Metallacarbenes from diazoalkanes: An experimental and computational study of the reaction mechanism

Article abstract of DOI:10.1021/ja028923c

PCP ligand (1,3-bis-[(diisopropyl-phosphanyl)-methyl]-benzene), and PCN ligand ({3-[(di-tert-butyl-phosphanyl)-methyl]-benzyl}-diethyl-amine) based rhodium dinitrogen complexes (1 and 2, respectively) react with phenyl diazomethane at room temperature to give PCP and PCN-Rh carbene complexes (3 and 5, respectively). At low temperature (-70 °C), PCP and PCN phenyl diazomethane complexes (4 and 6, respectively) are formed upon addition of phenyl diazomethane to 1 and 2. In these complexes, the diazo moiety is η¹ coordinated through the terminal nitrogen atom. Decomposition of complexes 4 and 6 at low temperatures leads only to a relatively small amount of the corresponding carbene complexes, the major products of decomposition being the dinitrogen complexes 1 and 2 and stilbene. This and competition experiments (decomposition of 6 in the presence of 1) suggests that phenyl diazomethane can dissociate under the reaction conditions and attack the metal center through the diazo carbon producing a η¹-C bound diazo complex. Computational studies based on a two-layer ONIOM model, using the mPW1 K exchange-correlation functional and a variety of basis sets for PCP based systems, provide mechanistic insight. In the case of less bulky PCP ligand bearing H-substituents on the phosphines, a variety of mechanisms are possible, including both dissociative and nondissociative pathways. On the other hand, in the case of i-Pr substituents, the η¹-C bound diazo complex appears to be a critical intermediate for carbene complex formation, in good agreement with the experimental results. Our results and the analysis of reported data suggest that the outcome of the reaction between a diazoalkane and a late transition metal complex can be anticipated considering steric requirements relevant to η¹-C diazo complex formation.

Full text of DOI:10.1021/ja028923c

Published on Web 05/01/2003
Metallacarbenes from Diazoalkanes: An Experimental and
Computational Study of the Reaction Mechanism
Revital Cohen, Boris Rybtchinski, Mark Gandelman, Haim Rozenberg,
Jan M. L. Martin,* and David Milstein*
Contribution from the Department of Organic Chemistry, The Weizmann Institute of Science,
76100 RehoVot, Israel
Received October 13, 2002; E-mail: david.milstein@weizmann.ac.il; comartin@wicc.weizmann.ac.il
Abstract: PCP ligand (1,3-bis-[(diisopropyl-phosphanyl)-methyl]-benzene), and PCN ligand ({3-[(di-tert-
butyl-phosphanyl)-methyl]-benzyl}-diethyl-amine) based rhodium dinitrogen complexes (1 and 2, respec-
tively) react with phenyl diazomethane at room temperature to give PCP and PCN-Rh carbene complexes
(3 and 5, respectively). At low temperature (-70 °C), PCP and PCN phenyl diazomethane complexes (4
and 6, respectively) are formed upon addition of phenyl diazomethane to 1 and 2. In these complexes, the
diazo moiety is η1 coordinated through the terminal nitrogen atom. Decomposition of complexes 4 and 6
at low temperatures leads only to a relatively small amount of the corresponding carbene complexes, the
major products of decomposition being the dinitrogen complexes 1 and 2 and stilbene. This and competition
experiments (decomposition of 6 in the presence of 1) suggests that phenyl diazomethane can dissociate
under the reaction conditions and attack the metal center through the diazo carbon producing a η1-C bound
diazo complex. Computational studies based on a two-layer ONIOM model, using the mPW1K exchange-
correlation functional and a variety of basis sets for PCP based systems, provide mechanistic insight. In
the case of less bulky PCP ligand bearing H-substituents on the phosphines, a variety of mechanisms are
possible, including both dissociative and nondissociative pathways. On the other hand, in the case of i-Pr
substituents, the η1-C bound diazo complex appears to be a critical intermediate for carbene complex
formation, in good agreement with the experimental results. Our results and the analysis of reported data
suggest that the outcome of the reaction between a diazoalkane and a late transition metal complex can
be anticipated considering steric requirements relevant to η1-C diazo complex formation.
Introduction
the reactive intermediates in both cyclopropanation and C-H
bond activation.6,7
Diazoalkanes represent a very important class of reagents in
organometallic chemistry1-3 regarding metallacarbene synthe-
sis,1,4,5 cyclopropanation,6 and C-H bond activation.6, 7 Metal-
mediated synthesis of reactive I, N, S, Se, and O ylides using
diazoalkanes is also of significant synthetic value.8
Transition metal complex mediated cyclopropanation of
olefins using diazoesters is among the best methods for the
synthesis of cyclopropane derivatives.6 Computational studies
regarding various aspects of diazoalkane reactivity in transition
metal chemistry have been reported.9 Intramolecular C-H bond
activation in diazoesters promoted by Rh(II) complexes has been
utilized for the construction of medium size rings.6 Recently,
catalytic intermolecular C-H activation of alkanes with diazo-
acetates has been reported.7 Metallacarbenes are thought to be
Two important early discoveries introduced diazo compounds
into synthetic organometallic chemistry as useful reagents for
metal carbene synthesis. Herrmann utilized diazo compounds
as versatile carbene synthons in the case of unsaturated
manganese complexes,10 and Roper employed diazomethane in
the synthesis of Ru and Os carbenes, which were found to be
stable, isolable compounds.4,11 Nowadays, the synthesis of late
transition metal carbenes relies extensively on the use of
diazoalkanes. Grubbs et al. have synthesized a variety of Ru
carbenes using diazoalkanes,12 including (PCy3)2Cl2RudCH-
(9) For recent theoretical works, see: (a) A review on transition metal-main
group multiple bonding: Cundari, T. R. Chem. ReV. 2000, 100, 807. (b)
Diazoalkanes in the metal catalyzed formation of carbonyl and ammonium
ylides: Padwa, A.; Snyder, J. P.; Curtis, E. A.; Sheehan, S. M.; Worsencroft,
K. J.; Kappe, C. O. J. Am. Chem. Soc. 2000, 122, 8155. (c) Cyclopropa-
nation: Bernardi, F.; Bottoni, A.; Miscione, G. P. Organometallics 2001,
20, 2751; Rodriguez-Garcia, C.; Oliva, A.; Ortuno, R.; Branchadell, V. J.
Am. Chem. Soc. 2001, 123, 6157. (d) C-H bond activation: Nakamura,
E.; Yoshikai, M.; Yamanaka, M. J. Am. Chem. Soc. 2002, 124, 7181.
(10) (a) Herrmann, W. A. Angew. Chem., Int. Ed. Engl. 1974, 13, 639. (b)
Herrmann, W. A. Chem. Ber. 1975, 108, 486.
(1) Herrmann, W. A. Angew. Chem., Int. Ed. Engl. 1978, 17, 800.
(2) Mizobe, Y.; Ishii, Y.; Hidai, M. Coord. Chem. ReV. 1995, 139, 281.
(3) Putala, M.; Lemenovskii, D. A. Russ. Chem. ReV. 1994, 63, 197.
(4) Roper, W. R. J. Organomet. Chem. 1986, 300, 167.
(5) Werner, H. J. Organomet. Chem. 1995, 500, 331.
(6) (a) Padwa, A.; Weingarten, M. D. Chem. ReV. 1996, 96, 223. (b) Doyle,
M. P.; Forbes, D. C. Chem. ReV. 1998, 98, 911. (c) Davies, H. M. L.;
Panaro, S. A. Tetrahedron 2000, 56, 4871.
(11) Hill, A. F.; Roper, W. R.; Waters, J. M.; Wright, A. H. J. Am. Chem. Soc.
1983, 105, 3939.
(7) For a recent review on intermolecular C-H activation see: Davies, H. M.
L.; Antoulinakis, E. G. J. Organomet. Chem. 2001, 617-618, 47.
(8) For a recent review, see: Hodgson, D. M.; Pierard, F. Y. T. M.; Stupple,
P. A. Chem. Soc. ReV. 2001, 30, 50.
(12) (a) Schwab, P.; France, M. B.; Ziller, J. W.; Grubbs, R. H. Angew. Chem.,
Int. Ed. Engl. 1995, 34, 2039. (b) Schwab, P.; Grubbs, R. H.; Ziller, J. W.
J. Am. Chem. Soc. 1996, 118, 100. (c) Sanford, M. S.; Valdez, M. R.;
Grubbs, R. H. Organometallics 2001, 20, 5455.
9
6532
J. AM. CHEM. SOC. 2003, 125, 6532-6546
10.1021/ja028923c CCC: $25.00 © 2003 American Chemical Society

