Alkyl migration reactions - The direct observation of the preferential migration of branched over linear alkyl groups

Article abstract of DOI:10.1016/S0022-328X(00)00158-3

Three binuclear metal alkyl complexes were synthesised, viz. [Cp(CO)₂Fe{CH(CH₃)CH₂CH₂}Fe(CO) ₂Cp] (I), [CpFe(CO)₂{CH(CH₃)CH₂CH₂CH ₂}Fe(CO)_2<

Full text of DOI:10.1016/S0022-328X(00)00158-3

www.elsevier.nl/locate/jorganchem
Journal of Organometallic Chemistry 604 (2000) 7–11
Alkyl migration reactions—the direct observation of the
preferential migration of branched over linear alkyl groups
Emily J. Cammell, Jo-Ann M. Andersen *
School of Chemistry, Uni6ersity of Birmingham, Edgbaston, Birmingham B15 2TT, UK
Received 2 June 1999; received in revised form 28 February 2000
Abstract
Three binuclear metal alkyl complexes were synthesised, viz. [Cp(CO)2Fe{CH(CH3)CH2CH2}Fe(CO)2Cp] (I),
[CpFe(CO)2{CH(CH3)CH2CH2CH2}Fe(CO)2Cp] (II) and [CpMo(CO)3{CH(CH3)CH2CH2}Mo(CO)3Cp] (III) (Cp=h5-cyclopen-
tadienyl). These compounds were then reacted with PPh3 in order to induce alkyl migration (CO insertion) reactions. The products
obtained were monosubstituted phosphine acyl species. In all cases, the branched chain alkyl fragment (viz. MꢀCH(CH3)) was
observed to react preferentially over the linear fragment (viz. MꢀCH2CH2). These results are rationalised in terms of the relative
importance of steric and electronic effects in alkyl migration reactions. © 2000 Elsevier Science S.A. All rights reserved.
Keywords: Alkyl migration; Catalysis; Selectivity; Branched; Linear
1. Introduction
can occur via two routes, either anti-Markovnikov mi-
gration to give the linear alkyl species (reaction (ii)) or
Markovnikov migration to give the branched alkyl
species (reaction (vii)). It is the further reactions of this
alkyl species, either linear or branched, which deter-
mines the nature of the products obtained from the
reaction.
A very high proportion of asymmetric catalysts that
have been developed are organometallic compounds.
High selectivity can be achieved by selecting the proper
catalyst, substrate and reaction conditions. Many of the
elementary steps of enantioselective catalytic reactions
are reversible and it is the first irreversible step which
involves chiral transition states that determines the
nature of the products. To achieve a high degree of
stereoselectivity, efforts must be focused towards a
single, chirality-determining transition state.
An important application of asymmetric catalysis is
the hydroformylation of prochiral olefins such as vinyl
aromatics and other vinylic species, e.g. vinyl acetate, to
produce aldehydes which are intermediates in the syn-
thesis of steroids, pheromones, antibiotics and anti-infl-
ammatories [1]. A catalytic cycle for rhodium-catalysed
olefin hydroformylation is shown in Scheme 1. The
initial step in the cycle is loss of a CO ligand followed
by coordination of the substrate olefin to form a metal
hydrido olefin complex (reaction (i)). Hydride migra-
tion onto the coordinated olefin then occurs, but this
It has been shown that at higher temperatures, reac-
tion (vii) can be reversible for branched alkyl species
obtained from certain vinylic substrates [2,3]; the alkyl
intermediate undergoes a facile b-hydride elimination
reaction to revert back to the hydride olefin complex.
This can then react again to form either the linear or
branched alkyl species. In asymmetric catalysis, the
branched species is the desired product as it contains
the new chiral centre. The b-hydride elimination reac-
tion is thus detrimental to the selectivity of asymmetric
hydroformylation (and other) reactions and should be
minimised as it constitutes a ‘draining away’ of the
chiral alkyl intermediate. One way of overcoming this
problem is to facilitate the alkyl migration reaction
(reactions (iv) and (ix)) so that the branched alkyl
group rapidly migrates to form the branched acyl spe-
cies, for which b-hydride elimination does not occur.
Thus, there are two selectivities which need to be
optimised in asymmetric catalytic reactions, viz. re-
* Corresponding author. Fax: +44-121-4147871.
