Isomerism of [Ru(η3-allyl)Cl(CO)(PPh3) 2]

Article abstract of DOI:10.1021/om049646g

Treatment of [RuHCl(CO)(PPh₃)₃] with CH ₂=C=CMe₂ produced the new allyl complex [Ru(ζ³-CH₂CHCMe₂)Cl(CO)(PPh ₃)₂], which slowly isomerizes to [Ru(ζ³- CH₂CMeCHMe)Cl-(CO)(PPh₃)₂] in both solution and the solid state. While both endo and exo isomers can be observed by solution NMR for the allyl complex [Ru(ζ³-CH₂CHCMe ₂)Cl(CO)(PPh₃)₂], only the endo isomer was detected for other related analogous allyl complexes [Ru(ζ³- allyl)Cl-(CO)(PPh₃)₂] (e.g. allyl = CH₂CMeCHMe, CH₂CHCHPh). The isomeric behavior has been investigated by computational chemistry. The theoretical calculations show that the metal-ζ³-allyl interaction in an endo isomer of [Ru(ζ³-allyl)Cl(CO)(PR₃)₂] complexes is stronger than that in an exo isomer, because the structural arrangement provides an optimal situation to maximize the Ru(d)-to-CO(μ*) back-bonding interaction. However, substituents at the ζ³-allyl ligand also play an important role in determining the relative stability. It was found that an anti substituent (with respect to the central allylic substituent) at one of the terminal carbons destabilizes the endo isomer more significantly and therefore makes the endo and exo isomers comparable in terms of their relative stabilities. A syn substituent at one of the terminal carbons is, however, found to have a negligible effect on the relative stability. The endo-exo interconversion was found to proceed by intervention of ζ¹-allyl intermediates instead of a direct ζ³-allyl rotation.

Full text of DOI:10.1021/om049646g

Organometallics 2004, 23, 4735-4743
4735
Isom er ism of [Ru (η3-a llyl)Cl(CO)(P P h 3)2]
Peng Xue, Siwei Bi, Herman H. Y. Sung, Ian D. Williams, Zhenyang Lin,* and
Guochen J ia*
Department of Chemistry and Open Laboratory of Chirotechnology of the Institute of
Molecular Technology for Drug Discovery and Synthesis, The Hong Kong University of Science
and Technology, Clear Water Bay, Kowloon, Hong Kong
Received May 17, 2004
Treatment of [RuHCl(CO)(PPh3)3] with CH2dCdCMe2 produced the new allyl complex
[Ru(η3-CH2CHCMe2)Cl(CO)(PPh3)2], which slowly isomerizes to [Ru(η3-CH2CMeCHMe)Cl-
(CO)(PPh3)2] in both solution and the solid state. While both endo and exo isomers can be
observed by solution NMR for the allyl complex [Ru(η3-CH2CHCMe2)Cl(CO)(PPh3)2], only
the endo isomer was detected for other related analogous allyl complexes [Ru(η3-allyl)Cl-
(CO)(PPh3)2] (e.g. allyl ) CH2CMeCHMe, CH2CHCHPh). The isomeric behavior has been
investigated by computational chemistry. The theoretical calculations show that the metal-
η3-allyl interaction in an endo isomer of [Ru(η3-allyl)Cl(CO)(PR3)2] complexes is stronger
than that in an exo isomer, because the structural arrangement provides an optimal situation
to maximize the Ru(d)-to-CO(π*) back-bonding interaction. However, substituents at the
η3-allyl ligand also play an important role in determining the relative stability. It was found
that an anti substituent (with respect to the central allylic substituent) at one of the terminal
carbons destabilizes the endo isomer more significantly and therefore makes the endo and
exo isomers comparable in terms of their relative stabilities. A syn substituent at one of the
terminal carbons is, however, found to have a negligible effect on the relative stability. The
endo-exo interconversion was found to proceed by intervention of η1-allyl intermediates
instead of a direct η3-allyl rotation.
In tr od u ction
Ruthenium η3-allyl complexes have attracted much
attention, because they can mediate organometallic and
catalytic reactions.1 It is well established that π-allyl
complexes can exist in two isomers due to the relative
orientations of the allyl ligands.1d,2 Understanding the
factors controlling the relative energy of the isomers
should help to understand the regio- and stereochem-
istry of reactions mediated by π-allyl complexes.
This work concerns the isomerism of ruthenium η3-
allyl complexes of the type [Ru(η3-allyl)Cl(CO)(L)2]. It
has been demonstrated that [Ru(η3-allyl)Br(CO)3] exists
as both endo (1A) and exo (1B) isomers in a ratio of 40:1
in CDCl3.3 In the endo isomer, the central allylic
hydrogen points toward the CO; in the exo isomer, the
central allylic hydrogen points toward the halide. An
early theoretical study showed that the endo isomer is
ca. 1.7 kcal/mol more stable than the exo isomer.4
A
* To whom correspondence should be addressed. E-mail:
chzlin@ust.hk (Z.L.); chjiag@ust.hk (G.J .).
number of ruthenium η3-allyl complexes [Ru(η3-allyl)-
Cl(CO)(PR3)2] ((PR3)2 ) (PPh3)2,5 (PMe2Ph)2,6 dppf7)
have also been prepared. In principle, these complexes
(1) See for example: (a) Yamamoto, Y.; Nakagai, Y.; Itoh, K. Chem.
Eur. J . 2004, 10, 231. (b) Mbaye, M. D.; Demerseman, B.; Renaud, J .
