Ligand dependence of the indenyl ring slippage in [(η5-Ind)MoL2(CO)2]0,+ complexes: Experimental and theoretical studies

Article abstract of DOI:10.1021/om971005g

The reaction of a series of cationic and neutral complexes [(η⁵-Ind)Mo(CO)₂L₂]^0,+ (1, L = NCMe; 2, L = CNMe; 3, L = OPPh₃; 4, L = OP(OMe)₃; 5, L = PMe₃; 6, L₂ = (PhSCH₂)₂; 7, L₂ = (O₂CCF₃)^-; 8, L₂ = (O₂CCH₃)^-; 9, L₂ = (O₂CPh)^-; 10, L₂ = [(NPh)₂CH]^-; 11, L₂ = [(N-p-tolyl)₂CH]^-; 12, L₂ = (S₂CNEt₂)^-; 13, L₂ = (S₂P(OEt)₂)^-; 14, L₂ = acac^-; 15, L₂ = MeC-(O)CHC(NPh)CMe) with acetonitrile produces the ring-slipped adducts [(η³-Ind)MoL2(NCMe)-(CO)₂]^0,+ only in the cases where L = NCMe, OPPh₃, OP(OMe)₃ or L₂ = (O₂CCF₃)^-, acacr. The adducts are labile and have been characterized by means of the distinctive chemical shift of the central pseudo-allylic proton of the η³-indenyl ligand. In the other cases no ring slippage is observed. In NCMe solvent 6 undergoes substitution to give 1, but the macrocycle trithiacyclononane (ttcn) reacts with 1 to give [(η³-Ind)Mo(CO)₂(κ-³-ttcn)] ⁺ as the BF₄^- salt. PMe₃ reacts with 1 to give substitution products in a stepwise fashion. [(η⁵-Ind)Mo(NCMe)-(PMe₃)(CO)₂]BF ₄ is formed instantaneously at -70 °C, but 5 forms slowly at room temperature without any detectable ring-slipped intermediate. The crystal structure of the isonitrile cation 2 is presented. 2 adds neither NCMe nor CNMe but undergoes CO substitution with excess CNMe. Extended Hu?ckel and density function calculations predict the most stable conformation of [(η⁵-Ind)MoL₂(CO)₂]⁺ (L = NCR, CNR) to be the one with the ring trans to the carbonyls, while for the [(η³-Ind)MoL₂(NCR)(CO)₂]⁺ complexes the ring lies over the carbonyls. According to the dft calculations, acetonitrile addition induces η⁵ → η³ slipping of the ring in the NCR derivatives but not in the CNR analogues, because the reaction is enthalpically driven in the former and not in the latter. The dft calculations of part of the reaction path indicate that the indenyl ring starts to slip when the incoming ligand is still far from the metal (～400 pm), thus supporting earlier kinetically based proposals.

Full text of DOI:10.1021/om971005g

Organometallics 1998, 17, 2597-2611
2597
Liga n d Dep en d en ce of th e In d en yl Rin g Slip p a ge in
[(η5-In d )MoL2(CO)2]0,+ Com p lexes: Exp er im en ta l a n d
Th eor etica l Stu d ies
Maria J . Calhorda,†,‡ Carla A. Gamelas,†,§ Isabel S. Gonc¸alves,†,|
Eberhardt Herdtweck, Carlos C. Roma˜o,*,† and Lu´ıs F. Veiros#
Instituto de Tecnologia Quı´mica e Biolo´gica, Quinta do Marqueˆs, EAN, Apt 127,
2780 Oeiras, Portugal, Departamento Quı´mica e Bioquı´mica, Faculdade de Cieˆncias,
Universidade de Lisboa, 1700 Lisboa, Portugal, Escola Superior de Tecnologia,
Instituto Polite´cnico de Setu´bal, 2900 Setu´bal, Portugal, Departamento Quı´mica,
Universidade de Aveiro, Campus de Santiago, 3800 Aveiro, Portugal, Anorganisch-chemisches
Institut, Technische Universita¨t Mu¨nchen, D-85478 Garching, Federal Republic of Germany,
and Centro de Qu´ımica Estrutural, Instituto Superior Te´cnico, 1096 Lisboa Codex, Portugal
Received November 14, 1997
The reaction of a series of cationic and neutral complexes [(η5-Ind)Mo(CO)2L2]0,+ (1, L )
NCMe; 2, L ) CNMe; 3, L ) OPPh3; 4, L ) OP(OMe)3; 5, L ) PMe3; 6, L2 ) (PhSCH2)2; 7,
L2 ) (O2CCF3)-; 8, L2 ) (O2CCH3)-; 9, L2 ) (O2CPh)-; 10, L2 ) [(NPh)2CH]-; 11, L2 ) [(N-
p-tolyl)2CH]-; 12, L2 ) (S2CNEt2)-; 13, L2 ) (S2P(OEt)2)-; 14, L2 ) acac-; 15, L2 ) MeC-
(O)CHC(NPh)CMe) with acetonitrile produces the ring-slipped adducts [(η3-Ind)MoL2(NCMe)-
(CO)2]0,+ only in the cases where L ) NCMe, OPPh3, OP(OMe)3 or L2 ) (O2CCF3)-, acac-.
The adducts are labile and have been characterized by means of the distinctive chemical
shift of the central pseudo-allylic proton of the η3-indenyl ligand. In the other cases no ring
slippage is observed. In NCMe solvent 6 undergoes substitution to give 1, but the macrocycle
-
trithiacyclononane (ttcn) reacts with 1 to give [(η3-Ind)Mo(CO)2(κ3-ttcn)]+ as the BF4 salt.
PMe3 reacts with 1 to give substitution products in a stepwise fashion. [(η5-Ind)Mo(NCMe)-
(PMe3)(CO)2]BF4 is formed instantaneously at -70 °C, but 5 forms slowly at room
temperature without any detectable ring-slipped intermediate. The crystal structure of the
isonitrile cation 2 is presented. 2 adds neither NCMe nor CNMe but undergoes CO
substitution with excess CNMe. Extended Hu¨ckel and density function calculations predict
the most stable conformation of [(η5-Ind)MoL2(CO)2]+ (L ) NCR, CNR) to be the one with
the ring trans to the carbonyls, while for the [(η3-Ind)MoL2(NCR)(CO)2]+ complexes the ring
lies over the carbonyls. According to the dft calculations, acetonitrile addition induces η5 f
η3 slipping of the ring in the NCR derivatives but not in the CNR analogues, because the
reaction is enthalpically driven in the former and not in the latter. The dft calculations of
part of the reaction path indicate that the indenyl ring starts to slip when the incoming
ligand is still far from the metal (∼400 pm), thus supporting earlier kinetically based
proposals.
In tr od u ction
of IndMLn complexes in comparison to the reactions of
their Cp analogues. This rate acceleration, or indenyl
effect as it was called by Basolo,2 certainly plays an
important role in some catalytic reactions performed by
indenyl complexes.3
Kinetic studies revealed that the indenyl effect is felt
in both dissociative (SN1) and associative (SN2) reaction
pathways. As discussed by Basolo, the latter are more
common and involve a larger rate enhancement.4 The
widely accepted explanation of this effect was first
The haptotropic rearrangement of coordinated poly-
enyl ligands is a well-established process by which the
metal adjusts its total electron count (usually 18e) upon
addition or removal of other ligands (or electrons) to or
from the coordination sphere. Where cyclic polyenyl
ligands are concerned, this process is casually termed
“ring slippage” and has been reviewed for the cyclopen-
tadienyl (Cp) and indenyl (Ind) ligands.1 In the latter
case, ring slippage is considered as particularly impor-
tant, since it is associated with distinct and sometimes
rather large rate accelerations in substitution reactions
(2) Rerek, M. E.; J i, L.-N.; Basolo, F. J . Chem. Soc., Chem. Commun.
1983, 1208.
(3) (a) Marder, T. B.; Roe, D. C.; Milstein, D. Organometallics 1988,
7, 1451. (b) Borrini, A.; Diversi, P.; Ingrosso, G.; Lucherini, A.; Serra,
G., J . Mol. Catal. 1985, 30, 181. (c) Bo¨nnemann, H. Angew. Chem.,
Int. Ed. Engl. 1985, 24, 248. (d) Foo, T.; Bergman, R. G. Organome-
tallics 1992, 11, 1801. (e) Schmid, M. A.; Alt, H. G.; Milius, W. J .
Organomet. Chem. 1996, 514, 45. (f) Llinas, G. H.; Day, R. O.; Rausch,
M. D.; Chien, J . C. W. Organometallics 1993, 12, 1283.
† Instituto de Tecnologia Qu´ımica e Biolo´gica.
‡ Universidade de Lisboa.
§ Instituto Polite´cnico de Setu´bal.
| Universidade de Aveiro.
Technische Universita¨t Mu¨nchen.
Instituto Superior Te´cnico.
#
(1) O’Connor, J . M.; Casey, C. P. Chem. Rev. 1987, 87, 307.
(4) J i, L.-N.; Rerek, M. E.; Basolo, F. Organometallics 1984, 3, 740.
S0276-7333(97)01005-4 CCC: $15.00 © 1998 American Chemical Society
Publication on Web 05/13/1998