E-mail address: j.m.andersen@bham.ac.uk (J.-A.M. Andersen).
0022-328X/00/$ - see front matter © 2000 Elsevier Science S.A. All rights reserved.
PII: S0022-328X(00)00158-3

8
E.J. Cammell, J.-A.M. Andersen / Journal of Organometallic Chemistry 604 (2000) 7–11
gioselectivity (linear versus branched) and enantioselec-
tivity (R versus S). While there is a great deal of work
being done on optimising the enantioselectivity of a
reaction by, for example, the use of chiral ligands and
auxiliaries [1,4], the issue of regioselectivity has received
considerably less attention. The key intermediates
formed during catalytic hydroformylation reactions are
metal carbonyl alkyl species and the way in which they
react further, especially how linear species react relative
to branched species, is crucial in determining the overall
selectivity of the reaction.
which contains both a branched alkyl moiety attached
to the metal (MꢀCH(CH3)) and a linear alkyl moiety
attached to the metal (MꢀCH2CH2). Reactions of I, II
and III with one equivalent of PPh3 should result in
substitution at only one end of the bridging alkyl unit,
either linear or branched. The purpose of this study was
to determine (a) whether reaction would occur at the
linear or at the branched end and to rationalise this in
terms of steric and electronic factors and (b) whether
the iron complexes would react in the same way as the
molybdenum complexes.
It is thus surprising that very little is known about
the relative migratory aptitudes of linear versus
branched alkyl groups. A few isolated studies would
seem to indicate that branched alkyl groups migrate
more rapidly than linear alkyl groups [5–8]; however,
not all of these studies provided a complete analysis of
the results. Both steric and electronic parameters should
be taken into account and these can often give conflict-
ing expectations as to the outcome of a reaction. These
factors are discussed in more detail in the Section 2.
We now report our results on the synthesis and alkyl
migration reactions of three binuclear complexes, viz.
2. Results and discussion
2.1. Synthesis
Compound I, [Cp(CO)2Fe{CH(CH3)CH2CH3}Fe-
(CO)2Cp], has been reported previously [9]. We have
synthesised it again by an analogous route to that used
before, and also the new compounds [Cp(CO)2Fe-
{CH(CH3)CH2CH2CH2}Fe(CO)2Cp]
(II)
and
[Cp(CO)3Mo{CH(CH3)CH2CH2}Mo(CO)3Cp] (III). All
three compounds were synthesised by reaction of the
cyclopentadienyl metal carbonyl anion with the appro-
priate dibromoalkane, as shown in Scheme 2. I, II
[CpFe(CO)2Fe{CH(CH3)CH2CH2}Fe(CO)2Cp]
[Cp(CO)2Fe{CH(CH3)ꢀCH2CH2CH2}Fe(CO)2Cp] (II)
and [Cp(CO)3Mo{CH(CH3)CH2CH2}Mo(CO)3Cp]
(III). These complexes all contain a bridging alkyl unit
(I),
Scheme 1. Hydroformylation of olefins, showing both linear and branched products.

E.J. Cammell, J.-A.M. Andersen / Journal of Organometallic Chemistry 604 (2000) 7–11
9
Scheme 2.
Scheme 3. Alkyl migration. ꢀ, vacant coordination site.
and III were isolated as yellow crystalline solids and
were characterised by melting point, microanalysis, IR
and 1H-NMR. Characterisation data are reported in
Tables 1 and 2.
extent by branched alkyls than linear alkyls. The faster
migration of more electron-donating alkyl groups has
already been established [10,11].
However, on steric grounds the situation can be
reversed. A branched alkyl group is bulkier than its
linear analogue and might thus migrate on to the
adjacent carbonyl ligand more slowly than the smaller
linear group. This has also been established previously
[10–13]. In addition, the product containing the bulkier
branched alkyl group adjacent to the incoming nucle-
ophile (e.g. PPh3) will experience a greater degree of
unfavourable steric interactions. There is, however, an
alternative steric effect which might operate in some
cases, which is that if the entering group (e.g. PPh3) is
not involved in the reaction at all, then one could argue
that the bulkier (branched) alkyl should react faster
since steric strain is being alleviated on going from the
more hindered alkyl to the open acyl. These two oppos-
2.2. Reaction of I, II and III with PPh3
Alkyl migration onto a coordinated carbonyl ligand,
or CO insertion as it is often referred to, is an in-
tramolecular nucleophilic attack of the alkyl group on
to the carbon atom of the carbonyl ligand to form a
metal acyl species, as represented in Scheme 3.