L.; Toupet, L.; Bruneau, C. Angew. Chem., Int. Ed. 2003, 5066. (c)
Sasabe, H.; Nakanishi, S.; Takata, T. Inorg. Chem. Commun. 2003, 6,
1140. (d) Braun, T.; Mu¨nch, G.; Windmu¨ller, B.; Gevert, O.; Laubender,
M.; Werner, H. Chem. Eur. J . 2003, 2516. (e) Chamberlain, B.; Duckett,
S. B.; Lowe, J . P.; Mawby, R. J .; Stott, J . C. Dalton 2003, 2603. (f)
Castro, A.; Turner, M. L.; Maitlis, P. M. J . Organomet. Chem. 2003,
674, 45. (g) Kondo, T.; Ono, H.; Satake, N.; Mitsudo, T.; Watanabe, Y.
Organometallics 1995, 14, 1945.
(2) See for example; (a) Norman, D. W.; Ferguson, M. J .; Stryker,
J . M. Organometallics 2004, 23, 2015. (b) Cedeno, D. L.; Weitz, E.
Organometallics 2003, 22, 2652. (c) Takao, Y.; Takeda, T.; Watanabe,
J .; Setsune, J . Organometallics 2003, 22, 233 (d) Staveren, D. R.; Bill,
E.; Bothe, E.; Bu¨hl, M.; Weyhermu¨ller, T.; Metzler-Nolte, N. Chem.
Eur. J . 2002, 8, 1649.
(4) Sakaki, S.; Ohki, T.; Takayama, T.; Sugimoto, M.; Kondo, T.;
Mitsudo, T. Organometallics 2001, 20, 3145.
(5) (a) Sasabe, H.; Nakanishi, S.; Takata, T. Inorg. Chem. Commun.
2002, 5, 177. (b) Hiraki, K.; Matsunaga, T.; Kawano, H. Organome-
tallics 1994, 13, 1878. (c) Hill, A. F.; Ho, C. T.; Wilton-Ely, D. E. T.
Chem. Commun. 1997, 2207. (d) Hiraki, K.; Ochi, N.; Sasada, Y.;
Hayashida, H.; Fuchita, Y.; Yamanaka, S. J . Chem. Soc., Dalton Trans.
1985, 873.
(6) Barnard, C. F. J .; Daniels, B. J . A.; Holland, P. R.; Mawby, R. J .
J . Chem. Soc., Dalton Trans. 1980, 2418.
(7) Cadierno, V.; Crochet, P.; Diez, J .; Garcia-Garrido, S. E.; Gimeno,
J . Organometallics 2003, 22, 5226.
(3) Wuu, Y. M.; Wrighton, M. S. Organometallics 1988, 7, 1839.
10.1021/om049646g CCC: $27.50 © 2004 American Chemical Society
Publication on Web 08/27/2004

4736 Organometallics, Vol. 23, No. 20, 2004
Xue et al.
Sch em e 1
could also have the two isomers 2A and 2B. However,
the expected isomerism was not observed previously and
only the endo isomer was detected in solution for the
reported ruthenium η3-allyl complexes of the type [Ru-
(η3-allyl)Cl(CO)(PR3)2].
During the course of investigating the reactivities of
[RuHCl(CO)(PPh3)3] toward allenes, we have prepared
the allyl complex [Ru(η3-CH2CHCMe2)Cl(CO)(PPh3)2]
and found that the allyl complex [Ru(η3-CH2CHCMe2)-
Cl(CO)(PPh3)2] does exist in two isomers in solution. The
main purpose of this work is to address the question
why two isomers were observed for [Ru(η3-CH2CHCMe2)-
Cl(CO)(PPh3)2] but only one isomer was observed for
other related reported allyl complexes [Ru(η3-allyl)Cl-
(CO)(PR3)2].
Resu lts a n d Discu ssion
Syn th esis of Allyl Com p lexes. Treatment of [Ru-
HCl(CO)(PPh3)3] (3) with CH2dCdCMe2 produced the
new allyl complex [Ru(η3-CH2CHCMe2)Cl(CO)(PPh3)2]
(4) (Scheme 1). Related complexes of the same metal
fragment with other allyl ligands (e.g. CH2CHCH2, CH2-
CMeCH2, MeCHCHCHMe, EtCHCHCHCO2Et) are
known.5 In all the reported complexes, the substituents
at the terminal allylic carbons are syn to the central
allylic proton or substituent. Complex 4 appears to be
the first example of [Ru(η3-allyl)Cl(CO)(PPh3)2] with a
terminal substituent anti to the central allyl proton or
substituent. Complex 4 exists as two isomers (4A and
4B) in solution, as indicated by the NMR spectroscopic
data (see discussion below).
In solution and in the solid state, complex 4 slowly
isomerizes to the known allyl complex 5 (Scheme 1),
which was prepared previously alternatively from the
reaction of RuHCl(CO)(PPh3)3 with isoprene.5d The
structure of 5 has been confirmed by X-ray diffraction
studies. The molecular structure of 5 is shown in Figure
1. The crystallographic details and selected bond dis-
tances and angles are given in Tables 1 and 2, respec-
tively. The structure of complex 5 can be described as a
trigonal bipyramid with the two PPh3 groups and the
allyl ligand at the equatorial positions and the Cl and
CO ligands at the axial positions. The P(1)-Ru(1)-P(2)
angle is 107.73(3)°. The allyl complex is in the endo
F igu r e 1. Molecular structure of [Ru(η3-CH2CMeCHMe)-
Cl(CO)(PPh3)2].
form, in which the methyl on the central allylic carbon
points toward the CO. The C-C bond distances of the
allyl ligand together with Ru-C(allyl) are similar to
those of reported ruthenium allyl complexes.1b-d,g,7
A proposed mechanism for the isomerization of 4 to
5 is shown in Scheme 2. The bonding mode of the allyl
ligand in complex 4 can change from η3 to η1 to give
intermediate A, which could undergo â-H elimination
to give intermediate B and then C. Intermediate C could
undergo an insertion reaction to give 5.
NMR P r op er ties of Com p lex 4. At room tempera-
1
ture, both the 31P and the H NMR spectra of 4 exhibit
broad peaks, indicating that the complex is fluxional.
