization to the air-stable yellow complexes
7a,b in which the hydride ligand adopts the axial position.
The assumption that the amide complexes 5a,b split H2
heterolytically is supported by DFT calculations with the
model complex [Rh(cht2N)(PH3)] (I) (cht = cyclohepta-
1
2
3
4
5
6
8
9
9
9
9
10
5b
5b
6b
5b
5b
5b
100
100
100
1000
1000
100
16
16
16
1.5
16
16
>97
>97
>97
22
65
>97
[a] THF, 100 bar H2 , T=298 K. [b] Ratio substrate/catalyst.
6320 www.angewandte.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2005, 44, 6318 –6323

Angewandte
Chemie
1
662 nm (1000). 5b: H NMR (500.2 MHz, [D8]THF, 190 K): d = 4.74
hydride 7b is the only detectable rhodium- and phosphorus-
containing species and this hydride is catalytically inactive.
We therefore assume that the formation of 7b in course of the
catalytic reaction is responsible for the incomplete conversion
of 9 with low catalyst loadings (Table 1, entry 5).
In summary, rhodium amides with previously unreported
structures can be prepared from readily available starting
materials. These complexes react cleanly with H2 which is
(dd, 3JH,H = 7.3Hz, J = 7.3Hz, 2H; CH olefin), 5.68 ppm (d, JH,H
=
3
7.3Hz, 2H; CH olefin); 13C NMR (125.8 MHz, [D8]THF, 190 K): d =
75.7 (s, 2C, CHolefin), 83.9 ppm (d, 1JRh,C = 15.1 Hz, 2C; CHolefin);
31P NMR (121.5 MHz, [D8]THF): d = 38.7 ppm (d, 1JRh,P = 124 Hz).
6a: 1H NMR (400.1 MHz, [D8]THF): d = ꢀ8.15 (dd, 1JRh,H = 23.0 Hz,
2JP, H = 23.0 Hz, 1H; RhH), 3.55 (d, JH,H = 9.3Hz, 2H; CH olefin), 3.91
3
(dd, JH,H = 9.3Hz, 3JH,H = 4.7 Hz, 2H; CHolefin), 5.56 ppm (d, JP, H
=
3
3
4.9 Hz, 1H; NH); 13C NMR (100.6 MHz, [D8]THF): d = 57.8 (d,
1JRh,C = 8.0 Hz, 2C; CHolefin), 60.6 ppm (d, JRh,C = 8.6 Hz, 2C; CHol-
1
ꢀ
heterolytically split in a one-step process across the polar Rh
efin); 31P NMR (162.0 MHz, [D8]THF): d = 65.4 ppm (d, JRh,P
=
1
N bond. Both, the amide and hydride can be used directly as
catalysts for ketone or imine hydrogenations, which very
likely proceed via the metal–ligand bifunctional mecha-
nism.[17] Given that the isomerization of the catalytically
active hydride intermediate to an inactive one can be
suppressed, efficient new catalysts for hydrogenations are in
sight.
144 Hz); 103Rh NMR (12.6 MHz, [D8]THF, 230 K): d = ꢀ187 ppm
(d, 1JRh,P = 144 Hz). 6b: M.p.: > 1008C (decomp). 1H NMR
(300.1 MHz, [D8]THF): d = ꢀ8.19 (dd, 1JRh,H = 23.3 Hz, JP, H
=
2
23.3 Hz, 1H; RhH), 3.65 (d, 3JH,H = 9.4 Hz, 2H; CHolefin), 3.97 (dd,
3JH,H = 9.4 Hz, 3JP, H = 4.5 Hz, 2H; CHolefin), 5.09 ppm (d, 3JP, H = 5.5 Hz,
1H; NH); 13C NMR (75.5 MHz, [D8]THF): d = 57.0 (d, JRh,C
=
1
7.9 Hz, 2C; CHolefin), 60.0 ppm (d, 1JRh,H = 8.8 Hz, 2C; CHolefin);
31P NMR (121.5 MHz, [D8]THF): d = 63.0 ppm (d, 1JRh,P = 145 Hz);
ATR IR : n˜ = 3169 (m, NH), 1756 cmꢀ1 (m, RhH). [D2]-6b: 2H NMR
(46.1 MHz, THF): d = ꢀ8.19 (br s, RhD), 4.92 ppm (br s, ND);
31P NMR (121.5 MHz, [D8]THF): d = 61.3ppm. The ND and RhD IR
absorptions (expected at about 1580 and 880 cmꢀ1) are hidden by
intense absorptions from the ligand. 7a: M.p.: > 1508C (decomp).