2598 Organometallics, Vol. 17, No. 12, 1998
Calhorda et al.
of the C4 and C9 atoms in Ind2Ni11b was later confirmed
by the existence of a very sensitive correlation between
the indenyl distortion and the value of those chemical
shifts.11a
Ch a r t 1
η1 coordination has also been well characterized and
is associated with fluxional behavior between the C1
and C3 bonding positions.1,12 Coordination of the in-
denyl anion via the six-membered ring has been ob-
served in [(η5-Ind)(η6-Ind)Re]K13 as well as the simul-
taneous coordination of both rings to different transition
metals, as in Cr(CO)3(µ-η:η-indenyl)RhL2 species.14
The situation termed as η3 + η2 coordination in Chart
1 is intermediate between η5 and η3 coordination and
seems to be the most common one in indenyl chemistry.
Although almost undistorted η5 coordination has been
observed in the 18e metallocenes such as Ind2M (M )
Fe, Ru) and [Ind2Co]+,6a most 18-electron compounds
with formal η5 coordination show this kind of slip-fold
distortion as first noted and summarized by Faller.9
More recent examples comprise the systematic study
that revealed this type of distortion in all members of a
variety of d8-[IndRhL2] complexes10,11 as well as in the
formally 16e ethylene-selective polymerization catalyst
Cp*(Ind)ZrCl2.3e A comparative analysis of the struc-
tures of many IndMLnL′m complexes led to a rule of
thumb which states that the benzene ring of the indenyl
points away from the metal in the direction opposite to
the most strongly trans director ancillary L (or L′)
ligand.9 This quite general rule seems to imply that
electronic factors are the ones responsible for the slip-
fold distortion. In the comparative study carried out
with metallocenes,6a it could be clearly shown that
addition of electrons to the metal results in ring slippage
and indenyl distortion to avoid 19 and 20 formal electron
counts as in Ind2Co and Ind2Ni, respectively. Recently,
we have noted that the 2e oxidation of (η3-Ind)CpM(CO)2
leads to [(η5-Ind)CpMo(CO)2]2+, and preliminary results
show that this is a chemically and electrochemically
reversible process.7e,f,15 On the other hand, less infor-
mation is available on the influence of one incoming
ligand (2e donor) upon the ring slippage process. Fur-
thermore, not much is known about the influence of
steric factors on the ring-slippage and slip-fold pro-
cesses. A systematic variation of the degree of substitu-
tion of the indenyl ring in (η-C6H7-xRx)Rh(CO)2 com-
plexes led to essentially unchanged structural char-
acteristics of the indenyl distortion in the ground state.16
However, steric factors do play a role as the kinetic data
have already established, for associative substitution
reactions of several IndM(CO)n (M ) Mn, n ) 3;4 M)
Rh, n ) 216) complexes were found to be dependent on
the basicity, the size, and the chemical nature of the
entering nucleophile.
proposed by Hart-Davis and Mawby: attack of an
incoming nucleophile in an associative pathway is made
easier by a low-energy η5 f η3 ring slippage of the
indenyl ligand which avoids a transition state with more
than 18 electrons at the metal.5 However, the kinetic
data are also consistent with a preequilibrium between
η5 and η3 forms of the starting complex, where the latter
species reacts with the entering nucleophile in a rate-
determining step.4 In any case, the low energy barrier
for the slippage is provided by the resonance stabiliza-
tion of the benzene ring which is regained in the η3
coordination mode.
In keeping with these easy ring slippage rearrange-
ments crystallographic studies have revealed a variety
of coordination modes of the indenyl ligand. Chart 1
summarizes the most important situations when only
the five-membered ring coordination is concerned.
Pure η5 coordination, i.e, with all five carbons of the
ring equidistant from the metal and a planar indenyl
system, is rare but has been observed, inter alia, for
Ind2Fe and closely related metallocene analogues6a and
has been studied from a molecular orbital point of
view.6b Pure η3 coordination is also very well character-
ized in both transition-metal7 and main-group indenyl
complexes.8 In this situation the C1, C2, and C3 atoms
are equidistant and bound to the metal. The ring
junction atoms C4 and C9 lie outside the bond distance
range, and the indenyl ring loses planarity. A number
of geometrical parameters and NMR chemical shifts
have been used to describe these distortions. In the
present paper we use the set of parameters proposed
by Faller, Crabtree, and Habib,9 which are closely
related to the ones used in the works of Marder and
Taylor10 and Baker,11a as discussed in ref 6. The initial
observation of a downfield shift of the 13C chemical shifts
(5) Hart-Davis, A. J .; Mawby, R. J . J . Chem. Soc. A 1969, 2403.
(6) (a) Westcott, S. A.; Kakkar, A. K.; Stringer, G.; Taylor, N. J .;
Marder, T. B. J . Organomet. Chem. 1990, 394, 777. (b) Crossley, N.
S.; Green, J . C.; Nagy, A.; Stringer, G. J . Chem. Soc., Dalton Trans.
1989, 2139.
(7) (a) Merola, J . S.; Kacmarcik, R. T.; Van Engen, D. J . Am. Chem.
Soc. 1986, 108, 329. (b) Forschner, T. C.; Cutler, A. R.; Kullnig, R. K.
Organometallics 1987, 6, 889. (c) Kowalewski, R. M.; Rheingold, A.
L.; Trogler, W. C.; Basolo, F. J . Am. Chem. Soc. 1986, 108, 2460. (d)
Poli, R.; Mattamana, S. P.; Falvello, L. R. Gazz. Chim. Ital. 1992, 122,
315. (e) Ascenso, J . R.; Azevedo, C. G.; Gonc¸alves, I. S.; Herdtweck,
E.; Moreno, D. S.; Roma˜o, C. C.; Zu¨hlke, J . Organometallics 1994, 13,
429. (f) Ascenso, J . R.; Azevedo, C. G.; Gonc¸alves, I. S.; Herdtweck, E.;
Moreno, D. S.; Pessanha, M.; Roma˜o, C. C. Organometallics 1995, 14,
3901. (g) Gonc¸alves, I. S.; Roma˜o, C. C. J . Organomet. Chem. 1995,
486, 155.
(11) (a) Baker, R. T.; Tulip, T. H. Organometallics 1986, 5, 839. (b)
Ko¨hler, F. H. Chem. Ber. 1974, 107, 570.
(12) Herrmann, W. A.; Ku¨hn, F. E.; Roma˜o, C. C. J . Organomet.
Chem. 1995, 489, C56 and references therein.
(8) Viebrock, H.; Abeln, D.; Weiss, E. Z. Naturforsch. 1994, 49B,
89.
(9) Faller, J . W.; Crabtree, R. H.; Habib, A. Organometallics 1985,
4, 929.
(10) (a) Marder, T. B.; Calabrese, J . C.; Roe, D. C.; Tulip, T. H.
Organometallics 1987, 6, 2012. (b) Kakkar, A. K.; J ones, S. F.; Taylor,
N. J .; Collins, S.; Marder, T. B. J . Chem. Soc., Chem. Commun. 1989,
1454. (c) Kakkar, A. K.; Taylor, N. J .; Calabrese, J . C.; Nugent, W. A.;
Roe, D. C.; Connaway, E. A.; Marder, T. B. J . Chem. Soc., Chem.
Commun. 1989, 990. (d) Carl, R. T.; Hughes, R. P.; Rheingold, A. L.;
Marder, T. B.; Taylor, N. J . Organometallics 1988, 7, 1613.
(13) Green, M. L. H.; Lowe, N. D.; O’Hare, D. J . Organomet. Chem.
1988, 355, 315.
(14) (a) Ceccon, A.; Gambaro, A.; Santi, S.; Valle, G.; Venzo, A. J .
Chem. Soc., Chem. Commun. 1989, 51. (b) Ceccon, A.; Elsevier: C. J .;
Ernsting, J . M.; Gambaro, A.; Santo, S.; Venzo, A. Inorg. Chim. Acta
1993, 204, 15.
(15) Calhorda, M. J .; Gamelas, C. A.; Gonc¸alves, I. S.; Herdtweck,
E.; Lopes, J . P.; Roma˜o, C. C.; Veiros, L. Paper presented at the XIIth
FECHEM, Prague, Czech Republic, 1997.
(16) Kakkar, A. K.; Taylor, N. J .; Marder, T. B.; Shen, J . K.;
Hallinan, N.; Basolo, F. Inorg. Chim. Acta 1992, 198-200, 219.

Indenyl Ring Slippage in Molybdenum Complexes
Organometallics, Vol. 17, No. 12, 1998 2599
Ch a r t 2. Ca tion ic Com p lexes of Typ e
In the case of IndMn(CO)3, PR3 ligands, including the
very bulky PCy3, replace CO but 4-tert-butylpyridine is
unreactive under the same conditions. Similarly, the
addition of P(nBu)3 to (η5-Ind)Re(CO)3, to give (η1-Ind)-
Re(CO)3{(P(nBu)3}2 is much faster than the correspond-
ing addition of 2,2′-bipyridyl.17 Regardless of all efforts,
no intermediate (η3-Ind)Re(CO)3(P(nBu)3) species could
be detected. As a matter of fact, even despite the low
activation barrier to associative ligand substitution in
IndMLn complexes, provided by the ring slippage pro-
cess via some kind of transition state or intermediate
bearing a η3-indenyl ligand and the observed thermo-
dynamic and kinetic stability of many such complexes,
it is surprising that so few examples have been identi-
fied where ligand addition to a (η5-Ind)MLn species
results in isolable or spectroscopically identifiable [(η3-
Ind)MLnL′] derivatives. The first characterized ex-
amples all involve very electron-rich species, namely,
(η3-Ind)Ir(PMe2Ph)3 (from (η5-Ind)Ir(C2H4)2),7a [(η3-Ind)-
Fe(CO)3]- (from (η5-Ind)Fe(CO)2]-,7b and (η5-Ind)(η3-
Ind)V(CO)2 (from Ind2V).7c More recently we described
a number of octahedral complexes of general formula
[(η3-Ind)ML2L′(CO)2]BF4 which are easily formed by
addition of L′ to the parent cations [(η5-Ind)ML2(CO)2]-
BF4 (M ) Mo, W) (eq 1).18 Interestingly enough, addi-
[(η5-In d )MoL2(CO)2]BF 4 P r ep a r ed a s in Eq 2
Ch a r t 3. Neu tr a l Com p lexes of Typ e
[(η5-In d )Mo(L-L)(CO)2] P r ep a r ed a s in Eq 2
[(η5-Ind)ML2(CO)2]+ + L′ f [(η3-Ind)MoL2L′(CO)2]+
(i) HBF4/CH2Cl2
(1)
[(η5-Ind)Mo(η3-C3H5)(CO)2]
8
(ii) 2L or L2
tion takes place readily with L ) L′ ) NCMe but not
with L ) L′ ) PMe3, thereby showing an unexpected
dependence of the reactivity on the nature of L or L′.
Also somewhat unexpected is the fact that the cationic
complex [(η3-Ind)W(NCMe)3(CO)2]BF4 bears a rather
bent indenyl ligand with slip-fold angles of 24.1 and
27.4° for two independent molecules,18 a value close to
the one found for the very electron rich (η3-Ind)Ir(PMe2-
Ph)3 (28°).7a
In the present paper we extend these studies to other
18e neutral and cationic four-legged piano stool com-
plexes, [(η5-Ind)ML2(CO)2]0,+, and complete them with
EHMO and DFT calculations aimed at understanding
the electronic and stereochemical features that govern
the η5 f η3 haptotropic shift in this system as a function
of the nature of the ancillary ligands involved in the
process. In an attempt to separate electronic from steric
effects, the calculations are mainly performed on the
isomeric complexes [(η5-Ind)Mo(NCMe)2(CO)2]+ and [(η5-
Ind)Mo(CNMe)2(CO)2]+, which show strikingly different
behaviors in addition reactions.
[(η5-Ind)MoL2(CO)2]BF4 + C3H6 (2)
With the exception of 4, which could not be isolated
as a solid, and the isonitrile derivative 2, which was only
obtained in a rather poor yield from a thick red oil, the
cations 1-6 are easily isolated as analytically pure
crystalline materials. They all present two ν(CO)
stretching bands for the cis-(CO)2 fragment in their IR
spectra. The 1H NMR characterization was done in
CD2Cl2 in order to avoid possible solvent-induced in-
denyl rearrangements. The chemical shifts of the
indenyl protons in the η5 coordination mode assumed
for the 18e compounds 1-6 follow the typical pattern
exemplified in Figure 1a for the well-known case of 1
and are very easy to assign.18
Excess OPPh3 and OP(OMe)3 in the respective prepa-
rations of 3 and 4 did not change the nature of the final
product; i.e., no octahedral product such as [(η3-Ind)-
Mo(OPPh3)3(CO)2]BF4 or [(η3-Ind)Mo{(OP(OMe)3)3}-
(CO)2]BF4 was formed.
In the case of 2, two sets of resonances revealed the
presence of two isomers. The simplest geometries are
represented in 2a -c.
Ch em ica l Stu d ies
Syn th esis a n d Ch a r a cter iza tion of th e [(η5-In d )-
ML2(CO)2]0,+ Com p lexes. The cationic complexes
used in this work are presented in Chart 2, and their
synthesis follows the general method summarized in
eq 2.7e-g,19
Under the same reaction conditions, use of negatively
charged bidentate ligands (L2-), instead of neutral
monodentate L or bidentate L2, produces the neutral
complexes presented in Chart 3.
IR data showed that, in the solid state (KBr pellet),
only two ν(CO) and two ν(CN) vibrations were present,
therefore ruling out the isomer 2c with a trans arrange-
ment of both CO and CNR ligands. When a sample of
solid crystalline 2 was dissolved at -70 °C in CD2Cl2,
1
only one set of resonances was present in the H NMR
(19) (a) Markham, J .; Menard, K.; Cutler, A. Inorg. Chem. 1985,
24, 1581. (b) Almeida, J . M.; Gonc¸alves, I. S.; Roma˜o, C. C. An. Quim.
Int. Ed. 1997, 93, 8. (c) Azevedo, C. G.; Calhorda, M. J .; Carrondo, M.
A. A. F. d. C. T.; Dias, A. R.; Duarte, M. T.; Galva˜o, A. M.; Gamelas,
C. A.; Gonc¸alves, I. S.; Piedade, F. M.; Roma˜o, C. C. J . Organomet.
Chem. 1997, 544, 257.
(17) Casey, C. P.; O’Connor, J . M. Organometallics 1985, 4, 384.
(18) Ascenso, J . R.; Gonc¸alves, I. S.; Herdtweck, E.; Roma˜o, C. C. J .
Organomet. Chem. 1996, 508, 169.