To study the fluxional behavior, VT NMR experiments
were carried out. As an illustration, the variable-
temperature (298 to 242 K) 31P{1H} NMR spectra (in
toluene-d8) are shown in Figure 2. In the room-temper-

Isomerism of [Ru(η3-allyl)Cl(CO)(PPh3)2]
Organometallics, Vol. 23, No. 20, 2004 4737
Sch em e 2
Ta ble 1. Cr ysta l Da ta a n d Str u ctu r e Refin em en t
Deta ils for [Ru (η3-CH2CMeCHMe)Cl(CO)-
(P P h 3)2] (5)
formula
fw
C42H39ClOP2Ru
758.19
cryst syts
space group
a, Å
b, Å
c, Å
R, deg
â, deg
γ, deg
V, Å3
triclinic
P1h
9.9594(10)
10.1567(10)
18.8676(18)
87.811(2)
82.016(2)
67.2980(10)
1743.4(3)
2
F igu r e 2. Variable-temperature 31P{1H} NMR spectra of
[Ru(η3-CH2dCHCMe2)Cl(CO)(PPh3)2] in toluene-d8. The
signals marked with an asterisk are due to [Ru(η3-CH2-
CMeCHMe)Cl(CO)(PPh3)2].
Z
d
calcd, g cm-3
1.444
θ range, deg
no. of rflns collected
no. of indep rflns
2.17-25.00
11 918
5915 (R(int) ) 0.0293)
5915/0/424
R1 ) 4.60, wR2 ) 11.25
1.038
1.004
-0.896
to 50 °C, the broad signal changes to a sharp singlet at
34.6 ppm. When the temperature is lowered to 263 K,
the broad peak observed at room temperature is sepa-
rated into three broad peaks at 30.9, 34.2, and 38.6 ppm.
On further lowering of the temperature, these peaks
become sharp and then change to four signals at 30.1,
35.0, 38.4, and 38.8 ppm when the temperature reaches
242 K. The major signals at 30.1 and 38.8 ppm are
associated with each other with a coupling constant of
3.6 Hz, while the minor signals at 35.0 and 38.4 ppm
are associated with each other with no resolvable
coupling. The NMR data indicate that the allyl complex
[Ru(η3-CH2CHCMe2)Cl(CO)(PPh3)2] (4) exists as two
isomers which are in fast equilibrium in solution. The
more intense 31P signals at 242 K at 30.1 and 38.8 ppm
can be attributed to the endo isomer 4A, and the minor
signals at 35.0 and 38.4 ppm can be attributed to the
exo isomer 4B. Two 31P signals for each of the isomers
are expected because of the presence of the unsym-
metrical allyl ligand. The observation of a single 31P
signal in the 31P{1H} NMR spectra at room temperature
or above can be explained by the fast interconversion
of the isomers via η1-allyl intermediates.
no. of data/restraints/params
final R indices (I > 2σ(I)), %
goodness of fit
largest diff peak, e Å-3
largest diff hole, eÅ-3
Ta ble 2. Selected Bon d Len gth s (Å) a n d An gles
(d eg) for [Ru (η3-CH2CMeCHMe)Cl(CO)(P P h 3)2] (5)
Ru(1)-C(10)
Ru(1)-C(3)
Ru(1)-P(1)
Ru(1)-Cl(1)
C(1)-C(2)
1.813(5)
2.261(5)
2.3985(10)
2.4852(10)
1.399(8)
1.536(7)
Ru(1)-C(2)
Ru(1)-C(1)
Ru(1)-P(2)
O(1)-C(10)
C(2)-C(3)
C(3)-C(4)
2.245(4)
2.276(5)
2.3996(10)
1.105(5)
1.396(9)
1.506(8)
C(2)-C(5)
C(10)-Ru(1)-C(2)
C(2)-Ru(1)-C(3)
C(2)-Ru(1)-C(1)
C(10)-Ru(1)-P(1)
C(3)-Ru(1)-P(1)
C(10)-Ru(1)-P(2)
C(3)-Ru(1)-P(2)
P(1)-Ru(1)-P(2)
P(1)-Ru(1)-Cl(1)
C(2)-C(1)-Ru(1)
C(3)-C(2)-C(5)
C(3)-C(2)-Ru(1)
C(5)-C(2)-Ru(1)
C(2)-C(3)-Ru(1)
84.94(19) C(10)-Ru(1)-C(3)
98.57(19)
36.1(2)
36.1(2)
C(10)-Ru(1)-C(1) 101.25(19)
C(3)-Ru(1)-C(1)
64.8(2)
119.1(2)
87.41(14)
133.07(19)
159.04(15)
88.58(13) C(2)-Ru(1)-P(1)
152.10(17) C(1)-Ru(1)-P(1)
93.71(14) C(2)-Ru(1)-P(2)
98.73(17) C(1)-Ru(1)-P(2)
107.73(3) C(10)-Ru(1)-Cl(1) 176.97(14)
90.98(3) P(2)-Ru(1)-Cl(1)
83.57(3)
120.7(5)
117.9(6)
73.2(3)
124.6(5)
128.8(4)
70.8(3)
121.2(5)
72.6(3)
121.9(4)
71.3(3)
C(3)-C(2)-C(1)
C(1)-C(2)-C(5)
C(1)-C(2)-Ru(1)
C(2)-C(3)-C(4)
C(4)-C(3)-Ru(1)
1
The fluxional behavior is also reflected in the VT H
NMR spectrum. At 213 K, the 1H NMR spectrum
showed the proton signals of the allyl group of 4A at
1.38 (3 H, CH3), 1.88 (3 H, CH3), 3.16 (1 H, CH2), 3.51
(1 H, CH2), and 4.96 (central CH) ppm and those of 4B
at 1.19 (3 H, CH3), 1.65 (3 H, CH3), 2.95 (1 H, CH2),
3.31 (1 H, CH2), and 5.75 (1 H, central CH) ppm. On
O(1)-C(10)-Ru(1) 177.3(4)
ature (298 K) 31P{1H} NMR spectrum, complex 4
exhibits a broad peak at 35 ppm (the two minor doublets
marked with asterisks at 31.7 (d, J ) 2.7 Hz) and 38.5
ppm are due to [Ru(η3-CH2CMeCHMe)Cl(CO)(PPh3)2]
(5), which was formed by the isomerization of 4 as
described above). When the temperature is increased
1
the basis of the H NMR integration, the ratio of 4A to
1
4B was estimated to be 2.4:1. The H NMR spectrum

4738 Organometallics, Vol. 23, No. 20, 2004
Xue et al.