1H NMR (400.1 MHz, [D8]THF): d = ꢀ21.37 (dd, 1JRh,H = 17.4 Hz,
2JP, H = 17.4 Hz, 1H; RhH), 0.82 (s, 1H; NH), 4.40 (dd, 3JH,H = 9.0 Hz,
Experimental Section
The syntheses of 4a,b can be performed without any particular
precautions. In contrast, the amides 5a,b must be handled under
exclusion of moisture and oxygen. Detailed descriptions of the
syntheses and spectroscopic data are given in the Supporting
Information.
3JP, H = 7.8 Hz, 2H; CHolefin), 5.15 ppm (dd, 3JH,H = 9.0, 3JP, H = 5.1 Hz,
2H; CHolefin); 13C NMR (100.6 MHz, [D8]THF): d = 57.1 (dd, JP, C
=
2
14.7 Hz, 1JRh,C = 9.8 Hz, 2C, CHolefin), 61.2 ppm (dd, 1JRh,C = 8.6 Hz,
2JP, C = 4.90 Hz, 2C, CHolefin); 31P NMR (162.0 MHz, [D8]THF): d =
47.3ppm (d, 1JRh,P = 138 Hz); 103Rh NMR (12.6 MHz, [D8]THF): d =
ꢀ38 ppm (d, 1JRh,P = 138 Hz). 7b: 1H NMR (300.1 MHz, [D8]THF):
d = ꢀ21.49 (dd, 1JRh,H = 17.3Hz, 2JP, H = 17.3Hz, 1H; RhH), 0.96 (s,
1H; NH), 4.47 (dd, 3JH,H = 8.8 Hz, 3JP, H = 7.8 Hz, 2H; CHolefin),
5.20 ppm (dd, 3JH,H = 8.8, 3JP, H = 5.0 Hz, 2H; CHolefin); 13C NMR
(75.5 MHz, [D8]THF): d = 56.4 (dd, 2JP, C = 14.3Hz, 1JRh,C = 9.4 Hz,
The reaction of 2 with two equivalents 1 in CH2Cl2 gave 3·CH2Cl2,
which was obtained as red crystals from the reaction mixture (yield >
90%). Subsequent reaction with PPh2R (R = Ph or Tol) gave yellow
solutions, from which 4a,b were precipitated by addition of n-hexane
(yields > 80%; 4a: R = Ph, 4b: R = Tol). The reaction of 4a,b with
one equivalent of KOtBu in THF gave a deep green reaction solution,
to which was added toluene (50vol%). After all volatiles had been
evaporated under vacuum, the green residue was extracted with THF,
filtered, and concentrated. Layering this THF extract with toluene
and n-hexane (THF/toluene/n-hexane = 1:1:10) gave dark green
microcrystals of 5a,b (yields: 80%). Deep green solutions of 5a,b
in THF were treated with H2 (D2) gas at 2 atm. Layering of the
resulting yellow solutions with n-hexane led to the crystallization of
the hydrides 6a,b or [D2]-6b as yellow air-stable platelets (yields >
80%).