Branched alkyl groups should therefore migrate
faster in to the carbonyl ligand than linear alkyl groups
as they will be more electron-donating and will thus
experience a greater electrostatic attraction towards the
d+ carbon atom. In addition, the resultant acyl species
is electron-deficient and will be stabilised to a greater
Table 1
Data for compounds I, II and III and products of reaction with PPh3
a
Compound No.
Yield (%)
M.P. (°C)
IR w(CO) (cm−1
)
Elemental analysis
C; Found (calc.) H Found (calc.)
I
II
III
I+PPh3
II+PPh3
III+PPh3
I b
26
31
18
52
63
57
22
86–88
90–93
76–80
105–108
112–116
98–102
84–87
2006vs, 211vs, 1949vs
2006vs, 2001vs, 1947vs
52.9 (52.7)
53.5 (53.8)
44.3 (44.0)
64.1 (64.4)
64.5 (64.8)
57.1 (57.5)
52.4 (52.7)
4.7 (4.4)
4.5 (4.7)
3.5 (3.3)
4.3 (4.8)
4.6 (5.0)
4.0 (4.3)
4.7 (4.4)
2046m, 2019s, 1981s, 1931s
2005mw, 1948m, 1920mw, 1661w
2005mw, 1945m, 1919mw, 1657w
2046mw, 2020mw, 1932m, 1665w
2009vs, 2003vs, 1955vs
a KBr discs; vs=very strong, s=strong, m=medium, w=weak.
b Data taken from Ref. [5].
Table 2
1H-NMR data for compounds I, II and III a and products of reaction with PPh3
Compound no. MꢀCH or
MꢀCOCH
MꢀCH(CH3) or
MꢀCOCH(CH3)
MꢀCH2
MꢀCH2(CH2)n
Cp
PPh3
I
II
III
I+PPh3
II+PPh3
III+PPh3
I b
2.2 (c, 1H)
2.1 (c, 1H)
2.5 (c, 1H)
1.9 (c, 1H)
1.86 (c, 1H)
2.7 (c, 1H)
2.48 (br, 1H)
1.05 (d, 3H, J=7Hz)
1.09 (d, 3H, J=6Hz)
1.42 (br, 3H)
0.81 (br, 3H)
0.82 (br, 3H)
1.35 (br, 2H)
1.38 (br, 2H)
1.75 (br, 2H)
1.33 (br, 2H)
1.38 (br, 2H)
1.75 (br, 2H)
1.63 (m, 4H)
1.56 (c, 2H)
1.63 (c, 4H)
2.18 (c, 2H)
1.59 (c, 2H)
1.65 (br, 4H)
2.18 (c, 2H)
1.63 (m, 4H)
4.73 (s, 10H)
4.72 (s, 10H)
5.30 (s, 10H)
4.72 (br, 10H)
4.72 (c, 10H)
5.28 (c, 10H)
4.70 (s, 10H)
7.62 (c, 15H)
7.66 (c, 15H)
7.82 (c, 15H)
1.25 (br, 3H)
1.36 (d, 3H, J=6Hz)
a In CDCl3, relative to TMS (l=0.00 ppm), s=singlet, d=doublet, br=broad, c=complex, m=multiplet
b Data taken from Ref. [5].

10
E.J. Cammell, J.-A.M. Andersen / Journal of Organometallic Chemistry 604 (2000) 7–11
comparison to in situ catalytic processes than perform-
ing two separate studies.
Compounds I, II and III were reacted with one
equivalent of PPh3, as represented in Scheme 5, in order
to determine which end of the molecule (linear or
branched) would react preferentially. All reactions were
performed in hexane at 45°C.