Sch em e 3
at room temperature showed only one set of proton
signals for the allyl group at 1.30 (3 H, CH3), 1.66 (3 H,
CH3), 3.08 (2 H, CH2), and 5.21 (1 H, central CH) ppm.
The NMR data suggest that the two terminal CH2
protons are exchanging more rapidly than the two
terminal CMe2 methyl groups, and the fluxional behav-
ior can be explained by the mechanism shown in Scheme
3. A similar mechanism was proposed previously for the
intercoversion of endo and exo isomers of [(MeSi(CH2-
PMe2)3)RuMe(η3-CH2CMeCH2)].8
NMR P r op er ties of Com p lexes 5 a n d 6. In con-
1
trast to 4, only the endo isomer can be detected by H
and 31P NMR in solution for the complex [Ru(η3-CH2-
1
CMeCHMe)Cl(CO)(PPh3)2] (5). In addition, the H and
31P NMR signals are sharp at room temperature,
indicating that the complex is not fluxional at this
temperature. We have also carefully examined the 31P
and H NMR spectra of the related complex [Ru(η3-CH2-
1
F igu r e 3. B3LYP optimized structures of 2a and 2b
together with selected bond distances (in Å) and relative
energies (in parentheses, in kcal/mol). For clarity, the
hydrogen atoms of PH3 are omitted.
CHCHPh)Cl(CO)(PPh3)2] (6),5a which was prepared from
the reaction of 3 with CH2dCdCHPh. Like [Ru(η3-CH2-
CMeCHMe)Cl(CO)(PPh3)2] (5), the 31P{1H} NMR spec-
trum of [Ru(η3-CH2CHCHPh)Cl(CO)(PPh3)2] (6) at room
temperature (298.7 K) showed two sharp doublets at
26.5 and 38.9 ppm with a coupling constant of 5.1 Hz,
indicating that the complex is not fluxional and has only
one isomer observable by NMR at this temperature.
Consistent with this proposition, the room-temperature
1H NMR spectrum showed four allylic proton signals
at 3.26 (CH2), 3.61 (CH2), 4.61 (CHPh), and 5.85 (central
CHCH2)Cl(CO)(PPh3)2]. Figure 3 shows the endo and
exo structures calculated for the model η3-allyl complex
[Ru(η3-CH2CHCH2)Cl(CO)(PH3)2] together with selected
bond distances. The endo isomer (2a ) is lower in energy
than the exo one (2b) by 2.64 kcal/mol, from which the
ratio of the endo and exo isomers is calculated to be 86:1
at room temperature (298.15 K) according to the Boltz-
mann distribution. Such a large ratio implies that the
amount of the exo isomer is too small to be detected
easily by NMR in solution, consistent with the experi-
mental observations mentioned above.5-7 Here, we
purposely label the structures with lower-case letters
to distinguish them from those described in the Experi-
mental Section, which are labeled with upper-case
letters.
1
CH) ppm. We also carried out VT H and 31P{1H} NMR
experiments for the compound, and the exo isomer was
also not observed in the temperature range 298-363 K
in toluene.
Th eor etica l Stu d y. It is interesting to note that two
isomers are observed for [Ru(η3-CH2CHCMe2)Cl(CO)-
(PPh3)2], but only one isomer is observed for other
related allyl complexes [Ru(η3-CH2CMeCHMe)Cl(CO)-
(PPh3)2] and [Ru(η3-CH2CHCHPh)Cl(CO)(PPh3)2]. To
better understand the substituent effect on the behavior,
a computational study has been carried out.
Before exploring the substituent effect on the isomer-
ism of substituted-allyl complexes, we first discuss the
isomeric behavior of the η3-allyl complex [Ru(η3-CH2-
CHCH2)Cl(CO)(PH3)2], a model complex for [Ru(η3-CH2-
Clearly, the relative stabilities of the endo and exo
isomers determine whether both isomers of the η3-allyl
complex can be detected in solution. The question now
is what factors determine the relative stability of the
two isomers. Comparing the two calculated structures
(Figure 3), we can see that the two structures have
similar Ru-CO, Ru-P, and Ru-Cl bond distances.
However, noticeable differences between the two struc-
tures can be found in the bond distances associated with
(8) Mcneill, K.; Anderson, R. A.; Bergman, R. G. J . Am. Chem. Soc.
1997, 119, 11244.

Isomerism of [Ru(η3-allyl)Cl(CO)(PPh3)2]
Organometallics, Vol. 23, No. 20, 2004 4739
Sch em e 4
Sch em e 5
Ru-C(terminal) bonds form larger angles (100.7°) with
CO. The optimal structural arrangement for bonding
interactions is achieved, and therefore, the Ru-C(ter-
minal) bonds are relatively stronger. In the exo isomer,
the two Ru-C(terminal) bonds have quite small angles
(88.3°) with the CO ligand. This structural arrangement
does not provide an optimal situation to maximizing the
Ru(d)-to-CO(π*) back-bonding interaction. Thus, the
optimal structural arrangement is not met. The overall
effect in maintaining the Ru-CO interaction is the
weakening of the Ru-C(terminal) bonds. We can con-
clude that the greater stability of the endo isomer is
related to its stronger Ru-C(terminal) bonds due to
relatively less weakening caused by the cis Ru-CO
bond.