Metallacarbenes from Diazoalkanes
A R T I C L E S
No examples of an η2 coordination mode through the C-N
bond as well as an η2-(CN) cycloadduct have yet been reported.
As described before, reactions of diazoalkanes with metal
complexes often result in the formation of metallacarbenes and
liberation of N2. As a result, only a few examples where well-
defined diazoalkane complexes undergo clean N2 loss to give
carbene complexes are known. In a review on the synthesis and
reactivity of diazoalkane complexes, Hidai pointed out that in
the course of carbene formation, “the coordination of a
diazoalkane to a metal complex precedes the N2 loss from the
diazoalkane, and it is of great importance to elucidate how the
N2 extrusion from the diazoalkane ligands proceeds in these
complexes”.2 In general, the reaction aptitudes of diazoalkanes
toward metal centers are not well understood, and the intimate
mechanism of the carbene formation from the diazo compounds
has not been investigated experimentally. Factors controlling
formation of metal carbenes from diazoalkanes are not well
understood and the rational choice of ligands, diazoalkanes, and
reaction conditions remains a significant challenge in metal-
lacarbene synthesis.
Here, we report an experimental and computational study on
metallacarbene formation using a diazoalkane (phenyl diazo-
methane) as a carbene donor. The phenyl diazomethane complex
observed at the initial stages of reaction appears to be a side
equilibrium compound rather than a true intermediate in the
process of the carbene complex formation. We show that the
key intermediate on the pathway of metallacarbene formation
is a complex with a diazoalkane carbon atom (η1-C) rather than
a nitrogen atom bound to the metal center. We suggest that the
outcome of the reaction between a diazoalkane and a late
transition metal complex can be anticipated using simple
electronic and steric arguments, and propose a generalization
of the known reactivity patterns.
Scheme 1
(Ph) which was synthesized by using phenyl diazomethane.13
This compound and its derivatives are widely used as very
efficient metathesis catalysts in organic synthesis and polymer
science. In seminal work on the synthesis of late transition metal
carbenes by Werner, diazoalkanes are also widely utilized as
carbene precursors. 5,14,15 It was demonstrated that careful choice
of the ligands and the diazo compounds is critical for the carbene
synthesis.5,14 In some catalytic cyclopropanation systems utiliz-
ing diazo compounds, various carbene complexes have been
isolated.16 We have recently reported the synthesis of a PCP-
Rh carbene complex using PhCHN2.17
Metallacarbenes are not always the products of the reaction
of a transition metal complex with a diazo compound. A
significant number of stable diazoalkane complexes have been
isolated in the reaction of certain transition metal complexes
with diazoalkanes.2,18 The coordination modes of diazoalkanes,
displayed at mono- and multimetallic centers, are quite diverse.
Monometallic complexes reported to date include either η2
-diazoalkane ligands bound through the N-N multiple bonds
(η2-NN)2,18,19 or η1 -diazoalkane ligands coordinated at the
terminal nitrogen atom (η1-N)2,5,18,20 (Scheme 1). Other possible
coordination modes of diazoalkane ligands include η2-binding
through the C-N bond (η2-CN), the four-membered ring
cycloadduct (η2-(CN)), and coordination through the carbon
atom (η1-C). However, unambiguous examples of these types
are yet unknown, although complexes of the type η1-C are
frequently assumed to be the initial intermediates preceding
metallacarbene species formation in the transition-metal-
catalyzed carbenoid reactions using diazoalkanes.21
Results
A. Experimental Results. Continuing our studies on PCP22
and PCN-type23 complexes we became interested in the mech-
anism of carbene complex formation from the corresponding
PCP and PCN type rhodium dinitrogen complexes and PhCHN2.
The rigid chelating PCP and PCN frameworks enhances
complex stability and introduces a range of special properties,
providing models for mechanistic studies and efficient cata-
lysts.24
PCP and PCN-Rh Dinitrogen Complexes. The PCP
dinitrogen complex 1 was synthesized as described in Scheme
2. The analogous PCN complex 2 was prepared by reacting the
corresponding methyl chloride complex with NaHBEt3 (see the
Experimental section). These complexes exist as mixtures of
monomer (1a, 2a) and dimer (1b, 2b) species, in a ratio which
depends on the concentration of the complexes and N2. Thus,
bubbling with argon or freeze-pump-thaw cycles accompanied
(13) For recent reviews, see: (a) Trnka, T. M.; Grubbs R. H. Acc. Chem. Res.
2001, 34, 18. (b) Fu¨rstner, A. Angew. Chem., Int. Ed. Engl. 2000, 39, 3012.
(14) (a) Werner, H.; Schwab, P.; Bleuel, E.; Mahr, N.; Steinert, P.; Wolf, Chem.
Eur. J. 1997, 3, 1375. (b) Ortmann, D. A.; Weberndorfer, B.; Ilg, K.;
Laubender, M.; Werner, H. Organometallics 2002, 21, 2369.
(15) (a) Weberndo¨rfer, B.; Henig, G.; Werner, H. Organometallics 2000, 19,
4687. (b) Werner, H.; Stu¨er, W.; Wolf, J.; Laubender, M.; Weberndo¨rfer,
B.; Herbst-Irmer, R.; Lehmann C. Chem. Eur. J. 1999, 11, 1889. (c) Werner,
H.; Stu¨er, W.; Weberndo¨rfer, B.; Wolf, J. Chem. Eur. J. 1999, 10, 1707.
(16) For recent examples, see: (a) Galardon, E.; Le Maux, P.; Toupet, L.;
Simonneaux G. Organometallics 1998, 17, 565. (b) Lee, H. M.; Bianchini,
C.; Jia, G.; Barbaro, P. Organometallics 1999, 18, 1961.
(17) Vigalok, A.; Milstein, D. Organometallics 2000, 19, 2061.
(18) (a) Sutton, D. Chem. ReV. 1993, 93, 995. (b) Dartiguenave, M.; Menu, M.
J.; Deydier, E.; Dartiguenave, Y.; Siebald, H. Coord. Chem. ReV. 1998,
178-180, 623.
(19) (a) Polse, J. L.; Kaplan, A. W.; Andersen, R. A.; Bergman, R. G. J. Am.
Chem. Soc. 1998, 120, 6316. (b) Polse, J. L.; Andersen, R. A.; Bergman,
R. G. J. Am. Chem. Soc. 1996, 118, 8737. (c) Kaplan, A. W.; Polse, J. L.;
Ball, G. E.; Andersen, R. A.; Bergman, R. G. J. Am. Chem. Soc. 1998,
120, 11 649.
(20) Albertin, G.; Antoniutti, S.; Bordignon, E.; Carrera, B. Inorg. Chem. 2000,
39, 4646.
(21) (a) Maxwell, J. L.; Brown, K. C.; Bartley, D.; Kodadek T. Science 1992,
256, 1544. (b) Bartley, D.; Kodadek T. J. Am. Chem. Soc. 1993, 115, 1656.
(22) For example: (a) Ohff, M.; Ohff, A.; van der Boom, M. E.; Milstein, D.
J. Am. Chem. Soc. 1997, 119, 11 687. (b) Vigalok, A.; Uzan, O.; Shimon,
L. J. W.; Ben-David, Y.; Martin, J. M. L.; Milstein, D. J. Am. Chem. Soc.
1998, 120, 12 539. (c) Cohen, R.; van der Boom, M. E.; Shimon, L. J. W.;
Rozenberg, H.; Milstein, D. J. Am. Chem. Soc. 2000, 122, 7723.
(23) (a) Gandelman, M.; Vigalok, A.;. Shimon, L. J. W.; Milstein, D.
Organometallics 1997, 16, 3981. (b) Gandelman, M.; Vigalok, A.;.
Konstantinovski, L.; Milstein, D. J. Am. Chem. Soc. 2000, 122, 9848. (c)
Gandelman, M.; Milstein, D. Chem. Commun. 2000, 1603.
(24) Reviews: (a) Rybtchinski, B.; Milstein, D. Angew. Chem., Int. Ed. Engl.
1999, 38, 870. (b) Jensen, C. M. Chem. Comm. 1999, 2443. (c) Albrecht,
M.; van Koten, G. Angew. Chem., Int. Ed. Engl. 2001, 40, 3750. (d) Vigalok
A, Milstein D. Acc. Chem. Res. 2001, 34, 798.
9
J. AM. CHEM. SOC. VOL. 125, NO. 21, 2003 6533

A R T I C L E S
Cohen et al.
Scheme 2
Figure 1. ORTEP plot of the X-ray structure of 1b. Hydrogen atoms are
omitted for clarity. Selected bond lengths (Å) and bond angles (degrees):
N(1)-N(1A) ) 1.108(3), Rh(1)-N(1) ) 1.9687(17), Rh(1)-C(4) ) 2.039-
(2), Rh(1)-P(2) ) 2.2680(9), Rh(1)-P(3) ) 2.2688(8); Rh(1)-N(1)-
N(1A) ) 177.3(2), C(4)-Rh(1)-N(1) ) 178.18(7), P(2)-Rh(1)-N(1) )
95.25(5), P(3)-Rh(1)-N(1) ) 98.70(5), P(2)-Rh(1)-P(3) ) 165.82(2),
C(4)-Rh(1)-P(2) ) 83.10(6), C(4)-Rh(1)-P(3) ) 82.92(7); Rh(1)-
N(1)-N(1A)-Rh(1A) ) -147(4); C(4)-Rh(1)-N(1)-N(1A) ) -11(4).
with argon refill resulted in almost exclusive formation of the
dimers 1b and 2b. At the typical concentrations of the
experiment (∼0.4M) under 1atm. N2 at room temperature, the
following ratios are observed: 1a:1b ) 8:3, 2a:2b ) 3:7; at
-50 °C 1a:1b ) 1:1, 2a:2b ) 1:2. These ratios are most
probably determined by a balance of steric and electronic factors.
Similar dimer-monomer equilibria in the case of various
dinitrogen complexes have been reported.25,26 For the sake of
simplicity, we depict complexes 1 and 2 as the monomers in
all the schemes, although dimer species are present. Crystals
of 1b suitable for X-ray analysis were grown at -35 °C from
a toluene-pentane solution of 1a,b. Only 1b crystallizes under
these conditions. The X-ray structure of 1b reveals that the
dinitrogen molecule is bridging between two square planar
PCP-Rh units, which are twisted around the RhNNRh axis with
a dihedral angle of 33(4)° (Figure 1). The coordinated dinitrogen
molecule is not activated, the N-N bond length (1.108(3) Å)
being close to the one in free N2.
Scheme 3
PCP System: Carbene and Diazoalkane Complex Forma-
tion. When complex 1 is reacted at room temperature with 1
equiv of PhCHN2 in benzene, N2 evolution takes place and
formation of the carbene complex 3 is observed (Scheme 3).
Preparation of this complex using a sulfur ylide precursor has
been recently reported by us.27 When an equimolar amount of
PhCHN2 is added to complex 1 at -70 °C in toluene the dark
green diazo complex 4 is formed quantitatively. It was
characterized by multinuclear NMR at -60 °C. Complex 4 gives
rise to a doublet at 66.46 ppm (1JRhP ) 154.4 Hz) in 31P{1H}
NMR. The Rh atom is bound to the aryl ring of the ligand as
heteronuclear 2D 15N-1H correlation experiment. In the cor-
relation spectrum, two 15N cross-peaks (related to a PhCHN2
hydrogen signal) appear: a doublet centered at 421.77 ppm
(1JRhN ) 31.5 Hz, Rh-NtNCHPh), and a singlet at 300.03 ppm
(Rh-NtNCHPh).29 Surprisingly, when heated to 30 °C,
complex 4 does not convert cleanly to the rhodium carbene 3,
but gives only 30% of 3, the other products being the Rh
nitrogen complexes 1 and stilbene. Formation of the carbene
complex was observed to start at -50 °C, the process being
very slow. It should be noted that when an excess of PhCHN2
is added to the carbene complex at -70 °C and the solution
slowly warmed, stilbene formation starts to take place at -50
°C (the temperature at which the carbene formation was
obserVed to begin), concurrently with formation of the dinitrogen
complex. Stilbene can be formed also in the process of carbene
decompostion (probably by dimerization); however, at room
temperature this process was observed to proceed for hours and
does not occur at all at -50 °C. Thus, stilbene formation can
be explained by the reaction of the diazoalkane with the carbene
evident from the splitting pattern of the ipso carbon in the 13
C
NMR spectrum, which appears as a doublet of triplets centered
at 173.02 ppm (1JRhC ) 38.3 Hz, 2JPC ) 4.9 Hz). The signal of
the PhCHN2 group appears in the 1H NMR spectrum as a singlet
centered at 3.76 ppm, the chemical shift being similar to the
reported one (4.13 ppm) for a PhCHN2-Rh complex.28 In the
13C NMR spectrum, the coordinated PhCHN2 gives rise to a
singlet at 57.95 ppm, ruling out the η1-C coordination mode
(see Scheme 1). Coordination through the terminal N atom (end-
on η1-N mode) of the diazoalkane was confirmed by a
(25) (a) van der Boom, M. E.; Liou, Sh.-Y.; Ben-David, Y.; Shimon, L. J. W.;
Milstein D. J. Am. Chem. Soc. 1998, 120, 6531. (b) Gusev, D. G.;
Dolgushin, F. M.; Antipin, M. Yu. Organometallics 2000, 19, 3429.
(26) Yoshida, T.; Okano, T.; Thorn, D. L.; Tulip, T. H.; Otsuka S.; Ibers, J. A.
J. Organomet. Chem. 1979, 181, 183.
(27) Gandelman, M.; Rybtchinski, B.; Ashkenazi, N.; Gauvin, R.; Milstein D.
J. Am. Chem. Soc. 2001, 123, 5372.
(28) Werner, H.; Mahr, N.; Frenking, G.; Jonas, V. Organometallics 1995, 14,
619.
(29) For the 15N NMR data of the diazo compounds, see: Regitz, M.; Maas, G.
Diazo Compounds Properties and Synthesis; Academic Press Inc.: New
York, 1986, pp 47-56.
9
6534 J. AM. CHEM. SOC. VOL. 125, NO. 21, 2003