2600 Organometallics, Vol. 17, No. 12, 1998
Calhorda et al.
F igu r e 2. ORTEP plot of the [(η5-Ind)Mo(CNMe)2(CO)2]+
cation (2a ) with the atomic numbering scheme. Thermal
-
ellipsoids are drawn at the 50% probability level. The BF4
anion is disordered in two positions.
Ta ble 1. Selected In ter a tom ic Dista n ces (p m ),
An gles (d eg), a n d Slip P a r a m eter s for
[(η5-In d )Mo(CNMe)2(CO)2]BF 4 (2a )
Mo-C(11)
Mo-C(12)
Mo-C(13)
230.2(9)
228.2(8)
230.2(8)
Mo-C(13A)
Mo-C(17A)
239.3(8)
240.0(7)
C(11)-C(12)
C(11)-C(17A)
C(12)-C(13)
C(13)-C(13A)
C(13A)-C(14)
141.8(14)
141.7(14)
141.0(15)
144.3(12)
143.3(15)
C(13A)-C(17A)
C(14)-C(15)
C(15)-C(16)
C(16)-C(17)
C(17)-C(17A)
142.1(15)
136.3(16)
141(3)
132(2)
141.8(13)
F igu r e 1. Typical resonances for coordinated indenyl in
the complexes [(η5-Ind)MoL2(CO)2]0,+ (a) and [(η3-Ind)MoL3-
(CO)2]0,+ (b).
Mo-C(1)
Mo-C(2)
199.7(11)
197.8(8)
Mo-C(3)
Mo-C(4)
212.9(8)
211.9(8)
C(1)-Mo-C(2)
C(1)-Mo-C(4)
75.3(4)
76.4(4)
C(2)-Mo-C(3)
C(3)-Mo-C(4)
78.7(3)
76.9(3)
∆ ) |S|, pm
σ, deg
Ψ, deg
13.1
0.4
3.8
∆(M-C), pm
Ω, deg
10
3.9
the allyl-ene distortion is obvious and is shown by the
small but definite values of ∆(M-C) of 10 pm for 2a . A
typical alternation of shorter and longer C-C distances
in the benzene part of the indenyl ligand indicates diene
character (Table 1).
As seen in Figure 2, the solid-state structure displays
the isomer 2a . Theoretical studies (EHMO and DFT
calculations) also predict the higher stability of 2a in
comparison to 2b,c (see below).
spectrum. Warming to room temperature slowly al-
lowed the appearance of the second isomer until a 1:1
equilibrium ratio was attained. No changes were
detected in the resonances of the first isomer. These
data are compatible with the presence of only one isomer
in the solid state, which was shown to be 2a on the basis
of a crystal structure investigation (Figure 2, Table 1).
Bond distances and angles together with the charac-
teristic slip parameters are listed in Table 1. The
compound (Figure 2) is monomeric and belongs to the
general family of four-legged piano-stool compounds
(Ind/Cp)MoL4. The geometry around the metal atom
is a pseudo square pyramid with the centroid (C11-
C17a) in the apical position. All intramolecular dis-
tances and angles are in the expected ranges and show
no particularities. In general, the observed values for
the ring-slippage parameters for 2a are consistent with
those given in the literature9,18 and are typical for the
η5 coordination of the C5 ring of the indenyl ligand.
Detailed analysis of the observed distortions are handi-
capped by the quality of the crystal data 2a . However,
The neutral complexes of Chart 3 were similarly
characterized by elemental analysis, IR (two ν(CO)
stretches), and 1H NMR (typical pattern of Figure 1a
for the indenyl resonances).
The trifluoroacetate (7) and the acetate (8) derivatives
are rather labile, as shown by the appearance of more
than two CO stretching bands in their IR spectra in KBr
1
pellets. The H NMR spectrum of 7 in CD2Cl2 at room
temperature presents the typical resonances expected
for a η5-Ind complex of this family: two multiplets for
H5-8 (δ 7.48-7.45 and 7.40-7.30 ppm), a doublet for
H1/3 (δ 5.94 ppm), and a triplet for H2 (δ 5.45 ppm).
Therefore, a bidentate coordination (κ2-O2CCF3) is as-
sumed. The â-keto(phenyliminato) derivative 15 is
exceedingly moisture sensitive, cleanly producing the
â-diketonato (acac) congener 14 on hydrolysis in wet
solvents (1H NMR evidence).

Indenyl Ring Slippage in Molybdenum Complexes
Organometallics, Vol. 17, No. 12, 1998 2601
Sch em e 1
In a previous publication we reported that the triden-
tate ligand Me3tacn (1,4,7-trimethyl-1,4,7-triazacyclo-
nonane) adds to [(η5-Ind)Mo(NCMe)2(CO)2]BF4 (1) to
give [(η3-Ind)Mo(κ3-Me3tacn)(CO)2]BF4.18 Similarly, ttcn
reacts with 1 to give the ring-slipped complex [(η3-Ind)-
Mo(κ3-ttcn)(CO)2]BF4 (16) in excellent yield. 16 has the
expected two ν(CO) vibrations in the IR spectrum, and
the resonance pattern for the indenyl is consistent with
η3-Ind coordination, namely the resonance of H2 at δ
7.26 ppm (see Figure 1b and the discussion below).
Ad d ition Rea ction s of th e [(η5-In d )ML2(CO)2]0,+
Com p lexes. As mentioned in the Introduction, dis-
solution of 1 in NCMe readily forms the adduct 17 as
shown in eq 3.18
[(η5-Ind)Mo(NCMe)2(CO)2]BF4 yNCMe
z
CH2Cl2
1
[(η3-Ind)Mo(NCMe)3(CO)2]BF4 (3)
17
This process is conveniently followed by in situ 1H
NMR in NCCD3 and was studied for the cationic and
neutral complexes 1-15 in order to ascertain the
generality and possible dependence of the ring-slippage
reaction on the nature of the ligand L. The complexes
1-15 were dissolved in NCCD3 at low temperature (ca.
-30 °C) and their spectra immediately recorded. The
spectral changes were then followed with time, upon
temperature increase, or both. In this family of com-
pounds, the resonance pattern of the indenyl ligand in
both η5 and η3 coordination modes is rather distinctive,
as shown in Figure 1 for the typical case of 1 (Figure
1a) and 17 (Figure 1b),18,19 and allows the facile
monitoring of the formation of the adducts of type [(η3-
Ind)MoL2(NCCD3)(CO)2]0,+ without the need to isolate
the compounds. Particularly informative is the highly
deshielded H2 resonance at ca. δ 7.2 ppm, which was
taken as the spectral signature of the η3-Ind coordina-
tion in this family of compounds. In the η5-Ind coordi-
nation this triplet appears at ca. δ 5.5 ppm. The
spectral data for each compound are given in the
Experimental Section.
Sch em e 2
ing conditions, reaction of 2 with a 10-fold excess of
CNMe in CH2Cl2 at room temperature led to the loss of
CO (IR evidence). The essential transformations re-
ported so far are included in Schemes 1 and 2, which
also summarize the results reported in this paper
together with previously reported ones.18,19
Upon dissolution in NCCD3, the cation [IndMo-
{(PhSCH2)2}(CO)2]BF4 (4) readily loses the labile dithio-
ether ligand and forms 17.
The 1H NMR spectrum of [(η5-Ind)Mo(OPPh3)2(CO)2]-
BF4 (3) in NCCD3 at -30 °C is compatible with this
structural formulation. The doublet at δ 6.58 ppm and
the triplet at δ 5.51 ppm are typical of the η5 coordina-
tion mode of the indenyl. After 2 h at this temperature
only the adduct [(η3-Ind)Mo(OPPh3)2(NCCD3)(CO)2]BF4
(18) is present in solution, as indicated by the triplet at
δ 7.23 ppm, assigned to H2 of the η3-Ind ligand of 18.
1
In previous work we have shown that the phosphine
In contrast, the H NMR spectrum of 4, in NCCD3 at
cations [(η5-Ind)Mo(PR3)2(CO)2]+ (PR3 ) /3 triphos, /2
dppe, PPh3, PMe3) add neither NCMe nor a third
equivalent of PR3 to give ring-slipped adducts. How-
ever, addition of PR3 to the cation [(η5-Ind)Mo(NCMe)2-
(CO)2]+ proceeds with overall substitution to give [(η5-
Ind)Mo(PR3)2(CO)2]+. This reaction was now monitored
by 1H NMR for the case of PMe3, to try to detect possible
intermediates, namely [(η3-Ind)Mo(NCMe)2(PMe3)(CO)2]+,
as depicted in Scheme 2.
1
1
-30 °C after 3 h still shows the presence of a mixture
of the initial η5 species 4 and the ring-slipped adduct
[(η3-Ind)Mo{OP(OMe)3}2(NCCD3)(CO)2]BF4 (19). The
spectrum of a freshly prepared solution of 2 at -30 °C
in NCCD3 is consistent with the exclusive presence of
the isomer 2a . After ca. 2 h two isomers (2a and 2b)
are present in a 1:1 ratio, but no sign of addition of
NCMe to 2, with concomitant ring slippage, was ob-
served. This strongly contrasts with the reactivity of
the isomeric nitrile complex described in eq 3. Treat-
ment of 2 with excess CNMe (200 µL, 4.2 mmol), in
(CD3)2CO, also fails to give the putative ring-slipped
adduct [(η3-Ind)Mo(CNMe)3(CO)2]BF4. Under more forc-
Addition of 1 equiv of PMe3 to a solution of [(η5-
Ind)Mo(NCMe)2(CO)2]+ in CD2Cl2 at -70 °C leads to the
immediate formation of the substitution product [(η5-
Ind)Mo(NCMe)(PMe3)(CO)2]+ (20). All the starting