Ta ble 3. Resu lts of Na tu r a l Bon d Or d er a n d
P op u la tion An a lyses for
[Ru (η3-CH2CHCH2)Cl(CO)(P H3)2]
Wiberg bond index
natural atomic charge
isomer Ru-C(terminal) Ru-C(central) C(terminal) C(central)
2a (endo)
2b(exo)
0.4004
0.3764
0.1718
0.1831
-0.361
-0.350
-0.105
-0.109
the Ru-η3-allyl structural units. The Ru-C(terminal)
bonds (between Ru and the terminal carbons of the η3-
allyl ligand) in the endo isomer are stronger than those
in the exo isomer. In contrast, the Ru-C(central) bond
in the exo isomer is stronger than that in the endo
isomer. The greater stability of the endo isomer is
apparently related to its stronger Ru-C(terminal)
bonds. In contrast, the stronger Ru-C(central) bond in
the exo isomer does not render its stability. The reason
for this unusual behavior can be explained as follows.
An η3-allyl ligand is generally considered as a monoan-
ionic ligand. One of the two orbitals accommodating the
four π electrons of the η3-allyl anionic ligand has no
orbital contribution from the central carbon (Scheme 4).9
Therefore, the central carbon of the η3-allyl ligand
makes a smaller contribution to the Ru-η3-allyl bonding
interactions in comparison to the two terminal carbons,
although the Ru-C(central) bond is always shorter than
the Ru-C(terminal) bonds for a given η3-allyl complex.
Natural bond order (NBO) analyses indeed give a much
smaller Wiberg bond index for the Ru-C(central) bond
in comparison to the Ru-C(terminal) bonds (Table 3).
The calculated structural features of isomers of the
model complex [Ru(η3-CH2CHCH2)Cl(CO)(PH3)2] indi-
cate that the endo isomer has stronger Ru-C(terminal)
bonds than the exo isomer. Thus, stronger Ru-η3-allyl
bonding interactions are expected for the endo isomer,
because the Ru-C(central) bond makes a relatively
small contribution to the structural stability. A plausible
explanation for the weaker Ru-C(terminal) bonds in the
exo isomer is given below. In a carbonyl complex, the
ligands that are cis to CO are normally expected to bend
away from the carbonyl ligand in order to maximize the
M(d)-to-CO(π*) back-bonding interaction. In other words,
large angles between CO and its cis ligands would
increase the back-bonding interaction. Scheme 5 il-
lustrates the d-p orbital mixing, showing the increased
orbital overlap due to the bending of angles. The mixing
of metal p orbitals with d orbitals is commonly employed
to explain the enhancement of metal(d)-to-ligand(π*)
back-bonding interactions.10 In the endo isomer, the two
To understand how substituents on the η3-allyl ligand
affect the relative stabilities of the endo and exo isomers,
we discuss the isomerism of monomethyl-substituted η3-
allyl complexes. Figure 4 shows six isomeric structures
calculated for the monomethyl-substituted η3-allyl com-
plex. In the endo (7a ) and exo (7b) isomers, the methyl
substituent is syn with respect to the substituent at the
central carbon of the allyl ligand, while in the endo (7a ′)
and exo (7b′) isomers, the methyl substituent is anti
with respect to the substituent at the central carbon of
the allyl ligand. In the endo (7a ′′) and exo (7b′′) isomers,
the methyl substituent is at the central C2 atom of the
allyl ligand. The energy difference between 7a and 7b
(2.52 kcal/mol) is close to that between 2a and 2b (2.64
kcal/mol), indicating that the syn-methyl substituent
has a negligible effect on the energy difference between
the endo and exo isomers. The C4-C1-Ru-C5 dihedral
angle in 7a is 47.4° and the C4-C1-Ru-Cl angle in
7b is 48.5°, giving staggered arrangements of the methyl
substituent with respect to the Ru-CO or Ru-Cl and
Ru-PH3 bonds. The staggered arrangements in 7a and
7b minimize the steric repulsion caused by the methyl
substituent and are clearly responsible for the negligible
effect. For 7a ′ and 7b′, which have an anti methyl
substituent, the energy difference (0.77 kcal/mol) be-
tween the endo (7a ′) and exo (7b′) isomers is much
smaller than that between 2a and 2b (2.64 kcal/mol),
indicating that the anti methyl substituent has a
significant effect on the energy difference. The C4-C1-
Ru-Cl dihedral angle in 7a ′ is 5.3° and the C4-C1-
Ru-C5 one in 7b′ is 2.1°, giving eclipsed arrangements
of the methyl substituent with respect to the Ru-Cl or
Ru-CO bond. Clearly, steric repulsion due to the
eclipsed arrangements is expected to destabilize both
the endo (7a ′) and exo (7b′) isomers. The relatively small
energy difference between 7a ′ and 7b′ in comparison to
that between 2a and 2b, which do not have a methyl
substituent, suggests that the destabilization in 7b′ is
more significant than in 7a ′. In other words, the 1,4-
repulsion between a methyl group and a chloride ligand
is greater than that between a methyl group and a
(9) Crabtree, R. H. The Organometallic Chemistry of the Transition
Metals, 3rd ed.; Wiley: New York, 2001; p 122.
(10) Albright, T. A.; Burdett, J . K.; Whangbo, M.-H. Orbital Interac-
tions in Chemistry; Wiley: New York, 1985; p 321.

4740 Organometallics, Vol. 23, No. 20, 2004
Xue et al.