Metallacarbenes from Diazoalkanes
A R T I C L E S
Scheme 4
case of 6. To explore the above-mentioned pathways (I-III),
we performed competition experiments, designed to probe the
possibility of diazoalkane detachment from the metal center.
Competition Experiments. Complex 6 was obtained at -78
°C, as described above, and the dinitrogen PCP complex 1 was
added to it at -78 °C (Scheme 6). The solution containing a
mixture of 6 and 1 (6:1 ) 1:2) was placed in an NMR probe
pre-cooled to -70 °C. No reaction took place at this temperature,
the ratio remaining the same for hours. When heated to -50
°C slow formation of both PCP and PCN carbene complexes 3
and 5 was observed. Keeping the solution overnight at -50 °C
resulted in a mixture containing the carbene complexes 3 and
5, the PCP dinitrogen complex 1 and the PCN dinitrogen
complex 2 (about 50% conversion, 3:5 ) 6:5). A small amount
of the PCP diazo complex 4 (3-4%) was also observed.
In another experiment, a mixture of 1 and 6 (6:1 ) 1:2) was
prepared at -78 °C in toluene and placed in an NMR probe
pre-cooled to -70 °C. It was then heated to 0 °C, upon which
immediate formation of the carbene complexes 3 and 5 took
place, resulting in a ratio 3:5 ) 3:4 as observed by 31P and 1H
NMR. This ratio changed only slightly at 0 °C for hours, due
to slow carbene decomposition to produce the corresponding
dinitrogen complexes and stilbene. In the experiments described
above no carbene transfer from 5 to 1 took place neither at -50
°C nor at 0 °C.
complex. We suggest that as the rate of formation of the carbene
complex 3 is decreased at low temperatures and the lifetime of
the diazoalkane complex 4 is increased, the concentration of
free PhCHN2 formed by dissociation from 4 is sufficient for
significant carbene consumption. An intermolecular reaction
between bound diazoalkane and the carbene complex is also a
possibility, although it seems less probable due to steric factors
(see below).
Thus, there is a striking difference between the reaction of
PhCHN2 with 1 at room temperature and at low temperatures.
The end-on η1-N diazo complex decomposition does not lead
to clean carbene formation, in line with previous observations
that diazo complexes may resist N2 loss and hamper carbene
formation.2,3
Thus, formation of both PCP and PCN carbenes 3 and 5
(which do not interconvert at the given temperature and do not
undergo carbene transfer) starting from the PCN diazoalkane
complex 6 suggests that the diazoalkane can detach from the
metal center (pathway II). The possibility of the bimolecular
mechanism (pathway III, see Scheme 5), that can lead to the
same observations as for the pathway II, cannot be ruled out.
However, the following reasons suggest that it is less probable
than II: (1) decomposition of 4 was observed to start at the
same facility independent of its concentration (inconsistent with
pathway IIIA); (2) decomposition of 6 is not increased
significantly upon addition of a 2-fold access of the dinitrogen
complex 1 (inconsistent with pathways IIIB and IIIC); (3) a
small amount of the PCP diazo complex 4 is formed at low
temperatures upon mixing 6 and 1, indicating that PhCHN2
dissociates from 6 and is trapped by 1. For further clarification
concerning pathway III see the Computational Results section.
B. Computational Results. To clarify the reaction mecha-
nism, Density Functional Theory (DFT) calculations were
performed. Calculations were carried out first for the PCP
system with H substituents on the phosphines (H-PCP system)
and CH2N2 instead of PhCHN2 in order to reduce the computa-
tion time. Afterward, the calculations were performed for the
H-PCP-Rh/PhCHN2 system to identify plausible reaction
pathways. Finally, iPr groups were introduced in an ONIOM30
approach to obtain a computational description (iPr-PCP-Rh/
PhCHN2 system) close to the actual experimental system.
Search for Plausible Reaction PathwayssH-PCP-Rh/
PhCHN2 System. The system was initially investigated (ge-
ometry optimization) at the mPW1k/SDD level.31 To further
improve the accuracy of the calculated profile, single-point-
energies at this level of reference geometry were calculated at
the mPW1k/SDB-cc-pVDZ level. Finally, in the preferred
PCN System: Carbene and Diazo Complex Formation.
The PCN-Rh carbene 5 was quantitatively formed by reaction
of the dinitrogen complex 2 with PhCHN2 at room temperature,
analogously to formation of the PCP-Rh complex 3 (Scheme
4). Preparation of complex 5 by the sulfur ylide method was
recently reported by us.27 Analogously to the PCP system, when
an equimolar amount of PhCHN2 is added to complex 2 at -70
°C in toluene, the dark green diazoalkane complex 6 is formed
quantitatively. Complex 6 gives rise to a doublet at 88.90 ppm
(1JRhP ) 204.6 Hz) in 31P{1H} NMR. The signals of the PhCHN2
group appear in the 1H NMR spectrum as a singlet centered at
3.84 ppm and in the 13C NMR as a singlet at 68.89 ppm.
Formation of the carbene complex 5 from 6 was observed to
start at -50 °C. At this temperature and at higher temperatures
complex 6 decomposes to give the dinitrogen complex 2 along
with unidentified products, and only a small amount of the
carbene complex 5 is obserVed, in a sharp contrast to the reaction
at room temperature. Stilbene formation along with unidentified
organic products was observed.
Possible Pathways for Carbene Complex Formation. Three
distinct pathways for carbene complex formation can be
envisaged (Scheme 5). In pathway I, (the intramolecular “bound
diazo” mechanism), the end-on diazo transforms into the carbene
complex without detachment, i.e., it is bound to the metal center
throughout the reaction sequence. In pathway II (the intermo-
lecular “direct attack” mechanism), PhCHN2 dissociates from
the metal center and then attacks it via the diazo carbon. In
pathway III (the “bimolecular” mechanism), bound diazoalkane
reacts with the metal center of another complex which could
be a complex of diazoalkane, dinitrogen, or a T-shaped 14e
intermediate. Unfortunately, the kinetic follow-up of the de-
composition of the diazoalkane complexes 4 and 6 was
hampered by a significant overlap between the signals of 1 and
(30) Dapprich, S.; Komaromi, I.; Byun, K. S.; Morokuma, K.; Frisch, M. J. J.
Mol. Struct. (THEOCHEM) 1999, 461, 1 and references therein.
1
4 in 31P{1H} and H NMR and by an unclean reaction in the
9
J. AM. CHEM. SOC. VOL. 125, NO. 21, 2003 6535

A R T I C L E S
Cohen et al.
Scheme 5
reaction profiles all the stationary points were optimized at the
mPW1k/SDB-cc-pVDZ level for consistency. All of the energies
reported are relative to the starting complex 1H (Scheme 7)
unless indicated otherwise. A summary of the computational
results is given in Table 1.
On the basis of the relative free energy, the diazo coordination
(1Hf3H) can take place both by associative and dissociative
mechanisms, the latter being preferable at room temperature
(vide infra). the observed formation of η1-N diazo complex 4
prior to the carbene formation (see Scheme 3) may suggest that
the diazo into carbene transformation does not involve detach-
ment of the diazoalkane throughout the carbene formation
sequence (bound diazo mechanism, pathway I in Scheme 5).
Thus, both “bound diazo” and “direct attack” mechanisms were
explored computationally, and the following pathways were
found (see Scheme 7). Following the formation of the η1-N diazo
complex 2H from 1H, the reaction sequence proceeds by η1-N
into η2-NN transformation (3Hf4H) with a kinetic barrier of
13.4 kcal/mol. Then there are two alternative pathways for
formation of the carbene complex, while keeping the diazo
molecule bound to the metal center. In the first one, the diazirine
intermediate 5H is formed and converted into an intermediate
6H with a chelating four-membered ring (η2-(CN)), which in
turn yields η1-C intermediate 7H (Figure 2). An additional
pathway is represented by a cleavage of a C-N bond in the
four-membered ring of 6H to give carbene nitrogen complex
8H that loses N2 to form the carbene complex 9H. The latter
path is kinetically more favorable. Thus, the barrier for breaking
the N-C bond in 6H to form 8H is 4.4 kcal/mol. Furthermore,
N2 loss from 8H to form the carbene complex 9H appears to
(31) Except of the geometry of the intermediate 7H, which was optimized at
the mPW1k/SDB-cc-pVDZ level since no such stationary point was found
at the mPW1k/SDD level. Instead, only intermediate 7Hi, which is a η2-
CN complex, was found at this level (Table 1). This could be justified by
the fact that complexes (7H and 7Hi) have comparable energies, and they
are closely related geometrically, thus only one of them was located in a
given basis set. In the larger basis set (SDB-cc-pVDZ), however, only 7H
was located. This result supports the initial belief that 7H is the intermediate
involved in the mechanism.
9
6536 J. AM. CHEM. SOC. VOL. 125, NO. 21, 2003