2602 Organometallics, Vol. 17, No. 12, 1998
Calhorda et al.
materials disappear within ca. 30 min, and no signals
are observed in the regions expected for the H2 reso-
nance of a ring-slipped indenyl ligand. Addition of a
second equivalent of PMe3 at -70 °C did not produce
any reaction until the temperature was raised to room
temperature. Again, no η3-Ind species was observed
while the resonances of [(η5-Ind)Mo(PMe3)2(CO)2]+ started
to replace those of 20. This second substitution is much
slower than the first and only became complete after
some hours at room temperature. Further addition of
PMe3 (6 equiv) at room temperature did not produce any
change in the Mo complex [(η5-Ind)Mo(PMe3)2(CO)2]+,
as previously reported.18
However, this addition is easily reversed and we were
not able to isolate 22.
The enaminone derivative IndMo{κ2-MeCOCHC(NPh)-
Me}(CO)2 (15) also adds NCMe, but less extensively
than its acac congener 14. In fact, the 1H NMR
spectrum of a sample of 15 in NCCD3 shows the
presence of the starting material together with the ring-
slipped adduct [(η3-Ind)Mo(NCMe){κ2-MeCOCHC(NPh)-
Me}(CO)2] (23) after 1 h at -30 °C, and this equilibrium
is no longer disturbed at this temperature. Exploratory
studies with related enaminones were inconclusive due
to their extremely easy hydrolysis to give 14.
1
The well-resolved H NMR spectrum of the carboxy-
lato complex 7 in CD2Cl2 is not reproduced in NCCD3,
at room temperature, since only broad signals can be
seen, as in the case of the spectrum of [(η5-Ind)Mo-
(NCMe)2(CO)2]+ in the same solvent. At -45 °C the
resolution increases and the resonances are assigned
in the following way: a broad triplet (δ 7.28 ppm) for
H2, a multiplet (δ 6.46 ppm) for the benzenoid protons
H5-8, and two relatively broad peaks (δ 5.19 and 4.76
ppm) for the H1 and H3 protons. All these chemical
shifts and multiplicities are therefore indicative of the
presence of a η3-indenyl asymmetric complex, for which
we propose the structure [(η3-Ind)Mo(NCMe)(κ2-O2-
CCF3)(CO)2] (21), although NMR cannot clearly rule out
a formulation such as [(η3-Ind)Mo(NCMe)2(κ1-O2CCF3)-
(CO)2] due to the impossibility of integrating signals of
coordinated NCCD3.
Th eor etica l Stu d ies
Several theoretical studies have addressed the prob-
lem of understanding the electronic structure of indenyl
complexes.6b,10c,20 Faller and co-workers9 analyzed the
structural features of the η5-indenyl complexes available
at the time and concluded that the benzene ring of the
indenyl is oriented trans to the ligand exhibiting the
greater trans influence.
Our purpose is to understand the structural prefer-
ences of η5- and η3-indenyl complexes of the types [(η5-
Ind)MoL2(CO)2]+ and [(η3-Ind)MoL2L′(CO)2]+ and the
reactivity patterns of the η5-indenyl complexes described
above. Haptotropic rearrangements of cyclopentadienyl
groups and other polyenes started to be theoretically
studied long ago, but a considerable amount of new
experimental results is now available.21 Extended
Hu¨ckel22 and density functional theory (DFT)23 calcula-
tions will be used.
Bon d in g in In d en yl Com p lexes. The indenyl
ligand differs from the simpler cyclopentadienyl ligand
by the larger number of π orbitals. Those having lower
energies are shown in Chart 4, and the notation is that
used in ref 20a.
The lower energy orbital, 1πs, resembles the corre-
sponding one for the Cp and is unique. However, for
each orbital belonging to the Cp 1e set, there are two
for the indenyl, namely 2πs and 3πs, and 1πa and 2πa.
This duplication derives from the presence of the
adjacent benzene ring, and it can be noted that in 2πs
there is no contribution from the two carbons in the
Our proposal is supported by the observation that
addition of diethyl ether to a solution of 7 in NCMe
produces a red-brown crystalline complex which has the
1
same H NMR spectrum, in NCCD3, and the elemental
analysis and IR spectrum corresponding to the adduct
[(η3-Ind)Mo(NCMe)(κ2-O2CCF3)(CO)2] (21).
After these observations it was quite unexpected to
verify that no similar type of addition takes place for
the very similar complexes with other carboxylato (8,
9), amidinato (10, 11), dithiocarbamato (12), and dithio-
phosphato (13) ligands, in NCCD3. The 1H NMR
spectra of these complexes in NCCD3 between -40 °C
and room temperature indicate the typical η5-indenyl
coordination with the H2 resonance in the range be-
tween δ 5.09 ppm (8) and δ 5.29 ppm (12).
(20) (a) Bonifaci, C.; Ceccon, A.; Santi, S.; Mealli, C.; Zoellner, R.
W. Inorg. Chim. Acta 1995, 240, 541. (b) O’Hare, D.; Green, J . C.;
Marder, T.; Collins, S.; Stringer, G.; Kakkar, A. K.; Kaltsoyannis, N.;
Kuhn, A.; Lewis, R.; Mehnert, C.; Scott, P.; Kurmoo, M.; Pugh, S.
Organometallics 1992, 11, 48.
(21) (a) Anh, N. T.; Elian, M.; Hoffmann, R. J . Am. Chem. Soc. 1978,
100, 110. (b) Albright, T. A.; Hofmann, P.; Hoffmann, R.; Lillya, C. P.;
Dobosh, P. J . Am. Chem. Soc. 1983, 105, 3396.
(22) (a) Hoffmann, R. J . Chem. Phys. 1963, 39, 1397. (b) Hoffmann,
R.; Lipscomb, W. N. J . Chem. Phys. 1962, 36, 2179. (c) Ammeter, J .
H.; Bu¨rgi, H.-J .; Thibeault, J . C.; Hoffmann, R. J . Am. Chem. Soc. 1978,
100, 3686.
(23) Parr, R. G.; Yang, W. Density Functional Theory of Atoms and
Molecules; Oxford University Press: New York, 1989.
As an exception, the acac derivative (η5-Ind)Mo(κ2-
acac)(CO)2 (14) readily adds NCMe when dissolved in
NCCD3 to give the symmetrical adduct [(η3-Ind)Mo(κ3-
MeCOCHCOMe)(NCMe)(CO)2] (22), as in eq 4.
The 1H NMR spectrum of 22 can only be well-resolved
at -45 °C. At this temperature the H2 triplet appears
at δ 7.11 ppm, which represents a downfield shift of ca.
2.10 ppm relative to its chemical shift in the η5-indenyl
coordination of 14 in CD2Cl2. An upfield shift of ca. 1.00
ppm is simultaneously recorded for the H5-8 protons.

Indenyl Ring Slippage in Molybdenum Complexes
Organometallics, Vol. 17, No. 12, 1998 2603
Ch a r t 4
Ch a r t 6
Ch a r t 7
Ch a r t 5
junction, C4 and C9, while in 2πa this contribution is
very small. The result is that upon bonding to a metal
center these orbitals of similar nodal properties will mix.
In a (η5-Ind)M complex, the ligand formally donates
three pairs of electrons to the metal and they come from
1πs, 2πs plus 3πs (they mix), and 2πa (mixing from 1πa
is negligible here), interacting with z2, xz, and yz,
respectively (Chart 5, left side). Back-donation is not
relevant.
When the indenyl slips and becomes η3-coordinated
due to ligand addition, the metal gains two electrons
and one of the previous two-electron stabilizing interac-
tions becomes a four-electron destabilizing interaction.
As both the bonding and the antibonding molecular
orbitals are populated, one bonding component is lost
and only four electrons are formally donated to the
metal (see right side of Chart 5). The two lowest,
occupied energy d levels for a d4 metal are not shown
in this chart. Notice that the change in the orbitals of
the indenyl upon bending does not affect this qualitative
picture.
compounds (L ) NCH, CNH), showing that the benzene
ring preferred, in both cases, to lie trans to the two CO
ligands, as seen above in the crystal structure of 2a and
following the rule of thumb of Faller.9 The energies of
these compounds are shown in Chart 6 (eV; black circles
indicate N atoms in the nitrile and isonitrile ligands).
Extended Hu¨ckel calculations on [(η5-Ind)ML2(CO)2]+
(L ) NCMe, CNMe) led to the same structural prefer-
ences.
DFT calculations were done for the related η3-indenyl
derivatives, namely [(η3-Ind)Mo (NCH)3(CO)2]+ (17; see
eq 3)18 and the hypothetical isonitrile complex [(η3-Ind)-
Mo (CNH)2(NCH)(CO)2]+, which might be formed upon
addition of NCR to [(η5-Ind)M(CNR)2(CO)2]+. The re-
sults are depicted in Chart 7 (energy in eV), and as
observed in the crystal structure of 17, the benzene lies
cis to the two CO groups; that is, the conformation is
not the same as in the η5-indenyl complexes.
Str u ctu r a l P r efer en ces in In d en yl Com p lexes.
(24) Amsterdam Density Functional (ADF) Program, release 2.01;
Vrije Universiteit: Amsterdam, The Netherlands, 1995.
Dft calculations23,24 were done on [(η5-Ind)ML2(CO)2]+

2604 Organometallics, Vol. 17, No. 12, 1998
Calhorda et al.
F igu r e 3. Interaction diagram between a bent η3-indenyl ligand and the Mo(NCMe)3(CO)2 fragment for two possible
conformations: benzene ring trans (left) or cis (right) to the CO ligands in the model complex [(η3-Ind)Mo(NCMe)3(CO)2]+.
The lowest energy d level is not shown.
Ch a r t 8
Extended Hu¨ckel calculations were performed on the
[(η3-Ind)MoL2(NCMe)(CO)2]+ (L ) NCMe, CNMe, PH3)
model, taking 30° as the folding angle. An optimization
of this angle led to a value of 31.5°, not too far from the
experimental ones of 24.1 and 27.4° found in the two
independent molecules of the W analogue [(η3-Ind)W-
(NCMe)3(CO)2]+.18 The most stable conformation for L
) NCMe, CNMe was the same as in the DFT calcula-
tions, but for L ) PH3, the preferred conformation was
the opposite (Chart 8).
The reason one conformation is preferred can be more
easily understood from the EH results. The energy
differences between the conformers of the η3-indenyl
derivative are more pronounced than for the η5-indenyl
complex, and it is easier to trace their origin. Our
interpretation will be based on the interaction diagram
between the η3-indenyl ligand and the metal fragment
Mo(NCMe)3(CO)2 when they combine in the two limiting
orientations, as shown in Figure 3.

Indenyl Ring Slippage in Molybdenum Complexes
Organometallics, Vol. 17, No. 12, 1998 2605
F igu r e 4. HOMO of [(η3-Ind)Mo(NCMe)3(CO)2]+ in two
conformations: more stable (right) and less stable (left).
The energy is lower for the conformation shown on
the right side, and it reflects the lower energy of the
HOMO, which is a metal-indenyl antibonding orbital
(remember Chart 5 above) but is also metal-carbonyl
bonding. It arises from interaction between yz and the
3πs orbital. In this molecule, 2πs will mix mainly with
1πs and interact with the z2 type orbital. The two
HOMOs are shown in Figure 4 in a three-dimensional
CACAO picture.20b The HOMO for the less favored
geometry (left side) is much less Mo-CO bonding.
As pictured in the diagram, the fragment orbitals are
exactly the same on the left and on the right. When
the bond to the indenyl is established, the metal orbitals
must rehybridize in order to achieve a good overlap with
the indenyl, and they do it in different ways in the two
cases. Notice that, on the right side of the diagram, the
most stable isomer, the HOMO comes mostly from 1a′
(∼70%), which is Mo-CO bonding. On the other hand,
on the left, a second metal fragment orbital, 2a′, Mo-
CO antibonding, mixes in and weakens the Mo-CO
bond. The HOMO derives from 1a′ (47%) and from 2a′
(18%). The loss of Mo-CO bonding character accounts
for the higher energy and is a consequence of binding
the indenyl ligand in this orientation.
A similar picture is found for the hypothetical iso-
meric ring-slipped adduct [(η3-Ind)Mo(CNMe)2NCMe-
(CO)2]+, where the preferred conformation is still the
same, b in Chart 8. As the carbonyl is a stronger
π-acceptor than the linear isonitrile, the conformation
which allows stronger bonding to the strongest π-ac-
ceptor is chosen. In the third type of model complex,
[(η3-Ind)Mo(PR3)2NCMe(CO)2]+, the same conformation
should be chosen on electronic grounds, according to the
above arguments, but the bulkiness of the phosphines
prevents this arrangement, where the planar allylic
group is very close to the phosphines. The rotation of
the indenyl group allows the benzene substituent,
conveniently bent up, to lie above these bulky ligands
(Chart 8, c). The preferred conformation predicted for
the bis(phosphine) derivative is therefore induced by
steric rather than by electronic effects and the bond
between the metal and the indenyl ligand is not
optimized. This may explain why, as far as we know,
there are no examples of [(η3-Ind)Mo(PR3)2(NCMe)-
(CO)2]+ complexes.
F igu r e 5. Walsh diagram for the addition of a NCMe
ligand to [(η5-Ind)Mo(NCMe)2(CO)2]+ to give [(η3-Ind)Mo-
(NCMe)3(CO)2]+.
electronic situation related to the previous one may be
found. The HOMO has a slightly lower energy for the
most stable conformer. It originates from a Mo-CO
bonding orbital and the corresponding orbital in the η5-
indenyl. The metal fragment has one ligand less, so its
frontier orbitals are not exactly the same as those of
MoL2(CO)2, and the four ligands remain farther from
the indenyl than in the pseudo-octahedral geometry.
Liga n d Ad d ition Rea ction s. The reaction of [(η5-
Ind)Mo(NCMe)2(CO)2]+ with one NCMe was studied
initially using extended Hu¨ckel calculations, and the
main changes of the electronic levels are shown in the
Walsh diagram (Figure 5). The significant levels are
highlighted and labeled.
The σ lone pair of the nitrile (left side) is stabilized
as it becomes the Mo-N bonding orbital and therefore
provides the driving force for the reaction. The metal
orbital involved in this new Mo-N bond is z2, which is
also the main contributor to the HOMO of the η5-indenyl
complex (left side). It will give rise to the σ* Mo-N
orbital and is very much destabilized, falling outside the
energy window shown in Figure 5. As the reaction
proceeds, the two electrons which occupied the HOMO
of the η5-indenyl complex (left) will occupy the HOMO
of the η3-indenyl complex (right), which is an orbital
mainly located in the metal and in the indenyl and
antibonding in character (see Figure 3). This orbital
descends from higher energy orbitals on the left and
becomes the HOMO after some avoided crossings (no
symmetry is kept along the reaction). When the C5 ring
of the indenyl bends, going from η5 to η3, the two carbon
atoms in the junction, C4 and C9, no longer bind to the
metal. Thus, the antibonding character of the orbital
is greatly relieved. This stabilization of the π* Mo-
Ind orbital upon bending of the ring also contributes to
lower the energies of the final state, as can be seen by
comparing the energy of the first empty π* Mo-Ind
orbital on each side of the Walsh diagram. The indenyl
ligand is forced to bend, along the reaction pathway, to
Why do the [(η5-Ind)MoL2(CO)2]+ complexes prefer
another arrangement with the benzene trans to the CO
groups? The energy difference between the two isomers
of [(η5-Ind)Mo(NCMe2)(CO)2]+ is only 0.14 eV, at the EH
level, and is therefore not so easy to trace. The
molecular orbital diagram suggests, however, that an