F igu r e 4. B3LYP optimized structures of the monomethyl-substituted η3-allyl complex together with selected bond distances
(in Å) and relative energies (in parentheses, in kcal/mol). For clarity, the hydrogen atoms of PH3 are omitted.
carbonyl ligand. A plausible explanation for this obser-
vation is that the negative charge carried by the chloride
exerts a greater repulsive force. Further support to the
plausible explanation comes from the energy difference
calculated for 7a ′′ and 7b′′. The methyl substituent
eclipses the Ru-CO bond in 7a ′′ and the Ru-Cl bond
in 7b′′. The increased energy difference between 7a ′′ and
7b′′ is clearly a result of the large steric repulsion
between the methyl substituent and the Ru-Cl bond
in 7b′′.
Supports to the steric argument above can also be
found from our additional calculations on the relative
stabilities of the endo and exo isomers having an anti
substituent bulkier than methyl, such as tert-butyl. 7a ′
(endo) is more stable than 7b′ (exo) by 0.77 kcal/mol.
When the methyl groups in 7a ′ and 7b′ are replaced by
tert-butyl, the stability order is reversed, the exo isomer
being more stable than the endo isomer by 1.29 kcal/
mol. 7a is more stable than 7a ′ by 4.41 kcal/mol, while
7b is more stable than 7b′ by 2.66 kcal/mol. For the
methyl-substituted free allyl, the syn form adopted by
7a and 7b was calculated to be more stable by only ca.
1.0 kcal/mol than the anti form adopted by 7a ′ and 7b′.
These results suggest that the eclipsed arrangements
in 7a ′ and 7b′ (with respect to Ru-Cl and Ru-CO,
respectively) contribute significantly to their instability.
Figure 5 shows four isomeric structures calculated for
the dimethyl-substituted η3-allyl model complex. In the
endo (4a ) and exo (4b) isomers, the two methyl substit-
uents are at the same terminal carbon atom of the η3-
allyl ligand, while in the endo (5a ) and exo (5b) isomers
one methyl substituent is at one terminal carbon and
the other is at the central one. The results of calcula-
tions show that the exo (4b) isomer is higher in energy
than the endo (4a ) one only by 0.25 kcal/mol (4a :4b )
100:66, 298.15 K), consistent with the experimental
observations that both the endo and exo isomers coexist
in solution. From the discussion above, we know that a
syn-methyl substituent has a negligible effect on the
energy difference between the endo and exo isomers.

Isomerism of [Ru(η3-allyl)Cl(CO)(PPh3)2]
Organometallics, Vol. 23, No. 20, 2004 4741
F igu r e 5. B3LYP optimized structures of the dimethyl-substituted η3-allyl complex together with selected bond distances
(in Å) and relative energies (in parentheses, in kcal/mol). For clarity, the hydrogen atoms of PH3 are omitted.
Therefore, the energy difference between 4a and 4b is
small, similar to that found between 7a ′ and 7b′,
because an anti substituent destabilizes the endo isomer
(4a or 7a ′) more in comparison to the exo isomer (4b or
7b′). Experiments described above show that isomers
4A and 4B slowly isomerize to 5 having an endo
structural form. Our calculations indeed show that 5a
is the most stable among the four isomers (Figure 5).
For 5a and 5b, the energy difference (4.27 kcal/mol) is
large. These results are again consistent with the
experimental observation that only one isomer was
observed by NMR in solution. The energy difference
between 5a and 5b is also similar to that found between
7a ′′and 7b′′, consistent with the notion that a syn
substituent has a negligible effect on the energy differ-
ence, due to its staggered arrangements with respect
to the Ru-CO or Ru-Cl and Ru-PH3 bonds.
Calculations were also done for [Ru(η3-CH2CHCHPh)-
Cl(CO)(PH3)2], a model complex of 6. The energy dif-
ference between the endo (6a ) and exo (6b) isomers was
found to be 2.00 kcal/mol (Figure 5), comparable to that
between 2a and 2b, although it is slightly smaller. The
argument used to understand the energy difference
between 2a and 2b can be applied here also. Experi-
ments detect only the endo isomer of 6, suggesting that

4742 Organometallics, Vol. 23, No. 20, 2004
Xue et al.
F igu r e 6. Energy profile for the 2a (exo) f 2b(endo) interconversion process.
the amount of the exo isomer is too small to be detected
under the experimental conditions.
at the η3-allyl ligand also play an important role in
determining the relative stability. We found that an anti
substituent at one of the terminal carbons destabilizes
the endo isomer more significantly and therefore makes
the endo and exo isomers comparable in terms of their
relative stability. The comparable stability allows both
isomers to be observable in solution. A syn substituent
is, however, found to have a negligible effect on the
relative stability. The endo-exo interconversion was
found to proceed by intervention of η1-allyl intermedi-
ates instead of a direct η3-allyl rotation.
The mechanism of the exo-endo interconversion is
another interesting point to be addressed computation-
ally. The calculations support our proposal that η1-allyl
intermediates are involved in the interconversion pro-
cess. Figure 6 shows the energy profile for the 2a (exo)
f 2b(endo) process. Rearrangement of η3-allyl in 2a to
η1-allyl in the intermediate IN1 requires a barrier of
23.46 kcal/mol. A barrierless rotation of the η1-allyl
ligand along the Ru-η1-allyl bond in IN1 gives another
η1-allyl intermediate, IN2. IN2 differs from IN1 only
in the orientation of the η1-allyl ligand. Formation of
the endo isomer 2b from IN2 requires a barrier of only
7.54 kcal/mol. An attempt to locate transition states for
the possible mechanism involving a direct η3-allyl rota-
tion failed. Keeping the η3-allyl bonding mode with
constrained geometry optimizations, we obtained a
series of η3-allyl structures along the direct rotation
path. From these partially optimized structures, transi-
tion states were searched. These calculations gave either
the local minima we have already obtained or the
transition states that are relevant to the mechanism
presented in Figure 6. Therefore, a direct η3-allyl
rotation for the exo-endo interconversion seems un-
likely.