Metallacarbenes from Diazoalkanes
A R T I C L E S
Scheme 6
Table 1. Relative Energies (∆Ee) and Free Energies (∆G298) of
i
the Various Complexes of the H-PCP and Pr-PCP Systems
(kcal/mol)
complex
∆E
e
∆G
298
H-PCP System
(mPWlk/SDB-cc-pVDZ//mPWlk/SDD)
(H-PCP)RhN2 (1H)
0.0
23.1
0.0
12.8
5.9
47.2
20.6
6.2
26.7
13.6
25.6
19.3
6.1a
10.8
17.6
-6.3
-7.1
-6.9
11.1
15.6
-14.3
0.0
11.5
-0.1
13.3
7.6
46.3
18.8
7.7
28.0
15.6
25.6
20.0
4.0a
12.4
16.5
-6.3
-10.2
-18.8
19.7
25.3
-0.6
(H-PCP)Rh (2H)
(H-PCP)Rh-NdNdCHPh (3H)
TS-Phdiazo-to-Eta2 TS(3H-4H)
(H-PCP)PhDiazo-NN-eta2 (4H)
TS-PhDiazoEta2 to Diazirine TS(4H-5H)
TS-eta2NNtoEtalCdiazo TS(4H-7H)
(H-PCP)Rh-Diazirine (5H)
TS-Diazirine to 4memRing TS(5H-6H)
(H-PCP)Rh-4memRing (6H)
TS-4memRing to eta1C TS(6H-7H)
TS-4memRing to CarbeneN2 TS(6H-8H)
(H-PCP)Rh- -CHPhdNdNa (7H)
(H-PCP)PhDiazo-CN-eta2 (7Hi)
TS-DiazCarbon-to-Carbene TS(7H-9H)
(H-PCP)PhCarbeneN2 (8H)
Scheme 7
TS-CarbeneN2 to Carbene TS(8H-9H)
(H-PCP)PhCarbene (9H)
TS-RhN2toRhDiazoEtal TS(1H-3H)
TS-RhN2toRhEtalC TS(1H-7H)
(H-PCP)Rh-Diazo-Rh (10H)
H-PCP Svstem
(mPWlk/SDB-cc-pVDZ)
(H-PCP)RhN2 (1H)
0.0
23.0
0.2
12.6
5.5
20.3
9.2
8.7
17.5
-7.3
10.3
14.8
0.0
11.5
-0.9
13.3
7.0
18.7
9.2
8.8
16.6
-19.4
18.7
23.8
(H-PCP)Rh (2H)
(H-PCP)Rh-NdNdCHPh (3H)
TS-Phdiazo-to-Eta2 TS(3H-4H)
(H-PCP)PhDiazo-NN-eta2 (4H)
TS-eta2NNtoEtalCdiazo TS(4H-7H)
(H-PCP)Rh- -CHPhdNdN (7H)
(H-PCP)Rh- -(CH)PhdNdN Agostic (7Hii)
TS-DiazCarbon-to-Carbene TS(7H-9H)
(H-PCP)PhCarbene (9H)
TS-RhN2toRhDiazoEtal TS(1H-3H)
TS-RhN2toRhEtalC TS(1H-7H)
iPr-PCP System
(mPWlk/LANL2DZ+P//ONIOM(mPWlk/LANL2DZ:
mPWlk/LANLlMB)
(iPrPCP)RhN2 (1P)
0.0
27.5
1.4
0.0
15.8
2.0
(iPr-PCP)Rh (2P)
(iPr-PCP)Rh-NdNdCHPh (3P)
TS-iPr-Phdiazo-to-Eta2 TS(3P-4P)
iPr-PhDiazo-NN-eta2 (4P)
15.0
12.1
23.9
49.0
8.9
20.3
17.4
17.8
27.9
5.2
16.6
15.7
24.0
49.9
11.2
24.9
18.1
21.9
29.1
-5.6
-5.9
15.8
8.6
TS-iPr-eta2NNtoFreeDiazo TS(4P-2P)
TS-iPr-eta2NN-Diazirine TS(4P-5P)
(iPr-PCP)Rh-Diazirine (5P)
(iPr-PCP)Rh-4memRing (6P)
(iPr-PCP)Rh- -CHPhdNdN (7P)
iPr-PhDiazo-CN-eta2 (7Pi)
TS-iPr-DiazCarbon-to-Carbene TS(7P-9P)
(iPr-PCP)PhCarbene-perpendicular (9Pppd
)
be barrierless as evident from the fact that although TS for
8Hf9H was located, it has a lower energy on the ∆Ee surface
than of 8H (see Table 1). The high barrier for the conversion
of η2-NN diazo complex 4H into three-membered diazirine
complex 5H (38.7 kcal/mol) suggests that this pathway is not
operative under the actual reaction conditions. It should be noted
that a high barrier for the conversion of η2-NN complex 4H to
diazirine complex 5H is obtained even though both complexes
have a similar relative energy, suggesting that this kinetic effect
is due to a large amount of energy needed for the induction of
strain during the formation of the diazirine. No TS for the direct
3H or 4H into 6H conversion was found. It should be noted
that four-membered ring η2-(CN) diazoalkane complexes of the
type 6H were suggested to be reactive intermediates in the
course of carbene formation from diazo compounds.3
(iPr-PCP)PhCarbene-parallel (9Pprl
(iPr-PCP)Rh-Diazo-Rh (l0P)b
(iPr-PCP)Rh-Dimer (11P)
)
4.8
-2.7
4.4
a Optimized in SDB-cc-pVDZ basis set. b With respect to 3P+2P.
The second possible pathway in the “bound diazo” mecha-
nism involves η2-NN to η1-C transformation (4Hf7H) with
subsequent N2 loss (Scheme 7, Figure 3). The barrier for η1-N
into η2-NN is 13.4 kcal/mol, and η2-NN in η1-C is 11.2 kcal/
mol, values which are reasonable for this mechanism operating
in H-PCP system at room temperature and below.
Analysis of the temperature dependence of the ∆G surface
reveals that at room temperature the dissociative path (1H f
2H f 3H) is more likely to take place (11.5 kcal/mol for
dissociative vs 19.7 kcal/mol for associative path) and at low
9
J. AM. CHEM. SOC. VOL. 125, NO. 21, 2003 6537

A R T I C L E S
Cohen et al.
Figure 2. H-PCP system, bound mechanism-diazirine pathway, reaction profile (∆G298) at mPW1k/SDB-cc-pVDZ//mPW1k/SDD Level.
Figure 3. H-PCP system, bound mechanism-η2-NN pathway, reaction profile (∆G298) at mPW1k/SDB-cc-pVDZ Level.
temperature (i.e., -70 °C) the dissociative and associative (1H
f 3H) paths are energetically fairly comparable (14.8 vs 16.7
kcal/mol). Already in the H-PCP system, however, the as-
sociative reaction to form 7H via direct reaction of 1H with
PhCHN2, has a relatively high barrier both at low and room
temperatures (22.0 and 25.3 kcal/mol, respectively).
Because optimization of 7H was carried out in a larger basis
set (SDB-cc-pVDZ), all the intermediates involved in the “direct
attack” pathway and transition states were also optimized at
this level of theory for the sake of consistency (Scheme 7, Table
1, Figure 3). Optimization of all structures presented in Scheme
7 in this basis set was beyond our computational resources and
appears redundant due to the intrinsically high energy of the
diazirine pathway. It can be noticed that the net effect of using
the larger basis set is destabilization of 7H with respect to its
relative energy in the smaller basis set calculations. Thus, for
A direct attack of the PhCHN2 carbon atom on the metal
center of intermediate 2H (or 1H) to form 7H (pathway II,
Scheme 5) is definitely a plausible alternative to the “bound
diazo” scheme.
9
6538 J. AM. CHEM. SOC. VOL. 125, NO. 21, 2003

Metallacarbenes from Diazoalkanes
A R T I C L E S
Scheme 8
example, 7H is only slightly more stable (by 2.3kcal/mol) than
the Rh-T shape intermediate 2H. The feasibility of N2 dissocia-
tion from 1H and PhCHN2 coordination through the carbon atom
implies that there is a possibility of direct η1-C diazoalkane
binding to the metal center, the mode favorable for carbene
formation.
An additional pathway (pathway III, Scheme 5) leading to
carbene formation is represented by a bimolecular route. A
complete search for such a mechanism was beyond our
computational resources, but intermediate 10H (see Table 1),
which is a two-metal center complex containing one diazo
molecule, could be located. This intermediate, in which one Rh
center is coordinated to the carbon atom and the second Rh is
coordinated to the terminal nitrogen, could lead to the formation
of the carbene (together with a formation of the Rh-N2 starting
material, 1H) by N-C cleavage, since the carbon atom of the
diazo molecule is already activated by the second Rh-center.
Moreover, it seems to be comparable in energy (see Table 1)
to the two individual intermediates (3H and 2H) in the H-PCP
system.
In short, our computational results for the H-PCP-Rh/
PhCHN2 system can be summarized as follows. Formation of
carbene complex in a system with H-substituents on the
phosphine arms may proceed by two conceptually different
mechanisms, both leading to the η1-C intermediate 7H: (1) a
“bound diazo” mechanism in which the PhCHN2 molecule stays
bound to the metal center throughout the whole reaction; (2) a
“direct attack” pathway in which the η1-N complex 3H is a
resting state. Bimolecular mechanism cannot be ruled out for
the H-PCP system. However, in principle, it is very similar to
the ‘direct attack’ mechanism, as both of them require the prior
coordination of the Rh center to the carbon atom of the PhCHN2
as a key step leading the formation of the carbene. To verify
these findings and to investigate the feasibility of other
that complexes HRhL2(N2) (L ) PtBu3, PiPr3, PCy3) release
nitrogen to form three-coordinate HRhL2 species in the case of
PtBu3 and PCy3, and dimer dinitrogen complex HL2Rh(N2)-
RhHL2 (similar to 1b) in the case of PiPr3.26 Thus, dissociative
N2 loss most probably also takes place in our system, which is
electronically and sterically very similar to the above-mentioned
ones. The Rh dinitrogen complex dimer-monomer equilibrium
(Scheme 9) is also a feasible process to produce the T-shaped
compound 2P. The possible modes of the phenyl diazomethane
bonding are η1-N, η2-NN, η1-C, η2-CN, and η2-(CN) (four-
membered ring). Although η1-N bonding is degenerate, the
cyclic η2-(CN) complex 6P is higher in energy by ∼9 kcal/mol
than the η1-N diazo complex; thus, is not expected to be formed.
(Furthermore, a TS for the coordination of free PhCHN2 to Rh-T
shaped in a η2-(CN) bonding mode, was located and found to
be ∼25 kcal/mol higher in energy than the Rh-T shape
intermediate). Only η2-NN, η2-CN and η1-C modes were found
to be viable. The η2-CN complex (7Pi) is closely related to 7P
both geometrically and energetically (Table 1). Although no
TS(7Pf7Pi) was found, it is expected that the interconversion
between these two complexes is relatively low in energy. It
should be noted that for the H-PCP system, only the η2-CN
complex (7Hi) was found in the small SDD basis set, however
in the larger SDB-cc-pVDZ basis set, only 7H was found,
indicative of the large resemblance between these two complexes
already in the H-PCP model system.
i
mechanisms, we introduced Pr substituents in an ONIOM
method in order to analyze the system having the closest
similarity to the experimental one.
iPr-PCP-Rh/PhCHN2 System (ONIOM). Suggested Re-
action Pathway. Geometry optimizations were performed at
the ONIOM(mPW1k/LANL2DZ:mPW1k/LANL1MB) level,
starting from the previously optimized mPW1k/LANL2DZ
geometries for the H-PCP inner layer. The outer layer (mPW1k/
LANL1MB) included only the iPr substituents on the phosphine
arms. To improve the reliability of the results, all-atom single
point energies calculations at the mPW1k/LANL2DZ+P level
were performed at the optimized geometries. The results of the
calculations are presented in Scheme 8, Figure 4, and Table 1.
First, we considered direct attack of the phenyl diazomethane
on the metal center. Both dissociative (through T-shaped
intermediate 2P) and associative (through dinitrogen complex
1P) mechanisms were tested but no transition states could be
found for the associative mechanisms. It is conceivable that the
associative attack of PhCHN2 on iPrPCP-Rh-N2 complex 1P
will be unfavorable in comparison to dissociative loss of N2
due to the steric interference imposed by the bulky iPr
substituents. Because already for the H-PCP system the barrier
for the associative attack was relatively high (23.8 kcal/mol for
the carbon attack and 18.7 kcal/mol for nitrogen attack) to take
place at low temperatures, it is expected that this barrier will
increase with increasing steric bulk. It was shown experimentally
It appears that the coordination of phenyl diazomethane
through the carbon atom, although less favorable than in the
H-PCP case (probably due to the steric bulk), is an energetically
plausible process leading to the carbene complex. Our attempts
to find a TS connecting 4P, the η2-NN complex, and 7P, the
η1-C complex (TS similar to the one found for the H-PCP
system), invariably lead to detachment of the phenyl diazo-
9
J. AM. CHEM. SOC. VOL. 125, NO. 21, 2003 6539

