Unprecedented double nucleophilic addition of a hydride at a central carbon of an η3-allyl ligand

Article abstract of DOI:10.1039/b407976k

Treatment of η³-allyl compound [Cp ₂Mo(η³-C₃H₅)]⁺ (1; Cp = η⁵-C₅H₅) with MH (M = Li, Na) resulted in reduction of the allyl ligand to give propane. Deuterium-labeling studies were used to trace the origins and fates of the hydrogen atoms. The mechanism is discussed in light of the HSAB principle. The studies showed that the formation of propane can be explained by 1,2-hydrogen migration from the central to the terminal carbon of the allyl ligand, and the subsequent double nucleophilic addition of the hydride at the central carbon.

Full text of DOI:10.1039/b407976k

View Online / Journal Homepage / Table of Contents for this issue
F U L L P A P E R
Unprecedented double nucleophilic addition of a hydride at a central
carbon of an g3-allyl ligand
Makoto Minato,* Ryoko Sekimizu, Daisuke Uchida and Takashi Ito*
Department of Materials Chemistry, Graduate School of Engineering,
Yokohama National University, 79-5 Tokiwadai, Hodogaya-ku, Yokohama, Japan, 240-8501.
E-mail: minato@ynu.ac.jp; Fax: +81-45-339-3933; Tel: +81-45-339-3933
Received 26th May 2004, Accepted 22nd September 2004
First published as an Advance Article on the web 7th October 2004
Treatment of g3-allyl compound [Cp2Mo(g3-C3H5)]+ (1; Cp = g5-C5H5) with MH (M = Li, Na) resulted in reduction
of the allyl ligand to give propane. Deuterium-labeling studies were used to trace the origins and fates of the
hydrogen atoms. The mechanism is discussed in light of the HSAB principle. The studies showed that the formation
of propane can be explained by 1,2-hydrogen migration from the central to the terminal carbon of the allyl ligand,
and the subsequent double nucleophilic addition of the hydride at the central carbon.
hydrocarbon products were observed in this reaction. Thus,
Introduction
treatment of 1 with MH did not result in an acid–base reaction.7
Instead, reduction of the allyl ligand occurred. A key question
relates to the mechanism of the reduction. Since alkali hydrides
are generally regarded as Lewis bases they have not been utilized
as reducing agents. To elucidate the reaction mechanism, studies
with deuterium-labeled reagents were performed (Scheme 2).
Nucleophilic attack on an g3-allyl ligand that is coordinated to
transition metals is of great general interest since a wide range
of important catalytic and stoichiometric processes involve
such reaction.1 The chemistry of g3-allyl complexes is strongly
dependent on the nature of the metal centre. By far the most
extensively studied and most highly developed chemistry is
that of (g3-allyl)palladium complexes, because of their utility
in organic synthesis.2 On the other hand, g3-allyl complexes of
the group VI metals are also attracting considerable attention as
the regioselectivity in nucleophilic substitution is quite different
from that observed for the palladium system.3
We previously showed that dicyclopentadienyl allyl moly-
bdenum complexes, Cp2Mo(g3-allyl)+X− (1) are conveniently
prepared by treatment of molybdenum dihydride, Cp2MoH2
(2) with an excess of allyl alcohols in the presence of protonic
acids HX, such as TsOH (TsO = p-CH3(C6H4)SO3), HCl, HI and
CF3COOH.4 Recently this reaction was successfully extended to
a homoallyl alcohol system.5 We are currently studying the
reactivity of 1 toward nucleophiles and Lewis bases. It is well
known that in dicyclopentadienyl allyl molybdenum complexes,
nucleophiles attack preferentially at the central carbon of the
allyl ligand to give metallacyclobutane derivatives.6 In this
paper we report upon a study of propane formation from the
reaction between 1 and MH (M = Li, Na). The results indicate
that propane obtained from 1 does not occur via the usual
metallacyclobutane process.
Scheme 2 Reduction of 1 using deuterium-labeled reagents.
First, reduction was achieved by reaction with LiD. In this
mass spectrum, the peak of highest mass at 46 is the mole-
cular ion corresponding to a structure C3H6D2. Therefore two
deuterium units were incorporated. The strongest peak at m/z
31 is the characteristic base peak, which can be attributed to
+
1
the CH3CD2 fragment.8 The H NMR spectrum in CDCl3
confirmed the complete deuteration at the central carbon as
the methylene signal at d 1.30 ppm was no longer present. All
the above spectroscopic features are consistent with the formula
CH3CD2CH3. This result indicated that nucleophilic attack
occurs twice at the central carbon of the g3-allyl ligand.