Exp er im en ta l Section
All manipulations were carried out under a nitrogen atmo-
sphere using standard Schlenk techniques, unless otherwise
stated. Solvents were distilled under nitrogen from sodium
benzophenone (hexane, ether, THF), sodium (benzene), or
calcium hydride (CH2Cl2). The starting materials [RuHCl(CO)-
(PPh3)3]11 and phenylallene12 were prepared according to the
literature methods. All other reagents were used as purchased
from Aldrich Chemical Co.
Microanalyses were performed by M-H-W Laboratories
(Phoenix, AZ). 1H, 13C{1H}, and 31P{1H} NMR spectra were
collected on a Bruker ARX-300 spectrometer (300 MHz). 1H
and 13C NMR shifts are relative to TMS, and 31P chemical
shifts are relative to 85% H3PO4.
[Ru (η3-CH2CHCMe2)Cl(CO)(P P h 3)2] (4). To a stirred
solution of [RuHCl(CO)(PPh3)3] (0.512 g, 0.538 mmol) in
dichloromethane (24 mL) was added dropwise a solution of
3-methyl-1,2-butadiene (0.054 mL, 0.54 mmol) in dichlo-
romethane (9 mL) at room temperature. After the addition was
complete, the resulting mixture was stirred for 20 min. The
volatile materials were then removed under vacuum. Addition
of diethyl ether (5 mL) to the residue gave an orange solid,
which was collected by filtration, washed with ether (5 mL ×
3), a mixture of THF/ether (1:4 v/v, 5 mL), and hexane (5 mL
× 2) in turn, and then dried under vacuum. Yield: 342 mg,
84%. 1H NMR (300 MHz, CD2Cl2, 298 K): δ 0.81 (s, 3 H, CH3),
1.29 (s, 3 H, CH3), 2.52 (br d, J (HH) ) 10.1 Hz, 2 H, CH2),
4.77 (br t, J (HH) ) 10.1 Hz, 1 H, central CH), 7.15-7.30 (m,
18 H, phenyl), 7.35-7.41 (m, 12 H, phenyl). 31P{1H} NMR
Con clu sion s
The allyl complex [Ru(η3-CH2CHCMe2)Cl(CO)(PPh3)2]
has been prepared from the reaction of CH2dCdCMe2
with [RuHCl(CO)(PPh3)3]. In solution, both endo and exo
isomers were observed by NMR for the new allyl
complex [Ru(η3-CH2CHCMe2)Cl(CO)(PPh3)2]. In con-
trast, only one isomer was detected for other related
analogous allyl complexes [Ru(η3-allyl)Cl(CO)(PPh3)2]
(e.g. allyl ) CH2CMeCHMe, CH2CHCHPh). The theo-
retical calculations show that, for the [Ru(η3-allyl)Cl-
(CO)(PR3)2] complexes, the metal-η3-allyl interaction
in an endo isomer is stronger than that in an exo isomer.
In the endo isomer, the two Ru-C(terminal) bonds form
larger angles with the CO ligand and are stronger
because the structural arrangement provides an optimal
situation to maximize the Ru(d)-to-CO(π*) back-bonding
interaction. Intrinsically, an endo isomer is expected to
be more stable than its exo form. However, substituents
1
(121.5 MHz, CD2Cl2, 298 K): δ 35.8 (br). H NMR (300 MHz,
toluene-d8, 299 K): δ 1.30 (br s, 3 H, CH3), 1.66 (br s, 3 H,
CH3), 3.08 (d, J (HH) ) 9.9 Hz, 2 H, CH2), 5.21 (br t, J (HH) )
(11) Ahmad, N.; Levison, J . J .; Robinson, S. D.; Uttley, M. F.;
Wonchoba, E. R.; Parshall, G. W. Inorg. Synth. 1974, 15, 45.
(12) Trost, B. M.; Pinkerton, A. B.; Seidel, M. J . Am. Chem. Soc.
2001, 123, 12466.

Isomerism of [Ru(η3-allyl)Cl(CO)(PPh3)2]
Organometallics, Vol. 23, No. 20, 2004 4743
9.9 Hz, 1 H, central CH), 7.20 (m, 18 H, phenyl), 7.88 (m, 12
H, phenyl). 31P{1H} NMR (121.5 MHz, toluene-d8, 298 K): δ
35.0 (br). Selected 1H NMR data of 4A at 213 K (300 MHz,
toluene-d8): δ 1.38 (br s, 3 H, CH3), 1.88 (br s, 3 H, CH3), 3.16
(br s, 1 H, CH2); 3.51 (br d, J (HH) ) 11.8 Hz, 1 H, CH2), 4.96
(br m, 1 H, central CH). Selected 1H NMR data of 4B at 213
K (300 MHz, toluene-d8): δ 1.19 (br s, 3 H, CH3), 1.65 (br s, 3
H, CH3), 2.95 (br d, J (HH) ) 10.8 Hz, 1 H, CH2), 3.31 (br s, 1
H, CH2), 5.75 (br m, 1 H, central CH). 31P{1H} NMR (121.5
MHz, toluene-d8, 213 K): δ 31.3 (d, J (PP) ) 3.8 Hz, 4A), 36.1
(s, 4B), 38.0 (s, 4B), 39.3 (d, J (PP) ) 3.8 Hz, 4A). Anal.
Found: C, 66.91; H, 5.29. Calcd for C42H39ClOP2Ru: C, 66.53;
H, 5.18.
level of density functional theory.13 Frequency calculations at
the same level of theory have also been performed to identify
all stationary points as minima (zero imaginary frequency).