2C; CHolefin), 60.4 ppm (dd, 1JRh,C = 8.8 Hz, 2JP, C = 5.2 Hz, 2C;
1
CHolefin); 31P NMR (121.5 MHz, [D8]THF): d = 45.0 ppm (d, JRh,P
=
138 Hz).
Received: March 2, 2005
Revised: May 25, 2005
Published online: September 19, 2005
NMR data were recorded at 298 K when not specified otherwise:
3: M.p.: > 2508C (decomp). 1H NMR (400.1 MHz, CD2Cl2): d = 5.97
Keywords: amides · heterolytic H2 splitting · homogeneous
(dd, 3JH,H = 9.4 Hz, 2JRh,H = 2.1 Hz, 4H; CHolefin), 6.20 ppm (d, 3JH,H
=
.
9.4 Hz, 4H; CHolefin); 13C NMR (100.6 MHz, CD2Cl2): d = 71.4 ppm
(br s, 8C; CHolefin); 103Rh NMR (12.7 MHz, CD2Cl2): d = 2992 ppm
(s); UV/Vis (CH3CN): lmax (e) = 232 (sh), 289 (45700), 329 nm (sh).
4a: M.p.: > 2608C (decomp). 1H NMR (300.1 MHz, CDCl3): d = 1.59
(br, 1H; NH), 5.42 (dd, 3JH,H = 9.4 Hz, 3JP, H = 7.4 Hz, 2H; CHolefin),
catalysis · hydrogenation · rhodium
[1] For a review citing experimental and theoretical work see: M.
Torrent, M. Solà, G. Frenking, Chem. Rev. 2000, 100, 439.
[2] a) J. Halpern, Inorg. Chim. Acta 1981, 50, 11; b) For calculations
see: C. Daniel, N. Koga, J. Han, X. Y. Fu, K. Morokuma, J. Am.
Chem. Soc. 1988, 110, 3773.
5.66 ppm (ddd, 3JH,H = 9.4 Hz, 3JP, H = 5.8 Hz, 2JRh,H = 1.3Hz, 2H;
1
CHolefin); 31P NMR (101.2 MHz, CDCl3): d = 7.7 ppm (d, JRh,P
=
111 Hz). 4b: M.p.: > 2608C (decomp). In CDCl3 solution two
ꢀ
conformations (likely due to hindered rotation around the Rh P
[3] a) J. M. Brown, Chem. Soc. Rev. 1993, 22, 25; b) For calculations
bond) in an approximate 2:1 ratio are observed. 1H NMR
(300.1 MHz, CDCl3): d = 1.59 (br s, 1H; NHmaj), 1.71 (br s, 1H;
NHmin), 5.35–5.48 (m, 2H; CHolefin,min; and 2H; CHolefin,maj), 5.60–
5.67 ppm (m, 2H; CHolefin,min; and CHolefin,maj); 31P NMR (121.5 MHz,
see: “The Challenge of d and f Electrons: Theory and
Computation”: N. Koga, K. Morokuma, ACS Symp. Ser. 1989,
394, 77.
[4] a) M. D. Fryzuk, P. A. MacNeil, Organometallics 1983, 2, 682;
b) M. D. Fryzuk, P. A. MacNeil, S. J. Rettig, J. Am. Chem. Soc.
1987, 109, 2803; c) M. D. Fryzuk, C. D. Montgomery, Coord.
Chem. Rev. 1989, 95, 1. The heterolytic splitting of H2 by metal
complexes, for example Cu2+ + H2!CuH+ + H+ was pro-
posed much earlier: d) J. Halpern, J. Organomet. Chem. 1980,
200, 133, and references therein; it is one key step in the
enzymatic H2 activation by metal–sulfur clusters in hydro-
genases: e) D. J. Evans, C. J. Pickett, Chem. Soc. Rev. 2003, 32,
268.