The products of the reaction were the phosphine-sub-
stituted acyl complexes where reaction had occurred at
the branched end of the molecule. The reactions were
monitored by FTIR and the reaction products charac-
Scheme 4. Linear versus branched reactivity.
ing steric effects have been discussed in depth by Cot-
ton and Markwell [13], who introduce the concept of a
‘steric window’ within which reactivity is enhanced. The
more typical effect, however, is the first, viz. that
bulkier alkyl groups undergo migratory insertion more
slowly than do smaller groups; studies reported by
ourselves and other groups have shown this to be the
case [10–13].
1
terised by FTIR, H-NMR and elemental analysis, as
reported in Tables 1 and 2. 1H-NMR was used to
identify which end of the complex had reacted. In all
cases, the signals for MꢀCH and MꢀCH(CH3) had
shifted position, whereas the signal for MꢀCH2 re-
mained unchanged.
1H-NMR spectra were also run of the crude reaction
mixture prior to recrystallisation of the PPh3-substi-
tuted products in order to ascertain that we were not
selectively recrystallising only one product. In all cases,
the NMR spectra of the crude mixtures showed that
only one product had formed.
When comparing linear and branched alkyl groups
and their expected migratory aptitudes, there are thus
two opposing effects to consider, viz. on electronic
grounds, branched alkyl groups should migrate faster,
whereas on steric grounds, the linear alkyl groups
should migrate faster. We thus synthesised compounds
I, II and III as they contain both linear and branched
alkyl moieties. Reaction of these complexes with only
one equivalent of nucleophile should thus result in
reaction at only one end of the molecule. If reaction
occurs at the linear end of the molecule (MꢀCH2CH2),
this then indicates that steric factors dominate over
electronic factors, whereas if reaction occurs at the
branched end of the molecule (MꢀCH(CH3)), this
would indicate that electronic factors dominate. This is
represented in Scheme 4. These experiments thus repre-
sent a direct comparison between migratory insertion
reactions of linear and branched alkyl groups. In addi-
tion, these complexes approximate actual catalytic in-
termediates more closely than do separate studies on
linear and branched alkyl species. A catalytic reaction,
at any given time, will contain a mixture of both linear
and branched alkyl species and it is the relative rates of
the two different species, which will determine the
regioselectivity of reaction. Thus, performing a compe-
tition experiment, as we have done, is a more accurate
3. Conclusions
The alkyl migration reaction occurred, in every case,
at the branched end of the molecule. This is indicative
of electronic factors being of greater significance than
steric factors for the migration of an alkyl group on to
a coordinated carbonyl ligand. This has implications in,
for example, asymmetric hydroformylation reactions. If
the rate of the overall process is increased, this may
result in a greater amount of branched (chiral) product
being formed. The branched alkyl intermediate should
react faster than the linear intermediate, and will thus
drive the catalytic cycle towards forming more
branched product.
However, it should be noted (as was pointed out by
a referee) that the steric demands of the two metal
centres chosen in this study are not very high. The
steric demand in typical hydroformylation catalysts is
Scheme 5. Reaction of I, II and III with PPh3.

E.J. Cammell, J.-A.M. Andersen / Journal of Organometallic Chemistry 604 (2000) 7–11
11
usually higher, so steric effects may become more
important.
THF (20 ml), was added dropwise over ca. 10 min to
1,3-dibromobutane (6.12 mmol) at 0°C with stirring.
The yellow–brown solution was then stirred for 4 h at
r.t. The solvent was removed under reduced pressure,
leaving an orange–red oily residue. This was dissolved
in a minimum of hexane, transferred to a deactivated
alumina chromatography column (2 cm×10 cm) and
eluted with hexane. A pale yellow band containing the
product was eluted first, from which the solvent was
removed under reduced pressure to yield a yellow oily
solid. This was recrystallised from hexane at −15°C to
give a yellow solid in 18% yield. A second, orange–red
band containing the [CpMo(CO)3]2 dimer remained on
the column.
4. Experimental
All reactions were carried out under an atmosphere
of high purity nitrogen using standard Schlenk tube
techniques. Tetrahydrofuran (THF) was dried by dis-
tilling over sodium/benzophenone and hexane was
dried by distilling over sodium wire under nitrogen.
[CpFe(CO)2]2 and [CpMo(CO)3]2 were obtained from
Strem Chemicals Inc. and the compounds BrCH-
(CH3)CH2CH2Br and BrCH(CH3)CH2CH2CH2Br were
obtained from Aldrich and were used without further
purification. Alumina (BDH, active neutral, Brockman
grade I) was deactivated before use.