2606 Organometallics, Vol. 17, No. 12, 1998
Calhorda et al.
F igu r e 6. Relative energy when one NCH ligand is moved away from the metal in [(η3-Ind)MoL2(NCH)(CO)2]+ and for
the final η5-Ind complex.
provide a more stable HOMO for the resulting complex
[(η3-Ind)Mo(MeCN)3(CO)2]+.
a calculation owing to the size and absence of symmetry
of the system.
Finally, the opposite effect, namely the destabilization
of the π Mo-Ind levels when the C5 ring bends, can be
seen on the level signaled in the Walsh diagram. Here,
when two carbon atoms move away from the metal, part
of the bonding character is lost and the energy increases.
The η3-indenyl ligand, in the preferred conformation of
the final complex, is still strongly bound and therefore
the final product is stable. The reaction takes place.
In the related phosphine derivative [(η5-Ind)Mo(PH3)2-
(CO)2]+, the preferred conformation of the final [(η3-Ind)-
Mo(PH3)2(NCH)(CO)2]+ derivative should be less favored
on electronic grounds. In the presence of bulky ligands,
the η3-indenyl group would be forced to adopt a confor-
mation that prevents a strong bond to the metal (see
Chart 8). Therefore, instead of ligand addition ac-
companied by indenyl slipping, other reactions (or no
reaction) will take place. This suggests that the steric
effect caused by the bulk of the PR3 ligands in the
starting reagent is responsible for the experimentally
observed absence of reaction.
With the final product, [(η3-Ind)MoL2(NCH)(CO)2]+
with L ) NCH or CNH, as starting material, the nitrile
was allowed to move away from the metal, as sketched
in Figure 6 for L ) NCH. A mirror plane was kept along
this movement, thereby preventing the indenyl ligand
from rotating. It can bend, slip, and change its orienta-
tion relative to other ligands. Therefore, the η3 deriva-
tive exhibits the preferred conformation, but the con-
formation of the final η5 derivative is not the most stable
one.
The energy increases as the ligand moves away from
the metal and the molecule rearranges. Eventually the
final geometry is reached. The most obvious geometry
adjustment taking place is the closing up of the Nax
-
Mo-Neq angle, which goes from 85.74° (2.17 Å) to 84.39°
(2.67 Å), 83.43° (3.17 Å), 82.91° (3.67 Å), and 63.49 Å in
the η5 derivative. At the same time there are small
changes in the indenyl ligand, but it remains essentially
η3-coordinated up to a Mo-N distance of 3.67 Å. When
L ) CNH, the reaction profile is very similar, and the
energy increases in a similar way when the Mo-N
distance increases. This suggests that the approach of
the incoming ligand induces the distortion of the in-
denyl.
The situation becomes more difficult when [(η5-Ind)-
Mo(CNMe)2(CO)2]+ is considered, because no reaction
of nitrile with this complex is observed and steric factors
cannot be responsible. There is no difference between
CNMe and NCMe from this point of view, and no
significant differences could be traced easily at the EH
level. This led us to study the starting and final
reagents, as well as part of the reaction path, using DFT
calculations which provide more reliable energies. We
did not attempt to locate a real transition state and find
the reaction pathway, as this would be too demanding
It is also possible to calculate the enthalpy change
for reaction 5 from the results of the DFT calculations.
[(η5-Ind)MoL2(CO)2]+ + NCH f
[(η3-Ind)MoL2(NCH)(CO)2]+ (5)
Both the starting and the final products are considered

Indenyl Ring Slippage in Molybdenum Complexes
Organometallics, Vol. 17, No. 12, 1998 2607
particularly electron-rich species. Besides NCMe, simi-
lar reactions are observed with DMF and oxygen ligands
such as OPPh3 and OP(OMe)3, acac, and en but not with
CNMe, PMe3 and other phosphines, bipy, carboxylates,
etc. Although steric considerations may be important
in determining this reactivity behavior, their role
remains unclear. In this respect, the CNMe case is
particularly striking, since we were able to detect
neither [(η3-Ind)M(CNMe)3(CO)2]+ nor [(η3-Ind)M(CN-
Me)2(NCMe)(CO)2]+. In their studies on the addition/
slippage reactions of [(η7-C7H7)M(CNtBu)(CO)2]+ and
[(η5-C7H9)ML2(CO)2]+ (M ) Mo, W) Whiteley and co-
workers reported many similar observations.25 In the
first place, NCMe forms adducts with most complexes
of both types, e.g. [(η3-C7H7)M(NCMe)3(CO)2]+ from [(η7-
C7H7)M(NCMe)(CO)2]+ and [(η3-C7H9)ML2(NCMe)(CO)2]+
from [(η5-C7H9)ML2(CO)2]+. In the latter example, the
reversible adduct formation with NCMe is less favored
by L2 ) diphosphines relative to L2 ) bipy. On the other
hand, addition of CNtBu to [(η7-C7H7)M(CNtBu)(CO)2]+
leads to associative substitution via [(η3-C7H7)M(CNtBu)3-
(CO)2]+ and the similar reaction with [(η5-C7H9)ML2-
(CO)2]+ leads to CO substitution via isolable [(η3-C7H9)-
ML2(CNtBu)(CO)2]+. The coordination of a π-acceptor
to the metal decreases the ability toward ring slippage
because [(η5-C7H9)M(CNMe)(dppe)(CO)]+ is no longer
capable of adding NCMe as [(η5-C7H9)M(dppe)(CO)2]+
does.25
F igu r e 7. Change in enthalpy for the reaction [(η5-Ind)-
MoL2(CO)2]+ + NCH f [(η3-Ind)MoL2(NCH)(CO)2]+.
in their most stable geometry. They are respectively
-0.125 eV for L ) NCH and 0.416 eV for L ) CNH,
and this means that while the first reaction is exother-
mic, the second is endothermic. This result is schemati-
cally shown in Figure 7.
The comparison of the energies of the initial and final
states indicates that the largest contribution to the
global enthalpy change comes from the final η3 deriva-
tive, which is 0.80 eV less stable for the isonitrile
complex. The initial η5 complexes differ only by 0.26
eV, the nitrile complex being again the most stable. This
can be explained by the simultaneous presence of two
carbonyl ligands trans to both nitriles (or isonitriles).
As they are strong π-acceptors, the presence of the
nitrile π-donors will make back-donation from the metal
relatively efficient, while, on the other hand, isonitriles
will compete for the same back-donation. This competi-
tion takes place for both η5 and η3 complexes, but the
electronic balance is more delicate in the η3 species,
where the metal orbitals are also involved in back-
donation to the indenyl ring. This competition is
responsible for the lower stability of the final product
of the reaction when L ) CNH, compared to L ) NCH.
All these facts clearly point to the existence of an
electronic regulation of the addition/slippage reaction.
EHMO calculations correctly predict all the conforma-
tional preferences in the four-legged piano-stool species
[(η5-Ind)MoL2(CO)2]+ and six-coordinated [(η3-Ind)ML2L′-
(CO)2]+ complexes involved, but they are unable to
disclose the electronic control of the reaction and as-
sociated energy changes. A thermodynamic explanation
is, however, provided by DFT calculations on model
compounds, i.e. with NCH and CNH instead of NCMe
and CNMe. In fact, the indenyl slippage reaction is
exothermic for NCH addition and endothermic for CNH
addition to [(η5-Ind)ML2(CO)2]+ (L ) CNH, NCH).
Assuming that the entropy of the addition/slippage
reaction is negative (T∆S < 0), the process is only
thermodynamically possible for reactions that are exo-
thermic enough to give ∆G° < 0. These energetics seem
to be governed mainly by the M-CO bond energy
changes that take place during the addition and are
dependent on the nature of L. Destabilizing the M-CO
bond in the octahedral adduct with competing π-accep-
tors, e.g. CNMe and bipy, disfavors or prevents the
observation of the ring slippage process. Conversely,
addition of π-donors such as NCMe and oxygen ligands
assists the ring slippage adduct formation by electron
donation into the metal orbitals (yz, xz) that enhance
M-CO back-bonding and the π Mo-Ind interaction.
Certainly, minor differences between several L ligands
may lead to different stabilities of the ring-slipped
adducts. The lack of reactivity of the carboxylates [(η5-
Ind)M(O2CR)(CO)2] (R ) Me, Ph) toward NCMe addition
contrasts sharply with the extremely facile addition of
Discu ssion a n d Con clu sion s
The kinetic rate acceleration of substitution reactions
of indenyl complexes vs their cyclopentadienyl conge-
ners has long been attributed to the easy slippage of
the indenyl ring from a η5 to a η3 coordination mode.4,5
In view of this mechanistic proposal and the fact that a
number of η3-indenyl complexes are stable entities, one
might expect to be able to isolate and characterize η3-
indenyl complexes formed upon ligand addition to η5-
indenyl substrates. However, the number of η3-indenyl
complexes prepared in this way is surprisingly small.
The most clear-cut examples result from CO uptake by
the very electron rich complexes [(η5-Ind)Fe(CO)2]- and
(η5-Ind)2V to give respectively [(η3-Ind)Fe(CO)3]- 7b and
(η5-Ind)(η3-Ind)V(CO)2.7c The formation of (η3-Ind)Ir-
(PMe3)3 from (η5-Ind)Ir(COD) and excess PMe3 is also
a related example.7a The search for (η3-Ind)Re(CO)3L
derivatives from (η5-Ind)Re(CO)3 and excess L only led
to (η1-Ind)Re(CO)3L2 (L ) P(nBu)3, 1/2 bipy).17 Therefore,
the easy addition of NCMe to the cation [(η5-Ind)M-
(NCMe)2(CO)2]+ (M ) Mo, W) came as a surprise, for
NCMe is a weak, labile donor and the cation is not a
(25) (a) Beddoes, R. L.; Hinchliffe, J . R.; Souza, A.-L. A. B.;
Whiteley, M. W. J . Chem. Soc., Dalton Trans. 1994, 2303. (b)
Hinchliffe, J . R.; Ricalton, A.; Whiteley, M. W. Polyhedron 1991, 10,
267. (c) Green, M. L. H.; Ng, D. K. P. Chem. Rev. 1995, 95, 439.