1
1
CDCl3): d = 6.9 (d, JRh,P = 111 Hz, maj), 7.6 ppm (d, JRh,P = 111 Hz,
min). 5a: 1H NMR (400.1 MHz, [D8]THF, 200 K): d = 4.69 (ddd,
3JH,H = 9.0 Hz, 3JP, H = 6.2 Hz, 2JRh,H = 1.2 Hz, 2H; CHolefin), 5.62 ppm
(ddd, 3JH,H = 9.0 Hz, 2JRh,H = 3.3 Hz, 3JP, H = 2.9 Hz, 2H; CHolefin);
1
13C NMR (100.6 MHz, [D8]THF, 200 K): d = 76.2 (d, JRh,C = 6.7 Hz,
2C, CHolefin), 84.5 ppm (d, 1JRh,C = 14.7 Hz, 2C, CHolefin). 31P NMR
(162.0 MHz, [D8]THF, 200 K): d = 40.8 ppm (d, 1JRh,P = 124 Hz); 103Rh
NMR (12.6 MHz, [D8]THF, 200 K): d = 838 ppm (d, 1JRh,P = 124 Hz);
UV/Vis (THF): lmax (e) = 301 (20000), 352 (10000), 438 (3000),
Angew. Chem. Int. Ed. 2005, 44, 6318 –6323
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
6321

Communications
[5] Recent reviews: a) S. E. Clapham, A. Hadzovic, R. H. Morris,
Gottschalk-Gaudig, J. C. Huffman, H. GØrard, O. Eisenstein,
K. G. Caulton, Inorg. Chem. 2000, 39, 3957.
Coord. Chem. Rev. 2004, 248, 2201; b) R. Noyori, Angew. Chem.
2002, 114, 2108; Angew. Chem. Int. Ed. 2002, 41, 2008; ; c) R.
Noyori, T. Ohkuma, Angew. Chem. 2001, 113, 40; Angew. Chem.
Int. Ed. 2001, 40, 40; d) M. J. Palmer, M. Wills, Tetrahedron:
Asymmetry 1999, 10, 2045.
[13] [M(CO)4]: a) M = Fe, Os: M. Poliakoff, J. J. Turner, J. Chem. Soc.
Dalton Trans. 1974, 2276; b) M = Ru: P. L. Bogdan, E. Weitz, J.
Am. Chem. Soc. 1989, 111, 3163; c) for calculations see: J. Li, G.
Schreckenbach, T. Ziegler, J. Am. Chem. Soc. 1995, 117, 486. For
a general discussion of the isolobal relationship between d8ML4
and CH2, see: T. A. Albright, J. K. Burdett, M.-H. Whangbo,
Orbital Interactions in Chemistry, Wiley, 1985, pp. 360, 404.
[14] C. Böhler, N. Avarvari, H. Schönberg, M. Wörle, H. Rüegger, H.
Grützmacher, Helv. Chim. Acta 2001, 84, 3127.
[6] Mechanistic and theoretical work: a) C. A. Sandoval, T.
Ohkuma, K. Muæiz, R. Noyori, J. Am. Chem. Soc. 2003, 125,
13490, and references therein; b) K. Abdur-Rashid, S. E. Clap-
ham, A. Hadzovic, J. N. Harvey, A. J. Lough, R. H. Morris, J.
Am. Chem. Soc. 2002, 124, 15104.
[15] All calculations were carried out within the framework of DFT
using the Gaussian03[15a] set of programs and the B3PW91
functional. This functional employs a combination of exchange
terms: exact HF, the Becke 1988 nonlocal gradient correction[15b]
and the original Slater local exchange functional.[15c] In addition,
it uses the Perdew–Wang 1991 local correlation functional.[15d]
The 6-31 + G* basis set was employed for all the atoms
connected to the Rh center as well as for the four CH olefinic
protons. All the other atoms, except Rh, were calculated using
the 6-31G* basis set. A quasirelativistic effective core potential
operator was used to represent the 28 innermost electrons of the
rhodium atom. The basis set for the metal was that associated
with the pseudopotential, with a standard double-z LANL2DZ
contraction.[15e] Cartesian coordinates of the theoretical struc-
tures are given in the Supporting Information. The minimum
energy structures were characterized by full vibration frequen-
cies calculations. Natural charges were computed with the NBO
program implemented in Gaussian03.[15 f] a) Gaussian03(Revi-
sionB.04), M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E.