A R T I C L E S
Cohen et al.
Figure 4. iPr-PCP system, Detached Mechanism pathway, reaction profile (∆G298) at mPW1k/LANL2DZ+P//ONIOM(mPW1k/LANL2DZ:mPW1k/
LANL1mb) Level.
Scheme 9
Figure 5. TS(4H-7H) transition state between η2-NN to η1-C diazo
complexes of H-PCP system.
intermediate. Due to the apparently low barrier of the diazo
dissociation and energetic availability of η1-C intermediate 7P
i
the “direct attack” mechanism appears to operate in the Pr-
PCP-Rh/PhCHN2 system. The “bound diazo” mechanism
seems to be the less probable situation, although we cannot
completely rule out its operating concurrently with the “direct
methane. Thus, TS for the transformation of 7P (η2-NN) to Rh-T
shaped complex 2P and free PhCHN2 was found (Table 1). This
process has a barrier of 8.3 kcal/mol, suggesting that when the
attack” mechanism. The pathway proceeding through diazirine
η2-NN complex is formed, there is a low barrier for dissociation
of the diazo molecule.32 We believe that in the real system a
TS for 4Pf7P transformation most probably is not accessible
due to a steric effect hindering the approach of the carbon atom
from an η2-NN conformation of the diazo. Further support for
this assumption may be obtained by examining the geometry
of the TS(4Hf7H) found for the H-PCP system (Figure 5).
In this TS, the phenyl ring is situated in close proximity to
the phosphine arm, such that the phenyl group is expected to
i
intermediate is intrinsically high in energy also for the Pr-
PCP system, as can be observed from the high barrier for the
formation of the diazirine complex, TS(4P-5P) (Scheme 8,
Table 1).
As far as the bimolecular mechanism (pathway III, Scheme
5) is concerned, the complete search for such a mechanism was
beyond our computational resources. We were able to locate
the intermediate 10P, (see Figure 6) with two PCP-Rh moieties
coordinated to the nitrogen and carbon of the phenyl diazo-
methane. It is about 16 kcal/mol higher in energy than 2P+3P.
Formation of 10P is likely to be preceded by the formation of
2P, and thus, the formation of 7P from 2P and free diazoalkane
seems to be energetically preferable (∆G ) 2.3 kcal/mol,
Scheme 8), favoring the direct attack pathway. It should be noted
that our experimental observations suggest the bimolecular
pathway to be less probable than the direct attack one. Thus,
i
i
interfere with the Pr substituents in the case of the Pr-PCP
system. We believe that in the “direct attack” scheme the free
PhCHN2 is less geometrically restricted to form an η1-C
(32) TS’s are normally difficult to find for association and dissociation of ligands
since it mostly involves entropic contribution, which cannot be calculated
on the bottom-of-the-well energy surface. Here, it was found probably due
to the relatively large geometrical change in the diazo molecule when
coordinated to the metal.
9
6540 J. AM. CHEM. SOC. VOL. 125, NO. 21, 2003

Metallacarbenes from Diazoalkanes
A R T I C L E S
Figure 7. 3D structures of intermediates 3P, 4P, 7P, 9Pprl, and 9Pppd
.
Figure 6. 3D structures of 1P, 2P, 10P, and 11P.
general structure of the PCP-Rh core being almost identical
to that of 1P, reflecting very similar coordination behavior, e.g.,
trans-influence, of N2 and PhCHN2. The NNC axis in 3P is
linear, the coordinated PhCHN2 being very close in geometry
to a free PhCHN2. In η2-NN complex 4P the NN bond is
lengthened in comparison to 3P and free PhCHN2 (bond order
1.70 vs 2.06) as expected for side-on NN coordination. The
η1-C complex 7P has a weak Rh-C bond as evident from the
corresponding bond order (0.1) and length (2.45 Å). It should
be noted that ONIOM geometry optimization at the aforemen-
tioned computational level for this complex lead only to the
closely related C-H agostic form where as in the H-PCP
system, two distinct stationary points were found for the η1-C
and C-H agostic geometries (7H and 7Hii, respectively). In
the H-PCP system the two complexes differed in their relative
energy only in about 0.4 kcal/mol (see Table 1) and the TS for
conversion between them is expected to be very low. Thus, it
could be that the small energy difference together with small
geometry change lead to a flat energy surface in which only
one geometry could be located in the case of ONIOM calcula-
tion. The complex can be viewed as an agostic one due to the
relatively short Rh-H distance (2.03 Å), although the diazo
C-H bond is not lengthened substantially in comparison to the
one in complexes 3P, 4P and free PhCHN2.
we can tentatively rule out the bimolecular pathway III (Scheme
5). In any case, in the bimetallic intermediate 10P, the key
interaction leading to a carbene complex is η1-C coordination
of phenyldiazomethane.
Thus, the direct coordination of the diazo compound through
the carbon atom, followed by the loss of N2, appears to be the
most plausible reaction pathway. On the basis of the experi-
mental and computational studies the following arguments favor
this conclusion. (1) The metallacarbene formation takes place
at a temperature at which the diazoalkane is starting to
decoordinate from the metal center. (2) If the “bound diazo”
scheme were operating starting from the observed η1-N diazo
complex, no side reactions ought to take place in the course of
the metallacarbene formation, contrary to our experimental
observations. (3) The low computed barrier for η1-N - η2-NN
transformation and the low barrier of subsequent PhCHN2
dissociation suggest that the “direct attack” scheme is favored.
(4) Other pathways were computed to be intrinsically high in
energy. (5) Formation of η1-C complex 7P is energetically
feasible. (6) Complex 7P is a key intermediate (one step before
the carbene formation), in almost all of the computed reaction
pathways.
Structure of the Key Compounds. The calculated structures
of 1P, 2P, 10P, 11P (dimer), 3P, 4P, 7P, 9Pprl, and 9Pppd are
presented in Figures 6 and 7. The bond lengths and bond orders
(Wiberg NBO analysis) are given in Table 2.
Carbene Unit Rotation. Although for the H-PCP system
only one optimized geometry was found for the carbene complex
i
9H, for the Pr-PCP system we found two stationary points
The dinitrogen complex 1P has the square planar geometry
expected for 16e- Rh(I) complexes. The N2 molecule is
coordinated in an end-on mode and the N-N bond length is
close to the free N2, so it is not activated. The geometry of the
η1-N diazoalkane complex 3P is also square planar, with the
for the carbene complex, 9Pprl and 9Pppd (Figure 7), both
structures being nearly isoenergetic. this is also reflected in very
similar bond lengths and bond order values (see Figure 4, Table
2) Complex 9Pprl adopts a parallel configuration of the carbene
unit to the PCP surface and 9Pppd a perpendicular one. The
9
J. AM. CHEM. SOC. VOL. 125, NO. 21, 2003 6541

A R T I C L E S
Cohen et al.
i
Table 2. Bond Angles (deg) and Lengths (Å) as Well as Bond Orders of Selected Complexes of the Pr-PCP System at ONIOM(mPWlk/
LANL2DZ:mPWlk/LANL1MB) Level of Theory
free PhCHN2
bond angles N-N-Cdiazo 179.2 P-Rh-P
N-Cdiazo-CPh 122 Cipso-Rh-N 177.4 N-N-Cdiazo 144.7 Cipso-Rh-Cdiazo 174.3
N-Cdiazo-H 116.3 Rh-N-N 177.4 N-Cdiazo-CPh 122.5 N-N-Cdiazo 178.8 Rh-Ccarbene-CPh 134.3 Rh-Ccarbene-CPh 128.6
N-N-Cdiazo 175.5
1.16 Rh-P 2.33 Rh-P
1.3 Cipso-Rh
3P
4P
7P
9Pprl
9Pppd
164.4 P-Rh-P
162.7 P-Rh-P
163.9 P-Rh-P
164.4 P-Rh-P
148.6
CipsoRh-Ccarbene 154.7 Cipso-Rh-Ccarbene 176.2
bond lengths N-N
2.34 Rh-P
2.34 Rh-P
2.02 Cipso-Rh
2.45 Rh-Ccarbene
2.03 Ccarbene-CPh
1.15 Ccarbene-H
1.31
1.46
1.1
0.69 Rh-P
2.33 Rh-P
2.32
2.15
1.91
1.47
1.1
N-Cdiazo
Cdiazo-CPh
Cdiazo-H
2.05 Cipso-Rh
1.95 Rh-Nterminal
1.17 Rh-Ninternal
2.05 Cipso-Rh
2.13 Rh-Cdiazo
2.07 Rh-Hdiazo
1.24 N-N
1.31 N-Cdiazo
1.45 Cdiazo-CPh
1.07 Cdiazo-H
0.41 Cipso-Rh
0.58 Rh-Cdiazo
0.62 Rh-Hdiazo
0.33 N-N
2.1
Cipso-Rh
1.45 Rh-N
1.08 N-N
1.91 Rh-Ccarbene
1.48 Ccarbene-CPh
N-Cdiazo
Cdiazo-CPh
Cdiazo-H
1.3
N-N
1.1
Ccarbene-H
1.45 N-Cdiazo
1.08 Cdiazo-CPh
Cdiazo-H
bond orders N-N
2.29 Rh-P
0.43 Rh-P
0.41 Rh-P
0.44
0.39
0.92
1.15
N-Cdiazo
Cdiazo-CPh
1.4
1.1
Cipso-Rh
Rh-N
0.57 Cipso-Rh
0.55 Rh-Nterminal
2.06 Rh-Ninternal
1.38 N-N
1.11 N-Cdiazo
Cdiazo-CPh
0.1
Cipso-Rh
0.42 Cipso-Rh
1.09 Rh-Ccarbene
1.11 Ccarbene-CPh
0.05 Rh-Ccarbene
2.32 Ccarbene-CPh
1.34
N-N
N-Cdiazo
Cdiazo-CPh
1.7
N-Cdiazo
1.4
Cdiazo-CPh
1.08
1.13
parallel geometry is most probably electronically more favorable
(larger RhdC bond order is observed for the parallel carbene
complexes, and seem to play either competitive or spectator
roles in the metallacarbene formation.
9Pprl: 1.09 in 9Pprl vs 0.92 in 9Pppd), whereas the perpendicular
one is sterically preferred. We believe that the barrier for the
interconversion between the two conformations is low, although
no TS was located. Experimentally, the symmetric NMR pattern
of the carbene complex suggests that the rotation of the carbene
unit takes place and it is a very facile process. It is further
supported by the observation of a pattern characteristic of a
symmetric system even at -70 °C in the NMR spectra of 2. It
has been reported that no barrier for carbene rotation could be
detected even below -100 °C, implying it must be very low.33
The experimental and computed structures of the Rh-iPrPCP
dinitrogen dimer reactant are presented in Figures 1 and 6,
respectively. There is good agreement between the computed
and the X-ray structures (see details in Supporting Information).
It is known that careful choice of ligands and diazo
compounds is crucial in order to achieve carbene complex
formation. A balance is always struck between the kinetic
availability of the carbene complex and its thermodynamic
stability. Thus, one of the main strategies is reacting a
diazoalkane with a metal complex bearing a specific set of
ligands, to be substituted by another set of more bulky ligands
once the carbene complex is formed.5,12 The choice of the diazo
compounds and ligands that result in carbene complex forma-
tion, is normally very limited.
We surmise that the reactivity observed in our PCP and
PCN-Rh systems and in other late transition metal systems is
controlled mainly by steric factors. Because formation of the
carbene complex requires prior η1-C diazo coordination, this
process is sterically demanding, far more so than η1-N or η2-
NN coordination, as reflected in the relative stability of these
compounds. Thus, an important requirement for the carbene
formation is the viability of an η1-C coordinated diazo complex.
The process could not be facilitated by formation of an η1-N
coordinated diazo complex; on the contrary, its formation seems
to interfere in many cases with carbene formation. It has been
suggested that a cyclic η2-(CN) intermediate containing a four-
membered ring is an intermediate in the course of carbene
synthesis.3 Our study shows that in the case of bulky substituents
this mechanism as well as mechanisms proceeding through an
η2-NN intermediate (“bound diazo” scheme) are unlikely to be
operative. On the other hand, in the less bulky H-PCP ligand
system both “bound diazo” and “direct attack” pathways can
take place.
Discussion
Our experimental observations and computational studies
suggest that at low temperatures reaction of the Rh(I) complex
with phenyl diazomethane leads to formation of the η1-N diazo
complexes and that coordination of the diazoalkane is reversible
upon warming. Binding of the diazoalkane to the metal through
carbon (η1-C) and subsequent N2 loss result in metallacarbene
formation. The formation mechanism of the putative metalla-
carbene in the course of the cyclopropanation reaction was
studied computationally,9c,d concluding that it involves initial
diazoalkane coordination through the carbon atom, followed by
loss of N2. However, other pathways, involving coordination
through the nitrogen atoms, were not investigated. In one
experimental study, an intermediate in the cyclopropanation
reaction with a diazoalkane coordinated through carbon atom
was observed at low temperature.21
The importance of steric factors in the formation of the critical
η1-C bound diazo complex intermediate seems to be paramount.
Thus, adding iPr substituents in our DFT study diminished the
stability of the η1-C intermediate 7P due to steric repulsion.
Additional evidence for the importance of steric factors is best
demonstrated in the work of Werner et al.5,14 (Scheme 10). In
their approach, a Rh complex bearing SbiPr3 ligands is reacted
with Ph2N2 to produce a metallacarbene. Stibane ligands are
then substituted by PiPr3 to give a stable carbene complex. When
It is generally recognized that formation of diazoalkane
complexes coordinated through N (either by η1-N or η2-NN)
may stabilize the diazo compounds. There are many examples
of stable diazoalkane complexes demonstrating η1-N or η2-NN
coordination modes. Generally, they do not form carbene
(33) Manganiello, F. J.; Radcliffe, M. D.; Jones, W. M. J. Organomet. Chem.
1982, 228, 273.
9
6542 J. AM. CHEM. SOC. VOL. 125, NO. 21, 2003