We next explored the use of THF-d8 as the reaction solvent.
The reaction was achieved by heating 1 and LiH in THF-d8 in
a high-pressure NMR tube at 80 °C. The mass spectrum of the
resultant product contains the highest peak at m/z 45 corres-
Results and discussion
Stirring a heterogeneous solution of 1 in THF with strong
base MH (M = Li, Na) at room temperature resulted in no net
reaction. However heating a mixture at 80 °C caused dissolution
and formation of a white precipitate that was identified as
TsOM (Scheme 1).
1
ponding to C3H7D. The H NMR spectrum shows two signals
at d 1.30 (methylene) and 0.90 ppm (methyl), with a 2:5 relative
intensity. Consequently, the resultant propane has the formula
CH3CH2CH2D. In addition, we confirmed that the reaction
of 1 with LiD in THF-d8 forms CH3CD2CH2D. The GC-MS
spectrum contains the highest peak at m/z 47. The strong peaks
+
at m/z 32 and 31 can be attributed to CDH2CD2+ and CDHCD2 ;
CH3CD2+ fragments, respectively. The 1H NMR spectrum shows
only the methyl signals at d 0.90 ppm. These results indicate that
only one deuterium unit from the solvent was incorporated into
one of the methyl groups. Thus a question about the source of
the hydrogen of another methyl group has been raised. A more
plausible scenario involves 1,2-hydrogen migration from the
central to the terminal carbon. In order to prove this hypothesis,
a deuterium-labeled g3-allyl complex 3 was prepared from the
corresponding labeled allyl alcohol.9
Scheme 1 Reaction of 1 with MH.
No detectable complexes were obtained from the reaction
mixture; an unstable molybdenocene Cp2Mo seems to be
formed in this reaction (vide infra). The collaboration of GC
analysis, using methane as internal standard, GC-MS spectro-
metry (EI, 70 eV) and 1H NMR spectra of the volatiles revealed
that propane is the only volatile product (23–32%); no other
T h i s j o u r n a l i s
©
T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 4
D a l t o n T r a n s . , 2 0 0 4 , 3 6 9 5 – 3 6 9 8
3 6 9 5

View Online
In the mass spectrum of the reduction product of 3 with
LiH in THF-d8 (Scheme 3), the highest peak at m/z = 50 is
convincing evidence of the formation of propane containing
GC-MS spectrum contains the highest peak at m/z 47. We
observed a strong peak at m/z 31, which can be attributed
to a CDH2CDH+ fragment. On the other hand we could not
observe the expected CDH2CD2+ fragment at m/z 32 (vide supra).
Interestingly, unlike the reaction of 1 with MH, a substantial
amount of propene was formed, which is consistent with the pre-
vious result.6 In addition, CH2DCHDCH2D was also obtained
when 1 was directly reduced by NaBD4. We thus conclude that
the reduction of 1 with MH does not proceed via the usual
metallocyclobutane process. Taking into account the difference
in reactivity between ME (M = Li, Na; E = H, D) and NaBH4,
a mechanism is proposed that accounts for the 1,2-hydrogen
migration process, followed by the double nucleophilic addition
of the hydride (Scheme 5).
1
six deuterium atoms. The H NMR spectrum shows only the
methylene signals at d 1.30 ppm. These observations confirmed
the presence of six deuteriums located on the terminal carbon
atoms, so the expected product, CD3CH2CD3, was obtained.
Therefore the propane obtained from 3 occurs by a 1,2-
hydrogen migration process followed by double nucleophilic
addition of the hydride.
The reactivity of allyl carbons in dicyclopentadienyl allyl
molybdenum complexes has been investigated theoretically.10
It was pointed out that the central carbon is more reactive than
the terminal carbons toward nucleophiles. Recently, Fujimoto
and co-workers showed that the acidic hardness of the reaction
sites is a key to understand the higher reactivity of the central
carbon.11 They concluded that the central carbon is softer as
Lewis acid in this complex primarily due to its stronger polariz-
ability. Hence, it is reasonable to postulate that the hardness of
the hydride anion has profound effects on the reaction course.
NaBH4 is generally regarded as a hard species, but it is softer
than ME. The difference of hardness of these reagents provides
an explanation for their behaviour toward carbonyl compounds
such as ketones. ME always favors elimination (by attacking the
hard proton at the a-position) whereas NaBH4 tends to attack at
the carbonyl carbon. Furthermore, it is well known that alkali
hydrides react with hard silyl compounds to yield pentaco-
ordinate anions.12 In this reaction, the alkali hydride is utilized
as nucleophile in analogy with the present work.