2608 Organometallics, Vol. 17, No. 12, 1998
Calhorda et al.
this solvent to the related complexes [(η5-Ind)ML2(CO)2]+
(L ) acac, CF3COO).
the stability of the adducts will increase and they can
be observed with stronger donors and/or π-acceptors.
The combination of our results with those of White-
ley allows for a systematization of the characteristics
of the addition/ring slippage reactions in the family of
[(π-CnHm)MLy(CO)2]z+ complexes (M ) Mo(II), W(II);
π-CnHm ) η5-Cp, η5-Ind, η5-C7H9, η7-C7H7; y ) 1, 2; z )
0, 1). (i) The ease of ring slippage within π-systems
increases in the order η5-Cp < η5-Ind < η5-C7H9 < η7-
C7H7. (ii) NCMe adds reversibly to all these compounds
except η5-Cp. (iii) π-Acceptor ligands L (e.g. CNR)
destabilize the formation of ring-slipped adducts with
NCMe. (iv) As a result of the preceding observations,
reaction with CNR ligands only leads to isolation of ring
slipped adducts in the case of η5-C7H9 and η7-C7H7
derivatives; otherwise, overall substitution is observed.
This general picture seems at first glance counterin-
tuitive, for the stronger donors (PMe3, CNR, bipy, etc.)
seem not to favor addition reactions, whereas the weak,
labile NCMe promotes ring slippage in all cases except
for Cp. However, accumulated kinetic evidence on the
indenyl effect as well as the stepwise substitution
reactions of [(η7-C7H7)Mo(CNtBu)(CO)2]+ and [(η5-
C7H9)Mo(CNtBu)2(CO)2]+ with isolable ring-slipped in-
termediates shows that substitutions are associative
and entrain ring slippage. Therefore, the impossibility
of observing ring-slipped species such as [(η3-Ind)ML2L′-
(CO)2]+ simply means that they are too energetic and
probably very close to the transition state in the
respective substitution reactions toward [(η5-Ind)MLL′-
(CO)2]+. As revealed by the snapshots taken along the
reaction pathway by means of DFT calculations (Figure
6), indenyl slippage is clearly present at very long,
almost nonbonding M-NCH distances (3.67 Å). This is
more readily interpreted in terms of eq 6.
Exp er im en ta l Section
All preparations and manipulations were done with stan-
dard Schlenk techniques under an atmosphere of argon.
Solvents were dried by standard procedures (THF, Et2O,
toluene, and hexane over Na/benzophenone ketyl; CH3COCH3
over anhydrous K2CO3; CH2Cl2, NCMe, and triethylamine over
CaH2), distilled under argon, and kept over 4 Å molecular
sieves (3 Å for NCMe).
Microanalyses were performed at the ITQB. 1H and 13C
spectra were measured on a Bruker CXP 300 and on a Bruker
AMX 300. 1H chemical shifts are reported on the scale relative
to SiMe4 (δ 0.0). Infrared spectra were obtained using a
Unicam Mattson Mod 7000 FTIR spectrometer.
The following reagents were prepared as published: IndMo-
(η3-C3H5)(CO)2,7e,f (PhSCH2)2,26 CpMo(η3-C3H5)(CO)2,7e,f [IndMoI-
(CO)2]2,7e,f N,N′-diphenylformamidine ([HC(NPh)(NHPh)]),27
the silver derivative of N,N′-diphenylformamidine (Ag{HC-
(NPh)2}),28 N,N′-di-p-tolylformamidine ([HC(HN-p-tolyl)(N-p-
tolyl)]),29 4-anilinopent-3-en-2-one (MeCOCHC(NHPh)Me),30,31
triphenylphosphine oxide (OPPh3),32 [IndMo(NCMe)2(CO)2]-
BF4,7e,f and CNMe.33
The silver derivative of N,N′-di-p-tolylformamidine (Ag{HC-
(N-p-tolyl)2}) resulted from a preparation similar to that for
Ag{HC(NPh)2}. The lithium salt of 4-anilinopent-3-en-2-one
(Li{MeCOCHC(NPh)Me}) was prepared by adding a 1.6 M
solution of n-butyllithium in ether (1 equiv) to a cold solution
(-70 °C) of the enaminone in THF.
P r ep a r a tion of [In d Mo(CNMe)2(CO)2]BF 4 (2). A solu-
tion of IndMo(η3-C3H5)(CO)2 (0.20 g, 0.65 mmol) in CH2Cl2 (20
mL) was treated with HBF4‚Et2O (1 equiv). After 10 min,
CNMe was added (61 µL, 1.30 mmol) and the reaction
continued for 1 h at room temperature. After concentration
to ca. 5 mL and addition of ether, a yellow microcrystalline
product precipitated, mixed with a dark oil. This was further
recrystallized from CH2Cl2/Et2O to give the pure compound
in 40% yield. Anal. Found: C, 41.33; H, 3.04; N, 6.66. Calcd
for C15H13BF4N2O2Mo: C, 41.32; H, 3.00; N, 6.42. Selected
IR (KBr, cm-1): 2228, 2214, vs, ν(CN); 1995, 1927, vs, ν(CO).
1H NMR (NCCD3, 300 MHz, -40 °C, δ in ppm): isomer A,
7.67-7.64 (m, 2H, H5-8), 7.35-7.32 (m, 2H, H5-8), 6.22 (d, 2H,
H1/3), 5.47 (t, 1H, H2), 3.44 (s, 6H, Me); isomer B, 7.63-7.59
(m, 2H, H5-8), 7.28-7.25 (m, 2H, H5-8), 6.11 (d, 2H, H1/3), 5.61
(t, 1H, H2), 3.59 (s, 6H, Me).
P r ep a r a tion of [In d Mo(OP P h 3)2(CO)2]BF 4 (3). A solu-
tion of IndMo(η3-C3H5)(CO)2 (0.20 g, 0.65 mmol) in CH2Cl2 (20
mL) was treated with HBF4‚Et2O (1 equiv). After 10 min,
excess OPPh3 was added (0.39 g, 1.40 mmol) and the reaction
continued for 2 h at room temperature. After concentration
to ca. 5 mL and addition of ether, a dark orange microcrys-
talline product precipitated. This was further recrystallized
from CH2Cl2/Et2O to give the pure compound in 90% yield.
Anal. Found: C, 61.49; H, 4.51. Calcd for C47H37BF4O4P2-
Mo: C, 62.00; H, 4.10. Selected IR (KBr, cm-1): 1942, 1852,
vs, ν(CO); 1120, ν(PdO). 1H NMR (NCCD3, 300 MHz, -30
°C, after 1 h, δ in ppm): [(η5-Ind)Mo(OPPh3)2(CO)2]BF4, 7.63-
7.49 (m, 30H, Ph), 7.10-7.05 (m, 2H, H5-8), 6.86 (m (br), 2H,
[(η5-Ind)MoL2(CO)2]+ h [(η3-Ind)MoL2(CO)2]+ + L′
[(η3-Ind)MoL2(CO)2]+ + L′ f [(η3-Ind)MoL′2(CO)2]+
(6)
[(η3-Ind)MoL2L′(CO)2]+ f
[(η5-Ind)MoLL′(CO)2]+ + L
The rapid η5 h η3 “preequilibrium” gives rise to the
electronically unsaturated species [(η3-Ind)Mo(CO)2L2]0,+
,
which is intercepted by weak ligands to eventually form
stable entities, the ring-slipped adducts, which are “en
route” toward the substitution products. Presenting it
as an alternative possibility, Basolo and co-workers
considered this rapid preequilibrium as totally consis-
tent with the kinetic data on the substitution reaction
of (η5-Ind)Mn(CO)3 by phosphines, which can also be
interpreted on the basis of the more familiar addition/
slippage associative pathway.4 Strong donors and/or
π-acceptors destabilize the ring-slipped adducts and give
rapid substitutions (or back-reactions to the starting
materials). Also, the nodal characteristics of the dif-
ferent π-systems may subtly influence the stability of
the η3-indenyl derivatives. A more extended π-system
will shift the “equilibrium” to the right because the
energy of the different hapticities is “smoothed out” and
the reaction becomes less discriminative. As a result,
(26) Hartley, F. R.; Murray, S. G.; Levason, W.; Soutter, H. E.;
McAuliffe, C. A. Inorg. Chim. Acta 1979, 35, 265.
(27) Giacolone, A. Gazz. Chim. Ital. 1932, 62, 577.
(28) Bradley, W.; Wright, I. J . Chem. Soc. 1956, 640.
(29) Backer, H. J .; Wanmaker, W. L. Recl. Trav. Chim. Pays-Bas
1949, 68, 247.
(30) Collman, J . P.; Kittleman, E. T. Inorg. Chem. 1962, 1, 499.
(31) Iida, H.; Yuasa, Y.; Kibayashi, C.; Iitaka, Y. J . Chem. Soc.,
Dalton Trans. 1956, 640.
(32) Gilman, H.; Brown, G. E. J . Am. Chem. Soc. 1945, 67, 824.
(33) Schuster, R. E.; Scott, J . E.; Casanova, J ., J r. Org. Synth. 1966,
46, 76.