Metallacarbenes from Diazoalkanes
A R T I C L E S
Scheme 10
diazoalkanes. Design of systems for formation of transition metal
carbenes from diazoalkanes should involve a coordinatively
unsaturated metal precursor and careful consideration of steric
factors, i.e., a proper diazoalkane fitting to the steric require-
ments of the metal center.
Computational Methods
All of the calculations were carried out using the Gaussian 98
program revisions A.734 and A.1135 running on Compaq ES40 and
XP1000 workstations as well as on a mini-farm of Pentium IV Xeon
1.7/2.0 GHz PC’s running Red Hat Linux 7.2 in our group, on an
experimental Linux PC Farm at the Faculty of Physics, and on the
(Israel) Inter-University Computing Center (IUCC) SGI Origin 2000.
(The patch to Gaussian 98 rev A.7 as detailed in the Appendix to ref
36 was applied.)
For the H-PCP system the mPW1k (modified Perdew-Wang
1-parameter for kinetics) exchange-correlation functional of Truhlar
and co-workers37 was employed in conjunction with the SDD and SDB-
cc-pVDZ basis sets (see below). The mPW1k functional was very
recently shown37,38 to yield more reliable reaction barrier heights than
other exchange-correlation functionals.
a Rh complex bearing iPr ligands is reacted with Ph2N2, an η1-N
diazo complex is formed. Stibane ligands most probably serve
as a less crowded analogue of phosphines due to their smaller
cone angle resulting from the larger atomic radius of Sb
compared to P. As a consequence, the metal center is less
sterically impeded and the η1-C diazo intermediate can be
formed. The same probably holds for the Ru carbene synthesis
using diazo alkanes by Grubbs et al.12 In this synthesis, (PPh3)3-
RuCl2 and phenyldiazomethane are utilized. Following the
carbene formation, PCy3 is added to produce a more stable,
sterically shielded carbene. The direct synthesis starting with
the precursor bearing PCy3 ligands is most probably difficult
due to steric reasons. It should be noted that generally, with a
given set of ligands on a metal center, only a narrow selection
of diazo alkanes can be used to produce a carbene complex.
Smaller diazoalkanes normally produce unstable carbenes (that
decompose by dimerization), while larger diazoalkanes give
η1-N bound diazo complexes, due to the difficulty in formation
of the essential η1-C diazo complex, which is sterically
demanding. The η1-N binding is obviously less sterically
demanding and η1-N complexes can be formed with a large
variety of metal complexes. Thus, the fit between diazoalkane
size and ligand bulk is a factor requiring the utmost consider-
ation.
Regarding the electronic factors, the availability of a vacant
coordination site is crucial. In almost all cases, metal precursors
for carbene complexes have either a free coordination site or
an ability to create one. We believe that the design of systems
for the production of metallacarbenes using diazo alkanes should
include a coordinatively unsaturated metal precursor and careful
consideration of steric factors, i.e., diazoalkane fit to the steric
crowding on the metal. Then the stability of the resulting carbene
has to be considered: it can be increased with bulky ligands,
protecting the carbene from dimerization and external attack
by solvent or other reagents.
The SDD basis set is the combination of the Huzinaga-Dunning
double-ú basis set on lighter elements with the Stuttgart-Dresden basis
set-relativistic effective core potential (RECP) combination39 on the
transition metals. The SDB-cc-pVDZ basis set, combines the Dunning
cc-pVDZ basis set40 on the main group elements with the Stuttgart-
Dresden basis set-RECP combination39 on the transition metals, with
an f-type polarization exponent taken as the geometric average of the
two f-exponents given in the Appendix to ref 41.
i
For the Pr-PCP, as in our previous studies on the PCP,42 PCN,43
and PCO44 systems, the initial survey of the potential surface was carried
(34) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M.
A.; Cheeseman, J. R.; Zakrzewski, V. G.; Montgomery, J. A.; Stratmann,
R. E.; Burant, J. C.; Dapprich, S.; Millam, J. M.; Daniels, A. D.; Kudin,
K. N.; Strain, M. C.; Farkas, O.; Tomasi, J.; Barone, V.; Cossi, M.; Cammi,
R.; Mennucci, B.; Pomelli, C.; Adamo, C.; Clifford, S.; Ochterski, J.;
Petersson, G. A.; Ayala, P. Y.; Cui, Q.; Morokuma, K.; Malick, D. K.;
Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Cioslowski, J.; Ortiz,
J. V.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.;
Gomperts, R.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng,
C. Y.; Nanayakkara, A.; Gonzalez, C.; Challacombe, M.; Gill, P. M. W.;
Johnson, B. G.; Chen, W.; Wong, M. W.; Andres, J. L.; Head-Gordon,
M.; Replogle, E. S.; Pople, J. A.; Gaussian 98, Revision A.7; Gaussian,
Inc.: Pittsburgh, PA, 1998.
(35) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M.
A.; Cheeseman, J. R.; Zakrzewski, V. G.; Montgomery, J. A. Jr.; Stratmann,
R. E.; Burant, J. C.; Dapprich, S.; Millam, J. M.; Daniels, A. D.; Kudin,
K. N.; Strain, M. C.; Farkas, O.; Tomasi, J.; Barone, V.; Cossi, M.; Cammi,
R.; Mennucci, B.; Pomelli, C.; Adamo, C.; Clifford, S.; Ochterski, J.;
Petersson, G. A.; Ayala, P. Y.; Cui, Q.; Morokuma, K.; Salvador, P.;
Dannenberg, J. J.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.;
Foresman, J. B.; Cioslowski, J.; Ortiz, J. V.; Baboul, A. G.; Stefanov, B.
B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; Gomperts, R.; Martin,
R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara,
A.; Challacombe, M.; Gill, P. M. W.; Johnson, B.; Chen, W.; Wong, M.
W.; Andres, J. L.; Gonzalez, C.; Head-Gordon, M.; Replogle, E. S. and
Pople, J. A. Gaussian 98, Revision A.11; Gaussian, Inc., Pittsburgh, PA,
2001.
Conclusions
We have carried out an experimental and computational study
of carbene complex formation from phenyl diazomethane in
PCN and PCP-Rh systems. Our studies show that end-on η1-N
complex formation precedes the formation of the carbene
complex in both cases. Coordination of PhCHN2 to the metal
centers in PCP and PCN systems was found to be reversible
both experimentally and computationally. The diazo η1-N
complex appears to be rather a side equilibrium resting state
and not an essential intermediate for carbene formation. The
DFT calculations reveal that the key intermediate for the
metallacarbene formation is an η1-C diazo complex. Its forma-
tion proceeds most probably by a direct attack of the diazo
molecule on the metal center. As a consequence, steric factors
play an important role in the carbene complex formation from
(36) Martin, J. M. L.; Bauschlicher, C. W.; Ricca, A. Comput. Phys. Commun.
2001, 133, 189.
(37) Lynch, B. J.; Fast, P. L.; Harris, M.; Truhlar, D. G. J. Phys. Chem. A.
2000, 104, 4811.
(38) (a) Parthiban, S.; de Oliveira, G.; Martin, J. M. L. J. Phys. Chem. A. 2001,
105, 895. (b) Lynch, B. J.; Truhlar, D. G. J. Phys. Chem. A. 2001, 105,
2936. (c) Iron, M. A.; Lo, H. C.; Martin, J. M. L.; Keinan, E. J. Am. Chem.
Soc. 2002, 124, 7041.
(39) Dolg, M. In Modern Methods and Algorithms of Quantum Chemistry;
Grotendorst, J., Ed.; John von Neumann Institute for Computing: Ju¨lich,
2000; Vol. 1, pp 479-508.
(40) Dunning, T. H., Jr. J. Chem. Phys. 1989, 90, 1007.
(41) Martin, J. M. L.; Sundermann, A. J. Chem. Phys. 2001, 114, 3408.
(42) Sundermann, A.; Uzan, O.; Milstein, D.; Martin, J. M. L. J. Am. Chem.
Soc. 2000, 122, 7095.
(43) Sundermann, A.; Uzan, O.; Martin, J. M. L. Organometallics 2001, 20,
1783.
(44) Rybtchinski, B.; Oevers, St.; Montag, M.; Vigalok, A.; Rozenberg, H.;
Martin, J. M. L.; and Milstein, D. J. Am. Chem. Soc. 2001, 123, 9064-
9077.
9
J. AM. CHEM. SOC. VOL. 125, NO. 21, 2003 6543