Softer NaBH4 is able to attack at the central carbon of the allyl
group to give the metallocyclobutane derivative.6 On the other
hand, it is conceivable that 1 does not react with the hard base
ME without rearranging into an isomer that contains a harder
acid centre. Although the reaction of 1 with NaBH4 proceeds at
room temperature, the reaction with ME requires activation by
heating. The thermal reaction would induce a 1,2-shift of the
hydrogen atom giving the intermediate olefin complex A where
the positive charge is mainly located on the central carbon;
hence it now becomes a harder acid centre.13 In this case, we can
draw a resonance limiting form A in which the g2-vinyl complex
contains a metal carbene bond.14 The reaction of A with ME
results in the formation of the metallacyclopropane complex
B.15 Complex B undergoes a second attack by ME to give the
r-alkyl anionic complex C. In this case, the nucleophile will
attack preferentially at the central carbon.16 Proton abstraction
from the solvent,17 followed by reductive elimination, would lead
to the formation of propane and an unstable molybdenocene.
Scheme 3 Reduction of deuterium-labeled allyl complex 3.
As previously noted, in a dicyclopentadienyl allyl moly-
bdenum complex, nucleophilic reaction proceeds via a metalla-
cyclobutane derivative. However, in our case, support for the
formation of an intermediate metallacyclobutane species was
not obtained. Furthermore, in trying to adapt this mechanism
to the present results, we were puzzled by several points. The
most glaring one is the fact that the nucleophilic attack occurs
twice at the central carbon of the g3-allyl ligand. It is also hard
to appreciate why 1,2-hydrogen migration occurs. To establish
whether or not this pathway was operative here, the metallo-
cyclobutane compound 4 that has deuterium on the b-carbon
was prepared by reacting 1 with NaBD4.6 If the propane forma-
tion were to proceed via this metallacyclobutane derivative, the
positions of the deuterium labels in the reduction product would
be completely consistent with those observed for the product of
1 with LiD (see Scheme 2).
Reduction of 4 was achieved by reaction with LiD in THF-d8
1
(Scheme 4). Detailed analysis of GC-MS and the H NMR
spectrum revealed the formation of CH2DCHDCH2D. The
1H NMR spectrum shows two signals at d 1.30 (methylene)
and 0.90 ppm (methyl), with a 1:4 relative intensity. The
Scheme 4 The formation of propane via the metallocyclobutane.
Scheme 5 Proposed mechanism for the formation of propane from 1 and ME (M = Li, Na; E = H, D).
3 6 9 6
D a l t o n T r a n s . , 2 0 0 4 , 3 6 9 5 – 3 6 9 8

View Online
Consistent with this idea, the use of PPh3 allowed us to trap the
molybdenocene as a phosphine adduct Cp2Mo(PPh3).18
NMR-scale reaction of 4 with LiD in THF-d8
A high-pressure NMR tube fitted with a rubber valve was
charged with 4 (16.8 mg, 0.0625 mmol) and LiD (5.3 mg,
0.066 mmol) under argon. THF-d8 (0.5 ml) was added to
the solid via a syringe. Then, the tube was placed in a 80 °C
oil-bath. After 3 h of heating, the 1H NMR spectrum showed
two signals at d 1.30 ppm (methylene) and 0.90 ppm (methyl),
with a 1:4 relative intensity. A sample was removed by
syringe through the rubber valve. The mass spectrum of the
product contained the highest peak at m/z 47 corresponding
to C3H5D3.
Conclusions
In this work we have concentrated on the reaction of 1 with
MH. In the reaction, reduction of the allyl ligand occurred to
give propane. The formation of propane can be explained by
1,2-hydrogen migration from the central to the terminal carbon,
and the subsequent double nucleophilic addition of the hydride
at the central carbon. The key point is that this reaction requires
activation by heating. To the best of our knowledge, this type of
reduction of the g3-allyl ligand is unprecedented.
CH3CD2CH3
Experimental
General procedures
NMR (CDCl3): d 0.90 (s, 6 H). GC-MS: m/z (%) 46
(CH3CD2CH3 , 22), 45 (47), 31 (100), 29 (14), 27 (18) and
+
15 (11).