4.3. Reactions of Compounds I, II and III with PPh3
Compound I, II or III (0.20 mmol) was dissolved in
hexane (10 ml) at 45°C. One equivalent of PPh3 (0.20
mmol) was added to the solution. The solution was
then left to stir overnight at 45°C. After this time, the
Schlenk tube containing the reaction mixture was
cooled down to r.t. When cool, approximately one half
of the solvent was removed under reduced pressure,
and the product left to crystallise at −15°C. The
products were obtained as yellow crystalline solids in
ca. 60% yield.
Melting points were recorded on a Kofler hot-stage
microscope (Reichert–Thermovar) and are uncor-
rected. Infrared spectra were recorded on an Atavar
360 FTIR Spectrometer as either as KBr discs or as
Nujol mulls with NaCl windows.
1H-NMR spectra were recorded on a Bruker AC-300
300MHz Spectrometer. The chemical shifts are relative
to TMS (0 ppm).
4.1. Synthesis of [Cp(CO)2Fe{CH(CH3)CH2CH2}-
Fe(CO)2Cp] (I) and [Cp(CO)2Fe{CH(CH3)ꢀCH2-
CH2CH2}Fe(CO)2Cp] (III)
Acknowledgements
We thank the University of Birmingham for financial
support.
Compound I has previously been reported by Cooke
et al. [9]. The same general procedure was adopted by
us in this study. A solution of Na{CpFe(CO)2] (5.65
mmol), prepared by reductive cleavage of [CpFe(CO)2]2
over Na–Hg in THF (20 ml), was added dropwise over
ca. 10 min to the dibromoalkane (5.65 mmol) at 0°C
with stirring. The yellow–brown solution was then
stirred for 3 h at room temperature (r.t.). The solvent
was removed under reduced pressure, leaving an or-
ange–brown oily residue. This was dissolved in a mini-
mum of hexane, transferred to a deactivated alumina
chromatography column (2 cm×10 cm) and eluted
with hexane. A yellow band containing the product was
eluted first, from which the solvent was removed under
reduced pressure to yield a yellow oily solid. This was
recrystallised from hexane at −15°C to give a yellow–
orange solid in 26% yield (I) or 31% yield (III). A
second, red band containing the [CpFe(CO)2]2 dimer
remained on the column.
References
[1] R. Noyori, Asymmetric Catalysis in Organic Synthesis, Wiley-In-
terscience, New York, 1994.
[2] R. Lazzaroni, G. Ucello-Barretta, S. Scamuzzi, R. Settambolo,
A. Caiazzo, Organometallics 15 (1996) 4657.
[3] T. Horiuchi, E. Shirakawa, K. Nozaki, H. Takaya,
Organometallics 16 (1997) 2981.
[4] J. Seyden-Penne, Chiral Auxiliaries and Ligands in Asymmetric
Catalysis, Wiley-Interscience, New York, 1995.
[5] M. Green, D.J. Westlake, J. Chem. Soc. (A) (1971) 367.
[6] J.D. Cotton, G.T. Crisp, L. Latif, Inorg. Chim. Acta 47 (1981)
171.
[7] K.H. Pannell, T. Giasolli, R.N. Kapoor, J. Organomet. Chem.
316 (1986) 315.
[8] M. Bassetti, L. Mannina, D. Monti, Organometallics 13 (1994)
3293.
[9] M. Cooke, N.J. Forrow, S.A.R. Knox, J. Chem. Soc. Dalton
Trans. (1983) 2435.
[10] J.M. Andersen, J.R. Moss, S. Afr, J. Chem. 50 (1997) 144.
[11] J.N. Cawse, R.A. Fiato, R.L. Pruett, J. Organomet. Chem. 172
(1979) 405.
4.2. Synthesis of [Cp(CO)3Mo{CH(CH3)CH2CH2}-
Mo(CO)3Cp] (II)
[12] J.M. Andersen, J.R. Moss, Organometallics 13 (1994) 5013.
[13] J.D. Cotton, R.D. Markwell, J. Organomet. Chem. 388 (1990)
388.
A solution of Na[CpMo(CO)3] (6.12 mmol), prepared
by reductive cleavage of [CpMo(CO)3]2 over Na–Hg in