Scuseria, M. A. Robb, J. R. Cheeseman, J. A. Montgomery, Jr.,
T. Vreven, K. N. Kudin, J. C. Burant, J. M. Millam, S. S. Iyengar,
J. Tomasi, V. Barone, B. Mennucci, M. Cossi, G. Scalmani, N.
Rega, G. A. Petersson, H. Nakatsuji, M. Hada, M. Ehara, K.
Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y.
Honda,O. Kitao, H. Nakai, M. Klene, X. Li, J. E. Knox, H. P.
Hratchian, J. B. Cross, C. Adamo, J. Jaramillo, R. Gomperts,
R. E. Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli,
J. W. Ochterski, P. Y. Ayala, K. Morokuma, G. A. Voth, P.
Salvador, J. J. Dannenberg, V. G. Zakrzewski, S. Dapprich, A. D.
Daniels, M. C. Strain, O. Farkas, D. K. Malick, A. D. Rabuck, K.
Raghavachari, J. B. Foresman, J. V. Ortiz, Q. Cui, A. G. Baboul,
S. Clifford, J. Cioslowski, B. B. Stefanov, G. Liu, A. Liashenko, P.
Piskorz, I. Komaromi, R. L. Martin, D. J. Fox, T. Keith, M. A.
Al-Laham, C. Y. Peng, A. Nanayakkara, M. Challacombe,
P. M. W. Gill, B. Johnson, W. Chen, M. W. Wong, C. Gonzalez,
J. A. Pople, Gaussian, Inc., Pittsburgh, PA, 2003; b) A. D. Becke,
Phys. Rev. A 1988, 38, 3098; A. D. Becke, J. Chem. Phys. 1993,
98, 5648; c) J. C. Slater, Quantum Theory of Molecules and
Solids, McGraw-Hill, New York, 1974; d) J. P. Perdew, J. A.
Chevary, S. H. Vosko, K. A. Jackson, M. R. Pederson, D. J.
Singh, C. Fiolhais, Phys. Rev. B 1992, 46, 6671; e) P. J. Hay, W. R.
Wadt, J. Chem. Phys. 1985, 82, 299; f) A. E. Reed, L. A. Curtiss,
F. Weinhold, Chem. Rev. 1988, 88, 899.
[7] a) Hydrogenolysis of the iridium nitrogen bond is a key-step in
the catalyzed hydrogenation of imines see: R. Dorta, D.
Broggini, R. Kissner, A. Togni, Chem. Eur. J. 2004, 10, 4546,
and references therein. b) See also: Y. Ng, C. Chan, J. A. Osborn,
J. Am. Chem. Soc. 1990, 112, 9400.
[8] Intermediates in the related catalytic cycle for the transfer
hydrogenation of ketones with ruthenium arene complexes
could be isolated and crystallized, see: K.-J. Haack, S. Hashi-
guchi, A. Fujii, T. Ikariya, R. Noyori, Angew. Chem. 1997, 109,
297; Angew. Chem. Int. Ed. Engl. 1997, 36, 285.
[9] P. Maire, F. Breher, H. Schönberg, H. Grützmacher, Organo-
metallics 2005, 24, 3207.
[10] a) T. Büttner, F. Breher, H. Grützmacher, Chem. Commun. 2004,
2820; b) T. Büttner, J. Geier, G. Frison, J. Harmer, C. Calle, A.
Schweiger, H. Schönberg, H. Grützmacher, Science 2005, 307,
235.