All manipulations were performed under an inert atmosphere of
nitrogen or argon using standard Schlenk techniques. Commer-
cially available reagent-grade chemicals were used as such with-
out any further purification. Solvents were purified according to
standard procedures. NMR spectra were recorded on a JEOL
JNMEX-270 spectrometer. Chemical shifts are given from
Me4Si as internal standard. Reaction products were analyzed
by GC using a capillary column (Shimadzu Unibeads A) on a
Shimadzu GC-7A gas chromatograph equipped with a flame
ionization detector. Mass spectra (70 eV, EI) were obtained
using a Hewlett-Packard 5890 Series II gas chromatograph
equipped with a JMS-4X500 mass selective detector. Dicyclo-
pentadienyl allyl molybdenum complex 1 was prepared by litera-
ture procedures or modifications of these.4,5 Deuterium-labeled
allyl alcohol was prepared by reduction of DCCCOOD with
LiAlD4 by a standard method.9 The metallacyclobutane 4 was
prepared essentially according to the method described earlier,6
based on the reaction of 1 with NaBD4.
CH3CH2CH2D
NMR (THF-d8): d 1.30 (br, 2 H), 0.90 (t, 5 H). GC-MS: m/z (%)
45 (CH3CH2CH2D+, 100>), 44 (72), 43 (100>), 42 (52), 30 (66),
29 (100>), 28 (71), 27 (100>), 26 (base peak, 100>), 16 (36) and
15 (76).
CH3CD2CH2D
NMR (THF-d8): d 0.90 (br, 5 H). GC-MS: m/z (%) 47
(CH3CD2CH2D+, 21), 46 (70), 45 (51), 43 (16), 32 (97), 31 (100),
30 (11), 29 (23), 28 (13), 27 (33), 26 (11) and 15 (8).
CD3CH2CD3
NMR (THF-d8): d 1.30 (s, 2 H), GC-MS: m/z (%) 50
(CD3CH2CD3 , 11), 49 (6), 48 (100), 47 (4), 46 (62), 45 (3), 43
(4), 42 (12), 32 (8), 30 (28), 29 (3) and 18 (33).
+
Reaction of 1 with NaH
CH2DCHDCH2D
A Schlenk tube fitted with a high-vacuum stopcock and
a septum inlet was charged with 219 mg of 1 (0.503 mmol) and
118 mg of NaH (4.93 mmol). THF (5 ml) was added to the
solids via vacuum transfer, yielding a dark-red suspension. The
reaction mixture was immersed in a liquid-nitrogen bath. Three
freeze–pump–thaw cycles were performed, and the stopcock
was closed under vacuum. After warming to room tempera-
ture, the tube was placed in a 80 °C oil-bath with stirring for
3 h. A white solid precipitated, which was identified as TsONa
by comparison with an authentic sample by IR spectroscopy.
The mixture was then cooled to room temperature. A definite
volume of methane (internal standard) was introduced through
the septum inlet via a syringe. Next, an atmosphere of argon
was admitted to the tube through the stopcock. The resulting
reaction mixture was warmed to 30 °C in a water-bath for 2 h.
A volatile sample was removed by syringe through the septum
inlet. The product was identified as propane (32%, based on 1)
by comparison with an authentic sample by GC coinjection,
GC-MS analysis and 1H NMR. A similar experiment employing
LiH or LiD gave propane in a 23% yield. The same procedures
were used for each bulk scale experiment.
NMR (THF-d8): d 1.30 (br, 1 H), 0.90 (br, 4 H). GC-MS: m/z
(%) 47 (CH2DCHDCH2D+, 12), 46 (42), 45 (37), 43 (17), 31
(100), 30 (12), 29 (47), 28 (13), 27 (51), 26 (23), 16 (4), 15 (14)
and 14 (9).
Acknowledgements
We gratefully acknowledge several useful discussions with
Mr Takeo Kaneko of Yokohama National University.
References and notes
1 G. Consiglio and R. M. Waymouth, Chem. Rev., 1989, 89,
257; B. M. Trost and D. L. V. Vranken, Chem. Rev., 1996, 96,
395.
2 (a) J. Tsuji, Organic Synthesis with Palladium Compounds,
Springer-Verlag, New York, 1980; (b) J. Tsuji, Palladium Reagents
and Catalysts, Wiley, Chichester, 1995; (c) B. M. Trost and I. R.
Verhoeven, in Comprehensive Organometallic Chemistry, ed.
G. Wilkinson, F. G. A. Stone and E. W. Abel, Pergamon Press,
Oxford, 1982, vol. 8, ch. 57, pp. 799–938.