Indenyl Ring Slippage in Molybdenum Complexes
Organometallics, Vol. 17, No. 12, 1998 2609
H5-8), 6.58 (d, 2H, H1/3), 5.51 (t, 1H, H2); [(η3-Ind)Mo(OPPh3)2-
(NCMe)(CO)2]BF4, 7.63-7.49 (m, 30H, Ph), 7.23 (t, 1H, H2),
6.50-6.47 (m, 2H, H5-8), 6.42-6.39 (m, 2H, H5-8), 5.12 (d, 1H,
H2). 1H NMR (NCCD3, 300 MHz, -30 °C, after 2h, δ in ppm):
[(η3-Ind)Mo(OPPh3)2(NCMe)(CO)2]BF4, 7.64-7.47 (m, 30H,
Ph), 7.23 (t, 1H, H2), 6.50-6.47 (m, 2H, H5-8), 6.42-6.39 (m,
2H, H5-8), 5.12 (d, 2H, H1/3).
red microcrystalline complex in 90% yield. The product was
further recrystallized from ether/hexane (1/3) by prolonged
cooling at -30 °C. Anal. Found: C, 55.79; H, 3.16. Calcd for
C18
H12O4Mo: C, 55.69; H, 3.11. Selected IR (KBr, cm-1): 1956,
1863, vs, ν(CO); 1601, 1501. 1H NMR (NCCD3, 300 MHz, -40
°C, δ in ppm): 7.58 (d(br), 2H, Ph), 7.50-7.48 (m, 2H, H5-8),
7.33 (c, 3H, Ph), 7.16-7.14 (m, 2H, H5-8), 6.34 (d, 2H, H1/3),
5.18 (t, 1H, H2). 13C{1H} NMR (NCCD3, 75 MHz, room
temperature, δ in ppm): 248.28, CO; 133.82, 129.85, 129.05,
126.89, 94.00, 78.08.
P r ep a r a tion of [In d Mo{OP (OMe)3}2(CO)2]BF 4 (4).
A
solution of IndMo(η3-C3H5)(CO)2 (0.20 g, 0.65 mmol) in CH2-
Cl2 (20 mL) was treated with HBF4‚Et2O (1 equiv). After 10
min, excess OP(OMe)3 was added (0.16 mL, 1.40 mmol) and
the reaction continued for 2 h at room temperature. After
concentration to ca. 5 mL and addition of ether, a reddish
brown oil precipitates in ca. 90% yield. Several attempts to
solidify the compound by dissolving it in CH2Cl2 and slowly
adding of Et2O or by washing it with hexane and ether failed
to yield the solid compound. Selected IR (KBr, cm-1): 1956,
1873, vs, ν(CO); 1244, ν(PdO). 1H NMR (NCCD3, 300 MHz,
-30 °C, δ in ppm): [(η5-Ind)Mo{OP(OMe)3}2(CO)2]BF4, 7.81-
7.78 (m, 2H, H5-8), 7.64-7.61 (m, 2H, H5-8), 6.18 (d, 2H, H1/3),
5.02 (t, 1H, H2), 3.68 (s, 6H, Me); [(η3-Ind)Mo{OP(OMe)3}2-
(NCMe)(CO)2]BF4, 7.22 (t, 1H, H2), 6.47-6.40 (m, 4H, H5-8),
5.12 (d, 2H, H1/3), 3.68 (s, 6H, Me).
P r ep a r a tion of In d Mo{K2-HC(NP h )2}(CO)2 (10). A solu-
tion of IndMo(η3-C3H5)(CO)2 (0.20 g, 0.65 mmol) in CH2Cl2 (15
n
mL) was treated with HBF4‚Et2O (1 equiv). After 10 min Bu4-
NI (0.25 g, 0.70 mmol) was added and the reaction continued
for 1 h. The complex [IndMo(CO)2I]2 was isolated as described
in the literature7f and dissolved in CH2Cl2 (20 mL). A slight
excess of Ag{HC(NPh)2} (0.21 g, 0.70 mmol) was added, and
after 10 h of stirring at room temperature, the AgI precipitate
was filtered and the solution taken to dryness. The obtained
residue was efficiently extracted with hexane and an orange
powder separated by concentration of the extract solution.
Yield: 60%. Anal. Found: C, 61.95; H, 4.01; N, 6.16. Calcd
for C24H18N2O2Mo: C, 62.35; H, 3.92; N, 6.06. Selected IR
(KBr, cm-1): 1939, 1850, vs, ν(CO). 1H NMR (NCCD3, 300
MHz, -40 °C, δ in ppm): 8.31 (s, 1H, CH), 7.28-7.23 (m, 2H,
H5-8), 7.10-7.07 (m, 2H, H5-8), 7.00-6.94 (c, 10H, Ph), 6.32
(s (br), 2H, H1/3), 5.22 (t, 1H, H2). 13C{1H} NMR (NCCD3, 75
MHz, room temperature, δ in ppm): 237.34, CO; 151.55, CH;
P r ep a r a tion of [In d Mo{(K2-P h SCH2)2}(CO)2]BF 4 (6). A
solution of IndMo(η3-C3H5)(CO)2 (0.20 g, 0.65 mmol) in CH2-
Cl2 (20 mL) was treated with HBF4‚Et2O (1 equiv). After 10
min, PhSCH2CH2SPh was added (0.18 g, 0.75 mmol) and the
reaction left for 1 h at room temperature. After concentration
to ca. 5 mL and addition of ether, an orange product precipi-
tated. This was further recrystallized from CH2Cl2/Et2O to
yield the pure compound in 95% yield. Anal. Found: C, 49.84;
H, 3.48. Calcd for C25H21BF4O2S2Mo: C, 49.83; H, 3.52.
Selected IR (KBr, cm-1): 1983, 1912, vs, ν(CO). 1H NMR (CD2-
Cl2, 300 MHz, room temperature, δ in ppm): 7.55-7.48 (m,
10H, Ph); 7.41-7.38 (m, 4H, H5-8), 6.13 (d, 2H, H1/3), 5.20 (t,
1H, H2), 2.55 (m, 2H, CH2), 1.99 (m, 2H, CH2).
P r ep a r a tion of In d Mo(K2-OOCCF 3)(CO)2 (7). A solution
of IndMo(η3-C3H5)(CO)2 (0.22 g, 0.71 mmol) in CH2Cl2 (15 mL)
was treated with excess CF3COOH. After 2 h the solution was
taken to dryness and the residue was washed at low temper-
ature with hexane/pentane. The dark red microcrystalline
complex was isolated in 98% yield. Anal. Found: C, 41.24;
H, 1.75. Calcd for C13H7F3O4Mo: C, 41.08; H, 1.86. Selected
IR (KBr, cm-1): 1971, 1960, 1904, 1886, vs, ν(CO), 1672
ν(CdO). 1H NMR (CD2Cl2, 300 MHz, room temperature, δ in
ppm): 7.48-7.45 (m, 2H, H5-8), 7.40-7.30 (m, 2H, H5-8), 5.94
(d, 2H, H1/3), 5.45 (t, 1H, H2).
P r ep a r a tion of In d Mo(K2-OOCMe)(CO)2 (8). A solution
of IndMo(η3-C3H5)(CO)2 (0.20 g, 0.65 mmol) in CH2Cl2 (15 mL)
was treated with HBF4‚Et2O (1 equiv). After 10 min excess
anhydrous sodium acetate was added (0.06 g, 0.75 mmol) and
the reaction continued for 5 h, after which time the mixture
was filtered and the resulting solution taken to dryness. The
oily residue was washed with a small portion of cold hexane
and then efficiently extracted with Et2O to give the dark red
microcrystalline complex in 90% yield. The product was
further recrystallized overnight from Et2O/hexane (1/3) at -30
°C. Anal. Found: C, 48.27; H, 3.22. Calcd for C13H10O4Mo:
C, 47.87; H, 3.09. Selected IR (KBr, cm-1): 1956, 1867, 1844,
vs, ν(CO); 1512. 1H NMR (NCCD3, 300 MHz, -40 °C, δ in
ppm): 7.45-7.42 (m, 2H, H5-8), 7.39-7.36 (m, 2H, H5-8), 6.24
(d, 2H, H1/3), 5.09 (t, 1H, H2), 1.48 (s, 3H, Me).
P r ep a r a tion of In d Mo(K2-OOCP h )(CO)2 (9). A solution
of IndMo(η3-C3H5)(CO)2 (0.20 g, 0.65 mmol) in CH2Cl2 (15 mL)
was treated with HBF4‚Et2O (1 equiv). After 10 min excess
anhydrous sodium benzoate (0.11 g, 0.75 mmol) was added and
the reaction continued for 5 h, after which time the mixture
was filtered and the resulting solution taken to dryness. The
oily residue was washed with a small portion of cold hexane
and then very efficiently extracted with Et2O to give a dark
146.16, Ph (C1); 130.46, C5/8; 128.90, Ph(C3/5); 128.48, C6/7
124.18, Ph (C4); 122.01, C4/9; 119.62, Ph (C2/6); 80.13, C2; 77.31,
C1/3
;
.
P r ep a r a t ion of In d Mo{K2-H C(N-p -t olyl)2)}(CO)2 (11).
This complex is prepared by a process entirely similar to that
for IndMo{HC(NPh)2}(CO)2 using a slight excess of Ag{HC-
(N-p-tolyl)2} (0.23 g, 0.70 mmol). Yield: 60%. Anal. Found:
C, 63.85; H, 4.54; N, 5.91. Calcd for C26H22N2O2Mo: C, 63.68;
H, 4.52; N, 5.71. Selected IR (KBr, cm-1): 1946, 1854, 1840
vs, ν(CO). 1H NMR (NCCD3, 300 MHz, -40 °C, δ in ppm):
8.22 (s, 1H, CH), 7.09-7.04 (d (br), 6H, tolyl +H5/8), 6.98-
6.94 (m, 2H, H5/8), 6.67-6.64 (d, 4H, tolyl), 6.29 (d, 2H, H1/3),
5.20 (t, 1H, H2), 2.31 (s, 6H, Me). 13C{1H} NMR (NCCD3, 75
MHz, room temperature, δ in ppm): 257.53, CO; 148.29, CH;
142.19, Ph (C1); 130.64, Ph (C4); 129.00, Ph (C3/5); 128.39;
126.27; 123.88; 123.66; 118.31, Ph (C2/6); 89.71; 19.28, Me.
P r ep a r a tion of In d Mo{K2-S2CNEt2}(CO)2 (12). A solu-
tion of IndMo(η3-C3H5)(CO)2 (0.20 g, 0.65 mmol) in CH2Cl2 (15
mL) was treated with HBF4‚Et2O (1 equiv). After 10 min
excess NEt2CS2NH4 (0.11 g, 0.75 mmol) was added and the
reaction continued for 1 h, after which time the reaction
mixture was filtered and the resulting solution taken to
dryness. The oily residue was extracted with toluene to give
an oil which was extracted with a small portion of ether. By
slow concentration and cooling of the Et2O/hexane (1/5) solu-
tion the ruby red crystalline complex was isolated in 70% yield.
Anal. Found: C, 46.61; H, 4.10; N, 3.08. Calcd for C16H17NO2-
S2Mo: C, 46.27; H, 4.12; N, 3.37. Selected IR (KBr, cm-1):
1940, 1854, vs, ν(CO). 1H NMR (NCCD3, 300 MHz, -40 °C, δ
in ppm): 7.24 (s, 4H, H5-8), 6.11 (d, 2H, H1/3), 5.29 (t, 1H, H2),
3.73-3.46 (m, 2H, CH2), 3.48-3.41 (m, 2H, CH2), 1.12 (t, 6H,
Me).
P r ep a r a tion of In d Mo{K2-S2P (OEt)2}(CO)2 (13). A solu-
tion of IndMo(η3-C3H5)(CO)2 (0.20 g, 0.65 mmol) in CH2Cl2 (15
mL) was treated with HBF4‚Et2O (1 equiv). After 10 min
excess (OEt)2PS2NH4 (0.15 g, 0.75 mmol) was added and the
reaction continued for 10 h, after which time the mixture was
filtered and the resulting solution taken to dryness. The oily
residue was extracted with hexane to give the purple crystal-
line complex in 98% yield. Anal. Found: C, 39.72; H, 4.16.
Calcd for C15H17O4PS2Mo: C, 39.83; H, 3.79. Selected IR (KBr,
cm-1): 1956, 1879, vs, ν(CO); 1007. 1H NMR (NCCD3, 300
MHz, -40 °C, δ in ppm): 7.32-7.26 (m, 4H, H5-8), 6.16 (d,