The LANL2DZ effective core potentials and basis sets14 were
used to describe Ru, Cl, and P, while the standard 6-31G basis
set15 was used for C, H, and O atoms. Polarization functions16
were also added for C (úd ) 0.600), O (úd ) 1.154), P (úd
)
0.340), and Cl (úd ) 0.514). For computational simplicity, the
PPh3 ligand used in experiments was modeled by PH3. All
calculations were performed with the Gaussian 98 software
package.17
[Ru (η3-CH2CMeCHMe)Cl(CO)(P P h 3)2] (5). A solution of
[Ru(η3-CH2CHCMe2)Cl(CO)(PPh3)2] (35.4 mg) in toluene (4 mL)
was heated at 60 °C for 4 h. After the solvent was removed
under vacuum, the residue was washed with diethyl ether and
hexane in turn and dried in vacuo. Yield: 10.5 mg, 30%.
Compound 5 was also obtained in 41% isolated yield when a
solution of [Ru(η3-CH2dCHCMe2)Cl(CO)(PPh3)2] (111 mg) in
dichloromethane (2 mL) was allowed to stand at ambient
Ack n ow led gm en t. We acknowledge financial sup-
port from the Hong Kong Research Grants Council
(HKUST 6090/02P, HKUST 6087/02P, and DAG03/
04.SC15), the University Grants Committee of Hong
Kong, through the Area of Excellence Scheme (Aoe).
Su p p or tin g In for m a tion Ava ila ble: Tables of bond
distances and angles, atomic coordinates and equivalent
isotropic displacement coefficients, and anisotropic displace-
ment coefficients for [Ru(η3-CH2CMeCHMe)Cl(CO)(PPh3)2] (5)
and Cartesian coordinates of all the calculated structures
reported in this article; X-ray data for 5 are also given as a
CIF file. This material is available free of charge via the
Internet at http://pubs.acs.org.
1
temperature for 1 week. H NMR (300 MHz, CD2Cl2, 298 K):
δ 0.93 (br s, 3 H, CH3), 1.89 (s, 3 H, CH3), 2.60 (d, J (PH) ) 4.3
Hz, 1 H, CH2), 2.67 (br s, 1 H, CH2), 3.11 (br m, 1 H, CH),
7.18-7.36 (m, 24 H, phenyl), 7.48 (t, 6 H, phenyl). 31P{1H}
NMR (121.5 MHz, CD2Cl2, 298 K): δ 31.7 (d, J (PP) ) 2.7 Hz),
38.5 (d, J (PP) ) 2.7 Hz).
[R u (η3-CH2CHCHP h )Cl(CO)(P P h 3)2] (6). This complex
was prepared from [RuHCl(CO)(PPh3)3] and phenylallene as
a yellow powder, following the procedure for [Ru(η3-CH2-
CHCMe2)Cl(CO)(PPh3)2]. Yield: 87%. 1H NMR (300 MHz,
toluene-d8, 298 K): δ 3.25 (m, 1 H, CH2), 3.61 (dd, J (HH)
)12.4, J (PH) ) 5.6 Hz, CH2), 4.61 (dd, J (HH) ) 11.9, J (PH)
) 6.7 Hz, CHPh), 5.85 (m, 1 H, CdCHC), 7.08-7.22 (m, 23 H,
phenyl), 7.76-7.84 (m, 12 H, phenyl). 31P{1H} NMR (121.5
MHz, toluene-d8, 298 K): δ 26.5 (d, J (PP) ) 5.1 Hz), 38.9 (d,
J (PP) ) 5.1 Hz).
OM049646G
(13) (a) Becke, A. D. J . Chem. Phys. 1993, 98, 5648. (b) Miehlich,
B.; Savin, A.; Stoll, H.; Preuss, H. Chem. Phys. Lett. 1989, 157, 200.
(c) Lee, C.; Yang, W.; Parr, G. Phys. Rev. B 1988, 37, 785.
(14) (a) Hay, P. J .; Wadt, W. R. J . Chem. Phys. 1985, 82, 270. (b)
Wadt, W. R.; Hay, P. J . J . Chem. Phys. 1985, 82, 284. (c) Hay, P. J .;
Wadt, W. R. J . Chem. Phys. 1985, 82, 299.
(15) (a) Gordon, M. S. Chem. Phys. Lett. 1980, 76, 163. (b) Hariha-
ran, P. C.; Pople, J . A. Theor. Chim. Acta 1973, 28, 213. (c) Binning,
R. C., J r.; Curtiss, L. A. J . Comput. Chem. 1990, 11, 1206.
(16) Huzinaga, S. Gaussian Basis Sets for Molecular Calculations;
Elsevier Science: Amsterdam, 1984.
(17) Frisch, M. J .; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.;
Robb, M. A.; Cheeseman, J . R.; Zakrzewski, V. G.; Montgomery, J . A.,
J r.; 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.; J ohnson, B. G.; Chen,
W.; Wong, M. W.; Andres, J . L.; Head-Gordon, M.; Replogle, E. S.;
Pople, J . A. Gaussian 98, revision A.9; Gaussian, Inc.: Pittsburgh, PA,
1998.
Cr ysta l Str u ctu r e An a lyses. Crystals suitable for X-ray
diffraction were grown from CH2Cl2 solutions of 5 layered with
hexane. Data collections were performed on a Bruker Apex
CCD area detector, by using graphite-monochromated Mo KR
radiation (λ ) 0.710 73 Å). Empirical absorption corrections
(SADABS) were applied. All structures were solved by direct
methods, expanded by difference Fourier syntheses, and
refined by full-matrix least squares on F2 using the Bruker
SHELXTL (version 5.10) program package. All non-hydrogen
atoms were refined anisotropically. Further details on crystal
data, data collection, and refinement are summarized in Table
1.
Com p u ta tion a l Deta ils. Molecular geometries of the
model complexes were optimized at the Becke3LYP (B3LYP)