A R T I C L E S
Cohen et al.
out by a two-layer ONIOM30 approach, in which the iso-propyl groups
on the phosphorus were placed in the outer layer, and the remainder
of the system was placed in the inner layer. For technical reasons, the
mPW1k functional was used for both layers. For the inner layer the
valence Los Alamos National Laboratory double-ú (LANL2DZ) basis
set- RECP combination was used,45 whereas for the outer layer, the
LANL1MB basis set-RECP combination was employed.45 (As is
customary, first-row atoms in this approach were treated by means of
the Dunning-Hay valence double-ú46 basis set).
Geometry optimizations for minima were carried out using the
standard Schlegel algorithm47 in redundant internal coordinates until
in the neighborhood of the solution, and then continued using analytical
second derivatives.48 Optimizations for transition states were carried
out by means of the QST3 approach,49 with an initial guess for the
transition state being generated from either linear synchronous transit
(LST50) or from manual manipulation of the geometry using MOLD-
EN.51 In cases where this approach failed to converge, we used
analytical second derivatives at every step.
Where necessary, the Grid ) UltraFine combination, i.e., a pruned
(99 590) grid in the integration and gradient steps and a pruned (50 194)
grid in the CPKS (coupled perturbed Kohn-Sham) steps, was used as
recommended in ref 36.
Zero-point and RRHO (rigid rotor-harmonic oscillator) thermal
corrections were obtained from the unscaled computed frequencies.
Where necessary to resolve ambiguities about the nature of a
transition state, intrinsic reaction coordinate (IRC52) calculations were
carried out. In some cases where IRC calculation failed for technical
reason, displacement along the normal coordinate for the imaginary
frequencies and optimizations started from there. Although this
procedure is less unambiguous than IRC it at least offers some form
of corroboration.
we obtain a reaction energy of 43-45 kcal/mol. At the mPW1K/SDB-
cc-pVDZ level used in this study, we obtain 51.7 kcal/mol at 298 K,
and 52.2 kcal/mol at absolute zero. As the experimental heat of
formation for diazomethane would appear to be quite uncertain and
the systems are small enough for ab initio computational thermochem-
istry, we have recalculated the reaction energy using G3X theory58 as
well as the CBS-APNO and CBS-QB3 methods,59 all of which have
average uncertainties in the 1 kcal/mol range for molecular atomization
energies, as well as the rather more elaborate W1 and W2h methods60
which generally achieve sub-kcal/mol accuracy. At absolute zero, we
obtain the following reaction energies: G3 × 57.9, CBS-APNO 57.4,
CBS-QB3 57.4, W1 58.6, and W2h 58.8 kcal/mol. In other words, five
computational thermochemistry methods from three very different
families yield fundamentally the same conclusion, namely that the
experimental heat of formation of diazomethane is in error by 12-15
kcal/mol: The most accurate (and computationally demanding) method,
W2h theory, predicts heats of formation at 0 and 298 K of 66.4 and
65.0, respectively. (For comparison, the same level of theory reproduces
the experimental atomization energies of C2H4 and N2 to within 0.45
and 0.32 kcal/mol, respectively. We note that W2h theory is devoid of
parameters derived from experiment.)
For the study at hand, this means that the mPW1K/SDB-cc-pVDZ
level of theory overestimates the reaction energy by 5-6 kcal/mol.
We considered larger basis sets (up to cc-pVQZ40) as well as different
exchange-correlation functionals (such as B3LYP and B97-161) and
obtained qualitatively similar results. For example, at the B97-1/cc-
pVQZ level, we obtain a reaction energy of 49.5 kcal/mol at room
temperature and 49.9 kcal/mol at absolute zero. As B97-1/TZ2P
reproduces the atomization energies of C2H4, N2, and even singlet CH2
to within 1 kcal/mol,62 this suggests that diazomethane and similar
systems present special difficulties to DFT methods, and we shall
investigate this methodological issue in detail soon. Yet the mPW1K/
SDB-cc-pVDZ results should still be of sufficient accuracy to permit
insight into the reaction mechanism.
For interpretative purposes, Wiberg bond indices53 were derived from
the natural bond order (NBO) analysis54 at the mPW1k/SDB-cc-pVDZ
level for the H-PCP systems and at the mPW1k/LANL2DZ level for
i
the Pr-PCP systems.
The energetics for our final iPr-PCP profile were validated by single-
point energy calculations, using the ONIOM(mPW1k/LANL2DZ:
mPW1k/LANL1MB) reference geometries, at the higher level of theory
mPW1k/LANL2DZ+P, where the “+P” stands for the addition of
polarization functions with exponents taken from ref 55.
Although the sheer size of the systems studied more or less leaves
DFT as the only alternative, DFT has never been really benchmarked
for diazo compounds and carbenes. We have therefore considered the
reaction energy of
Experimental Methods
All experiments with metal complexes and phosphine ligands were
carried out under an atmosphere of purified nitrogen in an MBraun
MB 150B-G and Vacuum Atmospheres Nexus System gloveboxes.
PCP63 and PCN64 ligands, [Rh(COE)2Cl]2 and PhCHN2 were
65
66
prepared according to literature procedures. (COE ) cyclooctene).
1H, 13C, 31P, and 15N (in 2D 15N-1H correlation) NMR spectra were
recorded at 400.1, 100.6, 162.0, and 40.5 MHz, respectively, at 295 K
(if not specified otherwise), using a Bruker Avance-400 NMR
spectrometer. Long range 2D15N-1H correlation spectrum was mea-
sured using the Bruker standard microprogram GHMBC by an inverse
gradient selected experiment using a 5-mm Bruker inverse multinuclear
resonance probe with a single-axis (z) gradient coil. 1H NMR and 13C-
{1H} NMR chemical shifts are reported in ppm downfield from
tetramethylsilane. 1H NMR chemical shifts are referenced to the residual
(1/2) C2H4 + N2 f CH2NN
From the respective experimental heats of formation at 298 K of 12.52
( 0.12 kcal/mol,56 0 (by definition), and 49.3 ( 2.3 or 51.3 kcal/mol,57
(45) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 270, 284, 299.
(46) Dunning, T. H.; Hay, P. J. In Modern Theoretical Chemistry; Schaefer, H.
F., III, Ed.; Plenum: New York, 1976; Vol. 3.
(47) Schlegel, H. B. J. Comput. Chem. 1982, 3, 214, Peng, C.; Ayala, P. Y.;
Schlegel, H. B.; Frisch, M. J. J. Comput. Chem. 1996, 17, 49.
(48) Stratmann R. E.; Burant J. C.; Scuseria G. E.; Frisch M. J. J. Chem. Phys.
1997, 106, 10 175.
(49) Peng, C.; Schlegel, H. B. Isr. J. Chem. 1994, 33, 449.
(50) Halgren, T. A.; Lipscomb, W. N. Chem. Phys. Lett. 1977, 49, 225.
(51) Schaftenaar, G. Molden 3.6, 1999. URL: http://www.cmbi.kun.nl/∼schaft/
molden/molden.html.
(57) Afeefy, H. Y. Liebman, J. F.; Stein, S. E.; Neutral Thermochemical Data
in NIST Chemistry WebBook, NIST Standard Reference Database Number
69, Linstrom, P. J.; Mallard, W. G.; Eds.; (Gaithersburg: National Institute
of Standards and Technology, 2001): http://webbook.nist.gov.
(58) Curtiss, L. A.; Redfern, P. C.; Raghavachari, K.; and Pople, J. A.; J. Chem.
Phys. 2001, 114, 108.
(59) Montgomery, J. A., Jr.; Frisch, M. J.; Ochterski, J. W.; Petersson, G. A.;
J. Chem. Phys. 1994, 101, 5900; Montgomery, J. A., Jr.; Frisch, M. J.;
Ochterski, J. W.; Petersson, G. A.; J. Chem. Phys. 1999, 110, 2822.
(60) Martin, J. M. L.; De Oliveira, G.; J. Chem. Phys. 1999, 111, 1843; Parthiban,
S.; Martin, J. M. L.; J. Chem. Phys. 2001, 114, 6014.
(61) Hamprecht, F. A.; Cohen. A. J.; Tozer, D. J.; Handy, N. C.; J. Chem. Phys.
1998, 109, 6264.
(62) Boese, A. D., personal communication.
(63) B. Rybtchinski, Y. Ben-David, D. Milstein Organometallics 1997, 16, 3786.
(64) Gandelman, M.; Vigalok, A.; Shimon, L. J. W.; Milstein, D. Organome-
tallics 1997, 16, 3981.
(65) Herde, J. L.; Senoff, C. V.; Inorg. Nucl. Chem. Lett. 1971, 7, 1029.
(66) Creary, X. Organic Syntheses; Wiley: New York, 1990; Collect. Vol. VII,
p 438.
(52) Gonzalez, C.; Schlegel, H. B. J. Chem. Phys. 1989, 90, 2154; J. Phys.
Chem. 1990, 94, 5523.
(53) Wiberg, K. B. Tetrahedron 1968, 24, 1083.
(54) Reed, A. E.; Curtiss, L. A.; Weinhold, F. Chem. ReV. 1988, 88, 899.
(55) Hollwarth, A.; Bohme, M.; Dapprich, S.; Ehlers, A. W.; Gobbi, A.; Jonas,
V.; Kohler, K. F.; Stegmann, R.; Veldkamp, A.; Frenking, G. Chem. Phys.
Lett. 1993, 208, 237. Ehlers, A. W.; Bohme, M.; Dapprich, S.; Gobbi, A.;
Hollwarth, A.; Jonas, V.; Kohler, K. F.; Stegmann, R.; Veldkamp, A.;
Frenking, G. Chem. Phys. Lett. 1993, 208, 111.
(56) Thermodynamic Properties of IndiVidual Substances, 4th ed.; Hemisphere:
New York, 1989.
9
6544 J. AM. CHEM. SOC. VOL. 125, NO. 21, 2003