3 B. M. Trost and I. Hachiya, J. Am. Chem. Soc., 1998, 120, 1104.
4 (a) T. Igarashi and T. Ito, Chem. Lett., 1985, 1699; (b) T. Ito and
T. Igarashi, Organometallics, 1987, 6, 199; (c) T. Ito, K. Haji,
F. Suzuki and T. Igarashi, Organometallics, 1993, 12, 2325;
(d) M. Minato, J. Ren, H. Tomita, T. Tokunaga, F. Suzuki,
T. Igarashi and T. Ito, J. Organomet. Chem., 1994, 473, 149.
5 M. Minato, S. Hiratsuka, R. Sekimizu, T. Ito and K. Osakada,
J. Organomet. Chem., 2004, 689, 1025.
6 (a) M. Ephritikhine, M. L. H. Green and R. E. MacKenzie, J. Chem.
Soc., Chem. Commun., 1976, 619; (b) M. Ephritikhine, B. R. Francis,
M. L. H. Green, R. E. MacKenzie and M. J. Smith, J. Chem. Soc.,
Dalton Trans., 1977, 1131.
7 A. J. Pearson and V. D. Khetani, J. Chem. Soc., Chem. Commun.,
1986, 1772.
NMR-scale reaction of 1 with LiH in THF-d8
A pressure NMR tube fitted with a rubber valve was charged
with 1 (41.8 mg, 0.0958 mmol) and LiH (1.90 mg, 0.239 mmol)
under argon. THF-d8 (0.5 ml) was added to the solid via a
syringe. Then, the tube was placed in a 80 °C oil-bath. After
3 h of heating, the 1H NMR spectrum showed two signals at d
1.30 ppm (methylene) and 0.90 ppm (methyl), with a 2:5 relative
intensity. A sample was removed by syringe through the rubber
valve. The mass spectrum of the product contained the highest
peak at m/z 45 corresponding to C3H7D. The same procedures
were used for each NMR experiment.
8 C. Lifshitz and M. Shapiro, J. Chem. Phys., 1967, 46, 4912.
9 T. Joachim, M. Holger and H. Albert, Synthesis, 1985, 8, 775.
D a l t o n T r a n s . , 2 0 0 4 , 3 6 9 5 – 3 6 9 8
3 6 9 7

View Online
10 (a) J. W. Lauher and R. Hoffmann, J. Am. Chem. Soc., 1976, 98, 1729;
(b) M. D. Curtis and O. Eisenstein, Organometallics, 1984, 3, 887.
11 T. Suzuki, G. Okada, Y. Hioki and H. Fujimoto, Organometallics,
2003, 22, 3649.
12 (a) B. Becker, R. J. P. Corriu, C. Guérin, B. Henner and Q. Wang,
J. Organomet. Chem., 1989, 368, C25; (b) R. J. P. Corriu, C. Guérin,
B. J. L. Henner and Q. Wang, Organometallics, 1991, 10, 3574.
13 T.-L. Ho, Tetrahedron, 1985, 41, 1.
14 (a) C. J. Beddows, A. D. Burrows, N. G. Connelly, M. Green,
J. M. Lynam and T. J. Peget, Organometallics, 2001, 20, 231;
(b) M. L. Kuznetsov, A. J. L. Pombeiro and A. I. Dement’ev,
J. Chem. Soc., Dalton Trans., 2000, 4413.
15 A similar structure has been suggested for Cp2Mo(C2H4), see:
(a) J. L. Thomas, J. Am. Chem. Soc., 1973, 95, 1838; (b) F. W. S.
Benfield and M. L. H. Green, J. Chem. Soc., Dalton Trans., 1973,
1324.
16 This regioselectivity is best understood by considering the p-
complex form of B. With substituted olefins, nucleophilic attack
occurs predominantly at the more-substituted olefin terminus,
see: J. P. Collman, L. S. Hegedus, J. R. Norton and R. G. Finke, in
Principles and Applications of Organotransition Metal Chemistry,
University Science Book, Mill Valley, CA, 1987, pp. 409–410.
17 A referee has raised the possibility that the reaction may occur
by abstraction of hydrogen radical. We cannot rule this out in the
system. However taking into account the reactivity of THF toward
basic organometallic compounds at high temperature, we think that
the intermediate C can abstract the proton from THF at 80 °C,
see: M. Schlosser, in Organometallics in Synthesis A Manual, ed.
M. Schlosser, John Wiley & Sons, Baffins Lane, Chichester, 2002,
pp. 289–293.
18 T. Ito, T. Tokunaga, M. Minato and T. Nakamura, Chem. Lett.,
1991, 1893.
3 6 9 8
D a l t o n T r a n s . , 2 0 0 4 , 3 6 9 5 – 3 6 9 8