[11] Crystal structure of 5b: C49H39NPRh, monoclinic, space group
P2(1)/c; a = 11.897(1), b = 14.932(1), c = 20.862(1) Š, b =
104.046(1)8; V= 3595.3(2) Š3; Z = 4; 1calcd = 1.433 Mgmꢀ3; crys-
tal dimensions 0.47 ” 0.31 ” 0.25 mm; diffractometer Bruker
SMART Apex; MoKa radiation, 200 K, 2Vmax = 80.168; 97685
reflections, 22235 independent (Rint = 0.0537), direct methods;
empirical absorption correction SADABS (version 2.03); refine-
ment against full matrix (versus F2) with SHELXTL (version
6.12) and SHELXL-97,[18] 471 parameters, R1 = 0.0560 and wR2
(all data) = 0.1471, max./min. residual electron density 1.900/
ꢀ1.523eŠ ꢀ3. The hydrogen atoms were placed in idealized
positions and included as riding atoms. Crystal structure of 6b:
¯
C49H41NPRh·2THF, triclinic, space group P1; a = 10.935(1), b =
11.864(1), c = 19.658(1) Š, a = 91.408(3)8, b = 100.019(3)8, g =
109.948(3); V= 2350.8(3) Š3; Z = 2; 1calcd = 1.301 Mgmꢀ3; crys-
tal dimensions 0.62 ” 0.53” 0.43mm; diffractometer Bruker
SMART Apex; MoKa radiation, 200 K, 2Vmax = 82.468; 61525
reflections, 29179 independent (Rint = 0.0239), direct methods;
empirical absorption correction SADABS (ver. 2.03); refine-
ment against full matrix (versus F2) with SHELXTL (ver. 6.12)
and SHELXL-97,[18] 471 parameters, R1 = 0.0471 and wR2 (all
data) = 0.1465, max./min. residual electron density 2.757/
ꢀ1.422 eŠꢀ3. The hydrogen atoms were placed in idealized
positions and included as riding atoms. The positions of the
hydrogen atoms H1 (NH) and H2 (RhH) were obtained from a
difference Fourier synthesis, they were refined isotropically. The
para-methyl moiety of the PPh2Tol ligand was “disordered” due
to phenyl/tolyl interchange and had to be split over two
positions, which were refined against each other (FVAR =
0.47). Consequently, one hydrogen atom in the para-position
of a phenyl ring was not included in the refinement. CCDC-
264482 (5b) and CCDC-264483( 6b) 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.
[16] We could not locate any minimum structure with a h2-bonded H2
molecule as ligand. On the contrary, the heterolytic splitting of
H2 by ruthenium(ii) phosphane amides is a two-step process
including first the formation of a s-bonded H2 complex followed
by an intramolecular proton transfer to the amide nitrogen atom
(energy barrier (DFT) from the amide to the transition state for
H2 cleavage: 13.4 kcalmolꢀ1 [7b]
) A one-step heterolytic cleavi-
.
age of H2 has been calculated for ruthenium(ii) arene amides
with a transition state at 25.2 kcalmolꢀ1 (MP2): M. Yamakawa,
H. Ito, R. Noyori, J. Am. Chem. Soc. 2000, 122, 1466.
[12] a) M. Ogasawara, S. A. Macgregor, W. E. Streib, K. Folting, O.
Eisenstein, K. G. Caulton, J. Am. Chem. Soc. 1996, 118, 10189;
b) T. Gottschalk-Gaudig, J. C. Huffman, K. G. Caulton, H.
GØrard, O. Eisenstein, J. Am. Chem. Soc. 1999, 121, 3 242; c) T.
=
[17] The complete catalytic cycle for the hydrogenation of H2C O
with I was calculated. The results strongly support the assump-
tion that this reaction follows the metal–ligand bifunctional
6322
www.angewandte.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2005, 44, 6318 –6323

Angewandte
Chemie
mechanism and that the heterolytic H2-splitting is the rate-
ing the calculation of the catalytic cycle are given in the
determining step. The transfer of Hd+, Hdꢀ to the H2C O
=
Supporting Information.
molecule, which is bonded in the second coordination sphere at
the place where the THF molecule is in 6b, has only a very small
barrier (Ea = 0.5 kcalmolꢀ1). However, our experimental set-up
did not allow the stirring of the reaction mixtures, which should
enhance considerably the turnover frequency. Details concern-
[18] SHELXTL, V. 6.12, Bruker AXS, Madison, Wisconsin, USA,
2002; G. M. Sheldrick, SHELXL-97, Program for the Refine-
ment of Crystal Structures, University of Göttingen, Germany,
1997.
Angew. Chem. Int. Ed. 2005, 44, 6318 –6323
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
6323