Metallacarbenes from Diazoalkanes
A R T I C L E S
hydrogen signal of the deuterated solvents, and in 13C{1H} NMR the
13C signal of the deuterated solvents was used as a reference. 31P NMR
chemical shifts are reported in ppm downfield from H3PO4 and
(s, CH3-CH2-N), 34.81 (dd, 1JPC ) 15.1 Hz, 2JRhC ) 1.6 Hz, (CH3)3C-
2
P), 30.28 (d, JPC ) 5.9 Hz, (CH3)3C-P), 21.78 (s, CH3-Ar), 13.47
(s, CH3-CH2-N). Assignment of 13C{1H}NMR signals was confirmed
by 13C DEPT. 15N NMR of 2b labeled with 15N2 (toluene-d8): 305.5
referenced to an external 85% solution of phosphoric acid in D2O. 15
N
1
NMR chemical shifts are reported as referenced downfield to liquid
ammonia. Abbreviations used in the description of NMR data are as
follows: Ar ) aryl, br ) broad, dist. ) distorted, s ) singlet, d )
doublet, t ) triplet, q ) quartet, m ) multiplet, and v ) virtual.
Elemental analysis was performed at Mikroanalytisches Laboratorium,
45470 Mu¨lheim, Germany.
(br d) JRhN ) 18.4, Rh-NtN).
Preparation of the Rhodium Carbene Complex 3. When a C6D6
solution of phenyldiazomethane (98 µL, 0.45M solution) was added
to a C6D6 solution (0.5 mL) of the dinitrogen complex 1 (20 mg, 0.044
mmol of monomer form) at room-temperature quantitative formation
of complex 3 accompanied by N2 evolution was observed. The
characterization of compound 3 was reported.27
Synthesis of 1. A THF solution (3 mL) of the PCP ligand, 1,3-bis-
[(diisopropyl-phosphanyl)-methyl]-benzene, (142 mg, 0.419 mmol) was
added to a THF solution (5 mL) of [Rh(COE)2Cl]2 (150 mg, 0.209
mmol). The reaction mixture was kept at room temperature for 1 h
and dried under vacuum. The residue was dissolved in THF (5 mL)
and a THF solution (4 mL) of KOtBu (70 mg, 0.624 mmol) was added
to it, resulting in a color change from orange to brown. The reaction
mixture was kept for 0.5 h at room temperature, after which the solvent
was removed under vacuum and the residue was washed with a small
amount of pentane. The resulting solid was extracted with toluene, and
the solution was filtered to remove insoluble inorganic particles and
dried under vacuum, resulting in 160 mg (84% yield) of 1 as a brown
solid. The compound is essentially clean (95% by NMR), and can be
used as is. Complex 1 was further purified by crystalization from a
toluene-pentane mixture (8 mL, tol:pen ) 1:3, v/v) at -35 °C resulting
in orange crystals. They were collected by filtration, washed with cold
pentane, and dried under vacuum resulting in 100 mg (52.7% yield) of
pure complex 1. It is slightly soluble in pentane and highly soluble in
other organic solvents such as benzene, toluene and THF.
Formation of 4. A C6D6 solution of phenyldiazomethane (196 µL,
0.45M solution) was added at -70 °C to a toluene-d8 solution (0.5
mL) of 1 (40 mg, 0.088 mmol of monomer form), resulting in the
quantitative formation of 4. Complex 4 was characterized in solution
at -60 °C.
31P{1H}-NMR (toluene-d8): 66.46 (d, 1JRhP ) 154.4 Hz). 1H NMR
(toluene-d8): 7.31 (m, 1H, Ar-H), 7.21 (m, 2H, Ar-H), 7.10 (br s, 2H,
Ar-H), 6.87 (m, 2H, Ar-H), 6.53 (m, 1H, Ar-H), 3.76 (s, 1H, N2CHPh),
2.96 (br s, 4H, Ar-CH2-P), 1.82 (m, 4H, P-CH(CH3)2), 1.23 (dist q,
JPH ) 7.4 Hz, 12H, P-CH(CH3)2), 0.98 (dist q, JPH ) 6.2 Hz, 12H,
1
P-CH(CH3)2). 13C{1H}-NMR (toluene-d8): 173.02 (dt, JRhC ) 38.3
Hz, 2JPC ) 4.9 Hz, Cipso, Rh-Ar), 151.80 (t, JPC ) 12.5 Hz, Ar), 132.41
(s, Ar), 126.77 (s, Ar), 123.63 (s, Ar), 122.98 (s, Ar), 121.09 (t, JPC
)
9.5 Hz, Ar), 120.56 (s, Ar), 57.95 (s, N2CHPh), 34.80 (vt, JPC ) 9.8
Hz, Ar-CH2-P), 25.20 (vt, JPC ) 9.4 Hz, P-CH(CH3)2), 19.60 (s,
P-CH(CH3)2), 18.65 (s, P-CH(CH3)2). The assignment of the 13C-
{1H} NMR signals was confirmed by 13C DEPT. 15N NMR from 15N-
1
1H correlation (toluene-d8): 421.77 (d, JRhN ) 31.5 Hz, Rh-Nt
1
1
1a: 31P{1H} NMR (C6D6): 66.20 (d, JRhP ) 157.9 Hz). H NMR
NCHPh), 300.03 (s, Rh-NtNCHPh).
(C6D6): 7.13 (br. s, 2H, Ar-H), 7.09 (m, 1H, Ar-H), 3.03 (vt, 4H, J )
Preparation of the Rhodium Carbene Complex 5. Addition of a
C6D6 solution of phenyldiazomethane (93 µL, 0.45M) to a C6D6 solution
(0.5 mL) of the dinitrogen complex 2 (20 mg, 0.042 mmol of monomer
form) at room-temperature resulted in quantitative formation of complex
5 accompanied by N2 evolution. The characterization of the compound
5 was reported.27
3.8 Hz, Ar-CH2-P), 2.03 (m, 4H, P-CH(CH3)2), 1.32 (dist q, JPH
)
8.0 Hz, 12H, P-CH(CH3)2), 1.09 (dist q, JPH ) 6.4 Hz, 12H, P-CH-
1
2
(CH3)2). 13C{1H}-NMR (C6D6): 172.22 (dt, JRhC ) 38.4 Hz, JPC
4.9 Hz, Cipso, Rh-Ar), 152.62 (td, J ) 13.4 Hz, J ) 3.2 Hz, Ar), 123.31
(s, Ar), 121.23 (t, J ) 9.4 Hz, Ar), 35.49 (vtd, JPC ) 11.5 Hz, JRhC
)
)
4.2 Hz, Ar-CH2-P), 25.83 (vt, JPC ) 9.5 Hz, P-CH(CH3)2), 20.34 (vt,
JPC ) 3.3 Hz, P-CH(CH3)2), 18.88 (s, P-CH(CH3)2). IR (film): 2165
Formation of 6. A C6D6 solution of phenyldiazomethane (93 µL,
0.45M) was added at -70 °C to a toluene-d8 solution (0.6 mL) of 2
(20 mg, 0.042 mmol of the monomer form), resulting in the quantitative
formation of 6. Complex 6 was characterized in solution at -70 °C.
cm-1, νNtN
.
1
Selected NMR data for 1b: 31P{1H} NMR (C6D6): 67.35 (d, JRhP
1
) 155.5 Hz). H NMR (C6D6): 2.98 (vt, 4H, J ) 3.6 Hz, Ar-CH2-
1
1
31P{1H}NMR (toluene-d8): 88.90 (d, JRhP ) 204.6 Hz). H NMR
(toluene-d8): 7.49-7.35 (m, 2H, Ar), 6.93 (t, 3JHH ) 7.2 Hz, 1H), 6.82
(t, 3JHH ) 7.6 Hz, 1H), 6.67 (s, 1H, aromatic proton of the PCN ligand),
6.46 (d, JHH ) 8.2 Hz, 1H, Ar), 3.84(s, 1H, N2CHPh), 3.72-3.48
(multiplets, 4H, Ar-CH2-N, Ar-CH2-P), 3.18 (ddq, 3JHH ) 7.4 Hz,
P), 1.85 (m, 4H, P-CH(CH3)2), 1.18 (dist q, JPH ) 8.2 Hz, 12H, P-CH-
(CH3)2), 0.98 (dist q, JPH ) 6.5 Hz, 12H, P-CH(CH3)2). Elemental
analysis of 1b; Anal. Calcd for C40H70N2P4Rh2: C, 52.87; H, 7.76.
Found: C, 52.75; H, 7.79.
Synthesis of 2. A THF solution of NaBH(C2H5)3 (1M, 0.1 mL) was
added to a THF solution (3 mL) of the PCN rhodium methyl chloride64
(50 mg, 0.1 mmol) (PCN: {3-[(di-tert-butyl-phosphanyl)-methyl]-
benzyl}-diethyl-amine). The reaction mixture was stirred for 20 min
at room temperature, after which it was filtered and the solvent removed
under vacuum. The resulting solid was extracted with benzene, and
the benzene solution was filtered and dried under vacuum resulting in
45 mg (95%) of 2 as a yellow-brown solid. Complex 2 is slightly soluble
in pentane and highly soluble in other organic solvents such as benzene,
toluene and THF. The characterization of the dimer 2b is given as it is
the main product at room temperature and below.
3
2JHH ) 11.7 Hz, JRhH ) 2.3 Hz, 2H, CH3-CH2-N), 2.64 (m, 2H,
CH3-CH2-N), 2.22 (s, 3H, Ar-CH3), 2.00 (s, 3H, Ar-CH3), 1.53 (t,
3
JHH ) 6.9 Hz, 6H, CH3-CH2-N), 1.16 (d, JPH ) 12.5 Hz, 18H,
(CH3)3C-P). 13C{1H}NMR (toluene-d8): 169.52 (dd, 1JRhC ) 39.8 Hz,
2JPC,cis ) 4.2 Hz, Cipso, Rh-Ar), 150.41 (bs,Ar), 148.1 (d, JPC ) 14.1
Hz, Ar), 147.25 (s, Ar), 136.3 (s, Ar), 127.83 (s, Ar), 124.11 (s, Ar),
123.24 (s, Ar), 121.50 (s, Ar), 121.12 (s, Ar), 68.89 (s, N2CHPh), 64.62
1
(s, Ar-CH2-N), 55.40 (s, CH3-CH2-N), 36.10 (d, JPC ) 15.8 Hz,
Ar-CH2-P), 34.51 (d, 1JPC ) 14.5 Hz, (CH3)3C-P), 30.99 (bs, (CH3)3C-
P), 25.70 (s, CH3-Ar), 22.15 (s, CH3-Ar), 14.96 (s, CH3-CH2-N).
2b (integration given for one Rh-PCN unit): 31P{1H}NMR (toluene-
Competition Experiments. In a typical experiment, complex 6 was
obtained as described before from complex 2 (15 mg, 0.0315 mmol)
and an equimolar amount of PhCHN2 (0.45M C6D6 solution) in toluene-
d8. A toluene-d8 solution of 1 (30 mg, 0.066 mmol) was added at -70
°C to a toluene-d8 solution of 6 in an NMR tube equipped with rubber
septum. The NMR tube was placed in the NMR probe pre-cooled to
-70 °C and the reaction was followed at various temperatures.
X-ray Structural Analysis of Complex 1b. Complex 1b was
crystallized from toluene-pentane (1:5, v/v) solution of 1 in C2/c space
group with half a dimer (PCP-Rh-N) per asymmetric unit.
1
1
d8): 99.27 (d, JRhP ) 217.4 Hz). H NMR (toluene-d8): 6.45 (s, 1H,
3
2
Ar), 3.79 (s, 2H, Ar-CH2-N), 3.08 (dq, JHH ) 7.2 Hz, JHH ) 12.3
2
Hz, 2H, CH3-CH2-N), 2.96 (d, JPH ) 8.36 Hz, 2H, Ar-CH2-P),
2.79 (ddq, 3JHH ) 7.0 Hz, 2JHH ) 12.1 Hz, 3JRhH ) 2.7 Hz, 2H, CH3-
CH2-N), 2.23 (s, 3H, Ar-CH3), 2.09 (s, 3H, Ar-CH3), 1.50 (t, JHH
)
3
7.2 Hz, 6H, CH3-CH2-N), 1.33 (d, JPH ) 12.3 Hz, 18H, (CH3)3C-
P). 13C{1H}NMR (toluene-d8): 174.72 (dd, 1JRhC ) 29.8 Hz, 2JPC,cis
3.6 Hz, Cipso, Rh-Ar), 147.81 (s, Ar), 146.83 (s, Ar), 146.25 (d, JPC
)
)
7.1 Hz, Ar), 127.01 (s, Ar), 126.13 (s, Ar), 64.46 (s, Ar-CH2-N), 56.18
9
J. AM. CHEM. SOC. VOL. 125, NO. 21, 2003 6545

A R T I C L E S
Cohen et al.
Crystal Data. (C20H35NP2Rh), yellow prisms, 0.2 × 0.2 × 0.1 mm3,
monoclinic, C2/c (no. 15), a ) 21.628(4) Å, b ) 10.970(2) Å, c )
20.463(4) Å, â ) 112.08(3)°, from 10 degrees of data, T ) 120 (2) K,
V ) 4499.0(16) Å3, Z ) 8, Fw ) 454.34, Dc ) 1.342 Mg/m3, µ )
Acknowledgment. This work was supported by the Israel
Science Foundation (Grant No 83/00), the MINERVA founda-
tion, Munich, Germany, the Tashtiyot program of the Ministry
of Science (Israel), the Israel Inter-University Computing Center
and the Helen and Martin Kimmel Center for Molecular Design.
We thank Dr. Leonid Konstantinovsky for the help with NMR
measurements and Mr. Yehoshoa Ben-David for technical
assistance. We also thank Mr. Mark Iron for valuable discus-
sions. D. M. holds the Israel Matz Professorial Chair in Organic
Chemistry. B. R. and M. G. are the recipients of The
International Precious Metals Institute Student Awards, grants
of which supported this work.
0.904 mm-1
.
Data Collection and Treatment. Nonius KappaCCD diffractometer,
MoKR (λ ) 0.71073 Å), graphite monochromator, 36758 reflections,
-27 e h e 25, 0 e k e 13, 0 e l e 25, frame scan width ) 1.4°,
scan speed 1° per 70 s, typical peak mosaicity ) 0.41°, 4762
independent reflections (Rint ) 0.071). Data were processed with Denzo-
Scalepack.
Solution and Refinement. The structure was solved by direct
methods with SHELXS-97, and refined by full matrix least-squares
based on F2 with SHELXL-97 (225 parameters with no restraints).
Idealized hydrogen atoms were placed and refined in riding mode. The
final R-factors are R1 ) 0.029 (based on F2) for data with I > 2σ(I),
and R1 ) 0.045 on all 4762 reflections. The goodness-of-fit on F2 is
1.051, and the largest electron density equals 0.411 e/Å3.
Supporting Information Available: X-ray data for 1b and
XYZ coordinates of all computed structures (PDF). This material
is available free of charge via the Internet at http://pubs.acs.org.
JA028923C
9
6546 J. AM. CHEM. SOC. VOL. 125, NO. 21, 2003