2610 Organometallics, Vol. 17, No. 12, 1998
Calhorda et al.
2H, H1/3), 5.18 (t, 1H, H2), 3.62-3.48 (m, 4H, CH2), 1.15 (t,
6H, Me). 13C{1H} NMR (NCCD3, 75 MHz, room temperature,
Ta ble 2. Cr ysta llogr a p h ic Da ta for
[(η5-In d )Mo(CNMe)2(CO)2]BF 4 (2a )
δ in ppm): 254.88, CO; 128.75, C5/8; 127.59, C6/7; 122.59, C4/9
;
chem formula C15H13BF4MoN2O2
88.76, C2; 78.86, C1/3; 77.20; 64.99 (CH2); 16.34 (Me).
fw
436.02
P r ep a r a tion of In d Mo(K2-a ca c)(CO)2 (14). A solution of
IndMo(η3-C3H5)(CO)2 (0.18 g, 0.58 mmol) in CH2Cl2 (15 mL)
was treated with HBF4‚Et2O (1 equiv). After 10 min, excess
acetylacetone was added (0.5 mL) and the reaction continued
for 2 h. Excess NEt3 was added, and the mixture was left for
1 h. After that the solution was taken to dryness and the
residue was extracted with toluene. The resulting solution
was taken to dryness and this powder was recrystallized
overnight from Et2O/hexane at -30 °C. Yield: 88%. Anal.
Found: C, 52.39; H, 3.80. Calcd for C16H14O4Mo: C, 52.48;
H, 3.85. Selected IR (KBr, cm-1): 1948, 1840, vs, ν(CO). 1H
NMR (CDCl3, 300 MHz, room temperature, δ in ppm): 7.33-
7.30 (m, 2H, H5-8), 7.09-7.04 (m, 2H, H5-8), 6.05 (d, 2H, H1/3),
5.03 (t, 1H, H2), 4.75 (s, 1H, -CH), 1.71 (s, 6H, Me). 1H NMR
(NCCD3, 300 MHz, -45 °C, δ in ppm): 7.11 (t, 1H, H2), 6.41 (s
(br), 4H, H5-8), 5.32 (s, 1H, -CH), 4.95 (s (br), H1/3), 4.84 (s
(br), H1/3), 2.31 (s, 3H, Me), 1.89 (s, 6H, Me).
color/shape
cryst size (mm)
cryst syst
space group
a (pm)
yellow/needle
0.46 × 0.15 × 0.08
monoclinic
P21/c
725.01(3)
1398.34(9)
1693.49(11)
90
90.990(6)
90
1716.6(2)
4
b (pm)
c (pm)
R (deg)
â (deg)
γ (deg)
V (106 pm3)
Z
T (K)
193
1.687
8.1
864
Fcalcd (g cm-3
)
µ (cm-1
F000
)
λ (pm)
scan method
θ range (deg)
data collcd (h,k,l)
abs cor
71.073
imaging plate
3.17-25.67
(8, (15, (20
none
P r ep a r a t ion of In d Mo{K2-MeCOCH C(NP h )Me}(CO)2
(15). A solution of IndMo(η3-C3H5)(CO)2 (0.20 g, 0.65 mmol)
in CH2Cl2 (15 mL) was treated with HBF4‚Et2O (1 equiv). After
transmissn coeff
no. of rflns collcd
no. of indep rflns
no. of obsd rflns
no. of params refined
Rint
17 303
2179
2179 (all data)
254
0.060
0.0697
0.1650
1.197
+0.56, -0.48
n
10 min Bu4NI (0.25 g, 0.70 mmol) was added and the reaction
continued for 1 h. The complex [IndMo(CO)2I]2 was isolated
as described in the literature7f and dissolved in THF (20 mL).
A slight excess of Li{MeCOCHC(NPh)Me} (0.13 g, 0.70 mmol)
was added and the reaction continued for 12 h at room
temperature, after which time the mixture was taken to
dryness. The oily residue thus obtained was washed with
small portions of cold toluene and ether to give a pink-red
powder. Yield: 65%. Anal. Found: C, 59.82; H, 4.71; N, 3.41.
Calcd for C22H19NO3Mo: C, 59.87; H, 4.34; N, 3.17. Selected
IR (KBr, cm-1): 1925, 1827, vs, ν(CO).1H NMR (NCCD3, 300
MHz, room temperature, δ in ppm): 7.60-7.53 (m, 2H, H5-8),
7.48-7.44 (m, 5H, Ph), 7.31-7.27 (m, 2H, H5-8), 6.16 (d, 2H,
H1/3), 5.58 (t, 1H, CH), 5.06 (t, 1H, H2), 1.96 (s, 3H, Me), 1.91
(s, 3H, Me).1H NMR (NCCD3, 300 MHz, -30 °C, after 1 h, δ
in ppm): Complex 15 and (η3-Ind)Mo{η2-MeC(O)CHC(NPh)-
Me}(NCMe)(CO)2, 7.20 (t, 1H, H2), 6.45-6.42 (m, 2H, H5-8),
6.35-6.33 (m, 2H, H5-8), 5.77 (t, 1H, CH), 4.87 (d, 2H, H1/3),
2.06 (s, 3H, coordinated NCMe), 1.96 (s, 3H, CH3), 1.91 (s, 3H,
Me).
P r ep a r a tion of [(η3-In d )Mo(K3-ttcn )(CO)2]BF 4 (16). A
slight excess of 1,4,7-trithiacyclononane (0.25 g, 1.40 mmol)
was added to a solution of [IndMo(NCMe)2(CO)2]BF4 (0.28 g,
0.65 mmol) in CH2Cl2 (20 mL). The reaction was continued
for 2 h at room temperature, after which time the solution was
concentrated to ca. 5 mL and ether added to precipitate a light
yellow powder. This was further recrystallized from CH2Cl2/
Et2O to give the pure compound in 90% yield. Anal. Found:
C, 38.03; H, 4.10. Calcd for C17H19BF4O2S3Mo: C, 38.22; H,
3.58. Selected IR (KBr, cm-1): 1966, 1892, vs, ν(CO). 1H NMR
(CD2Cl2, 300 MHz, -30 °C, δ in ppm): 7.23 (t, 1H, H2), 6.50-
6.46 (m, 2H, H5-8), 6.45-6.40 (m, 2H, H5-8), 4.64 (d, 2H, H1/3),
3.04-3.00 (m, 4H, CH2), 2.91-2.68 (m, 8H, CH2).
R1a
wR2b
GOFc
∆Fmax/min (e Å-3
)
R1 ) Σ(||Fo| - |Fc||)/Σ|Fo|. wR2 ) [Σw(Fo2 - Fc2)2/Σw(Fo2)2]1/2
.
a
b
c GOF ) [Σw(Fo2 - Fc2)2/(NO - NV)]1/2; w ) SHELXL-93 weights.
Neutral atom scattering factors for all atoms and anomalous
dispersion corrections for the non-hydrogen atoms were taken
from ref 34. All calculations were performed on a DEC 3000
AXP workstation with the STRUX-V35 system, including the
programs PLATON-92,36 PLUTON-92,36 SIR-92,37 and SHELXL-
93.38
Da ta Collection , Str u ctu r e Solu tion , a n d Refin em en t
for Com p lex 2a . A summary of the crystal and experimental
data is reported in Table 2. Preliminary examination and data
collection were carried out on an imaging plate diffraction
system (IPDS; Stoe & Cie) equipped with a rotating anode
(Nonius FR591; 50 kV; 80 mA; 4.0 kW) and graphite-mono-
chromated Mo KR radiation. Data collection was performed
at 193 (293) K within the θ range of 2.2° < θ < 25.7° with an
exposure time of 3.0 min per image (oscillation scan modus
from æ ) 0.0° to 360° (230°) with ∆æ ) 1°). A total number of
17 303 (12 913) reflections were collected; 941 (789) negative
and systematic absent 189 (122) reflections were rejected from
the original data set. After merging, a total of 2179 (2984)
independent reflections remained and were used for all
calculations. Data were corrected for Lorentz and polarization
effects. All crystals of compound 2a were found to be twinned.
P r ep a r a t ion of (η3-In d )Mo(K2-OOCCF 3)(NCMe)(CO)2
(22). By recrystallization of IndMo(η2-OOCCF3)(CO)2 from
NCMe/Et2O, dark red crystals of IndMo(η3-OOCCF3)(NCMe)-
(CO)2 are obtained in 98% yield. Anal. Found: C, 42.97; H,
2.24; N, 3.44. Calcd for C15H10F3O4NMo: C, 42.78; H, 2.39 N,
3.33. Selected IR (KBr, cm-1): 2315, 2288, ν(NC); 1962, 1984,
vs, ν(CO); 1692, ν(CdO). 1H NMR (NCCD3, 300 MHz, -45
°C, δ in ppm): 7.28 (t (br), H2), 6.46 (m (br), H5-8), 5.19 (s (br),
H1/3), 4.76 (s (br), H1/3).
(34) International Tables for Crystallography; Wilson, A. J . C., Ed.;
Kluwer Academic Publishers: Dordrecht, The Netherlands, 1992; Vol.
C, Tables 6.1.1.4 (pp 500-502), 4.2.6.8 (pp 219-222), and 4.2.4.2 (pp
193-199).
(35) Artus, G.; Scherer, W.; Priermeier, T.; Herdtweck, E. STRUX-
V, a Program System To Handle X-ray Data; TU Mu¨nchen: Mu¨nchen,
Germany, 1997.
(36) Spek, A. L. PLATON-92-PLUTON-92, An Integrated Tool for
the Analysis of the Results of a Single-Crystal Structure Determina-
tion. Acta Crystallogr., Sect. A 1990, 46, C34.
X-r a y Cr ysta llogr a p h y. Suitable single crystals for the
X-ray diffraction studies were grown by standard techniques
from saturated solutions of CH2Cl2/Et2O at room temperature.
The structure was solved by a combination of direct methods,
difference Fourier syntheses, and least-squares methods.
(37) Altomare, A.; Cascarano, G.; Giacovazzo, C.; Guagliardi, A.;
Burla, M. C.; Polidori, G.; Camalli, M. SIR-92; University of Bari: Bari,
Italy, 1992.
(38) Sheldrick, G. M. SHELXL-93 In Crystallographic Computing
3; Sheldrick, G. M., Kru¨ger, C., Goddard, R., Eds.; Oxford University
Press: London, 1993; pp 175-189.

Indenyl Ring Slippage in Molybdenum Complexes
Organometallics, Vol. 17, No. 12, 1998 2611
Therefore, raw data were reduced with the program Twin.39
The unit cell parameters were obtained by full-matrix least-
squares refinements of 4354 (1804) reflections with the
program Cell.39 All “heavy atoms” of the asymmetric unit were
anisotropically refined. Hydrogen atoms were calculated in
ideal positions (riding model), included in the structure factor
calculations, but not refined. Full-matrix least-squares refine-
ments were carried out by minimizing Σw(Fo2 - Fc2)2 with the
SHELXL weighting scheme and stopped at shift/error < 0.001.
2.00; Mo-N, 2.17; Mo-C(CNMe), 2.17; Mo-P, 2.47; C-O, 1.15;
C-N, 1.13; C-C (NCMe), 1.45; C-C, 1.40; C-H, 1.08; P-H,
1.4.
DF T Ca lcu la tion s. Density functional calculations23
were
carried out on model compounds based on the structures of
2a for the [(η5-Ind)MoL2(CO)2]+ complexes and that of [(η3-
Ind)Mo(NCMe)2(CO)2]+,18 under Cs symmetry, using the Am-
sterdam Density Functional (ADF) program24 developed by
Baerends and co-workers.41 The slipping, folding, and position
of the indenyl and the bending of the equatorial ligands were
optimized. Vosko, Wilk, and Nusair’s local exchange correla-
tion potential was used,42 with Becke’s nonlocal exchange43 and
Perdew’s correlation corrections.44 The geometry optimization
procedure was based on the method developed by Versluis and
Ziegler,45 using the nonlocal correction terms in the calculation
of the gradients. The core orbitals were frozen for Mo ([1-
3]s, [1-3]p, 3d) and C, N, and O (1s). Triple-ú Slater-type
orbitals (STO) were used for H 1s, C, N, and O 2s and 2p, and
Mo 4s and 4p. A set of polarization functions was added: H
(single ú, 2p); C, N, and O (single ú, 3d).
Molecu la r Or bita l Ca lcu la tion s
Exten d ed Hu1 ck el Ca lcu la tion s. EH calculations were
performed using the ICON8 program22a,b with modified Hij
values.22c The basis set for the metal atoms consisted of ns,
np, and (n - 1)d orbitals. The s and p orbitals were described
by single Slater-type wave functions, and the d orbitals were
taken as contracted linear combinations of two Slater-type
wave functions. Only s and p orbitals were considered for P.
The parameters used for Mo were as follows (Hii (eV), ú): 5s
-8.77, 1.960; 5p -5.60, 1.900; 4d -11.06, 4.542 (ú1), 1.901 (ú2),
0.5899 (C1), 0.5899 (C2). Standard parameters were used for
other atoms. Three-dimensional representations of orbitals
were drawn using the program CACAO.40
Ack n ow led gm en t. This work was supported by
PRAXIS XXI under projects 2/2.1/QUI/316/94 and PBIC/
C/QUI/2201/95. I.S.G. (BPD) and C.A.G. (BD) thank
PRAXIS XXI for grants.
The calculations were performed on model complexes with
idealized geometries and Cs symmetry, taken from the real
structures quoted along the text. Thus, [(η3-Ind)MoL2(CNMe)-
(CO)2]+ (L ) NCMe, CNMe, PH3) complexes were taken as
perfect octahedrons, with all L-Mo-L angles equal to 90°; the
η3 coordination mode of the indenyl ligand was accomplished
by bending the ring across C1-C3 by 30° (see Chart 1). [(η5-
Ind)MoL2(CO)2]+ species have a piano-stool geometry with
Ind-Mo-X angles of 115°. This value, as well as the rota-
tional conformation of the indenyl ligand in all the models,
was the result of a geometry optimization. The bond distances
(Å) were as follows: M-(C5 ring centroid), 2.00; M-C(CO),
Su p p or tin g In for m a tion Ava ila ble: Tables of data col-
lection parameters, bond lengths, bond angles, fractional
atomic coordinates, and thermal parameters (6 pages). Order-
ing information is given on any current masthead page.
OM971005G
(41) (a) Baerends, E. J .; Ellis, D.; Ros, P. Chem. Phys. 1973, 2, 41.
(b) Baerends, E. J .; Ros, P. Int. J . Quantum Chem. 1978, S12, 169. (c)
Boerrigter, P. M., te Velde, G.; Baerends, E. J . Int. J . Quantum Chem.
1988, 33, 87. (d) te Velde, G.; Baerends, E. J . J . Comput. Phys. 1992,
99, 84.
(42) Vosko, S. H.; Wilk, L.; Nusair, M. Can. J . Phys. 1980, 58, 1200.
(43) Becke, A. D. J . Chem. Phys. 1987, 88, 1053.
(44) (a) Perdew, J . P. Phys. Rev. 1986, B33, 8822. (b) Perdew, J . P.
Phys. Rev. 1986, B34, 7406.
(45) (a) Versluis, L.; Ziegler, T. J . Chem. Phys. 1988, 88, 322. (b)
Fan, L.; Ziegler, T. J . Chem. Phys. 1991, 95, 7401.
(39) IPDS Operating System, Version 2.8; Stoe & Cie GmbH,
Darmstadt, Germany, 1997.
(40) Mealli, C.; Proserpio, D. M. J . Chem. Educ. 1990, 67, 39.