Novel benzoferrocenyl chiral ligands: Synthesis and evaluation of their suitability for asymmetric catalysis

Article abstract of DOI:10.1016/j.jorganchem.2007.01.004

Four benzoferrocenyl phosphorus chiral ligands were conveniently prepared in good overall yields. These ligands were found to be stable in solid form and in solution. Two of the four ligands were resolved by chiral HPLC. Unlike a reported bis(phosphino-η<

Full text of DOI:10.1016/j.jorganchem.2007.01.004

Journal of Organometallic Chemistry 692 (2007) 1956–1962
www.elsevier.com/locate/jorganchem
Novel benzoferrocenyl chiral ligands: Synthesis and evaluation
of their suitability for asymmetric catalysis
*
Muralidhara Thimmaiah, Rudy L. Luck, Shiyue Fang
Department of Chemistry, Michigan Technological University, 1400 Townsend Drive, Houghton, MI 49931, USA
Received 16 October 2006; received in revised form 4 January 2007; accepted 4 January 2007
Available online 14 January 2007
Abstract
Four benzoferrocenyl phosphorus chiral ligands were conveniently prepared in good overall yields. These ligands were found to be
stable in solid form and in solution. Two of the four ligands were resolved by chiral HPLC. Unlike a reported bis(phosphino-g5-inde-
nyl)iron(II) complex, in which the indenyl ligands undergo ring flipping through an g1-intermediate, these two ligands were found to be
configurationally stable in solution and in solid state. The suitability of these ligands for enantioselective catalysis was assessed in studies
on allylic alkylation reactions. When the two less sterically hindered ligands were used, excellent chemical yields were obtained, but the
other two more sterically hindered ones gave lower yields. When the two enantiopure ligands were used, enantioselectivity of up to 51%
ee was observed. These findings suggest that benzoferrocene derivatives may be used as chiral ligands for asymmetric catalysis.
ꢀ 2007 Elsevier B.V. All rights reserved.
Keywords: Benzoferrocene; Phosphorus ligands; Catalysis; Planar chiral
1. Introduction
directly or from a carbon atom that is attached to the Cp
ring (Fig. 1, 1 and 2). Using 1 and 2 as subunits, by com-
bining the two in various ways, attaching chiral groups
to ‘‘D’’, or further functionalization of the Cps, numerous
chiral ligands have been designed and synthesized. In addi-
tion to ligands based on 1 and 2, although with much fewer
examples, chiral ligands based on structures 3 and 4, where
‘‘D’’ is a nitrogen or phosphorus atom, have been synthe-
sized; Fu and co-workers used such ligands for copper cat-
alyzed conjugate addition of diethylzinc to enones [3] and
for carbene insertion into O–H bonds [4]. Ferrocene deriv-
atives based on 4 have also been used for organocatalyses
such as kinetic resolution of secondary alcohols [5].
We recently initiated a project that is aimed toward the
design and synthesis of planar chiral benzoferrocene
ligands and their applications in asymmetric catalytic reac-
tions. This new type of ligand contains the common struc-
ture unit 5 (Fig. 1), in which ‘‘D’’ is attached to C-4 of the
benzoferrocene. Unit 5 by itself is a monodentate ligand,
but through functionalization of positions such as C-3
and C-5 and/or attaching groups to ‘‘D’’, a new family of
multidentate ligand can be designed and synthesized. This
Catalytic asymmetric reactions are most constructive for
the synthesis of enantiopure compounds because with this
method sub-stoichiometric amounts of enantiopure cata-
lyst can produce large amounts of enantiomerically pure
or enriched product [1]. In transition metal catalyzed asym-
metric reactions, to achieve high enantioselectivity, the
design and synthesis of chiral ligands are crucial. Conse-
quently, numerous ligands incorporating central, axial or
planar chiralities were developed; and in many cases, incor-
porating two or more types of chiralities in one ligand was
found to be beneficial. Planar chiral ligands based on ferro-
cene derivatives have found wide application because of
their well-defined rigid conformation and ability to bring
chirality close to reaction centers [2]. Most of these ligands
contain electron donating heteroatoms (D) such as phos-
phorus, sulfur, nitrogen or oxygen that can coordinate with
transition metals either from the cyclopentadienyl (Cp) ring
*
Corresponding author.
E-mail address: shifang@mtu.edu (S. Fang).
0022-328X/$ - see front matter ꢀ 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.jorganchem.2007.01.004

M. Thimmaiah et al. / Journal of Organometallic Chemistry 692 (2007) 1956–1962
1957
7
D
1
3
D
D
R
Fe
Fe
Fe
Fe
Fe
D
4
1
2
3
4
5
D
D: Electron donating heteroatoms such as P, N, S and O
Fig. 1. Basic structural units of ferrocenyl ligands.
type of ligand is expected to have the following favorable
structural features: the position of ‘‘D’’ shown in 5 is fixed
due to the rigid conformation of benzoferrocene, hetero-
atom ‘‘D’’ shown in 5 is in a biased environment created
by the ferrocenyl unit and the electron density on ‘‘D’’
can be tuned by attaching different functionalities at C-7.
Moreover, the size of the Cp in 5 can be adjusted to achieve
suitable steric hindrance and when multidentate ligands are
used, a bulky Cp group can shield one face of the catalyst
with a well-defined rigid conformation. In this paper, we
describe the synthesis and characterization of the simplest
ligands of this family 6a–d (Scheme 1) and their application
in allylic alkylation reactions. Our goal is to demonstrate
that these new ligands can be easily prepared, are chemi-
cally and configurationally stable, and are compatible with
transition metal catalyzed asymmetric reaction conditions.
tive decomposition of 8 during workup. Because 8, 7 and
H–Cp* are all non-polar and have very similar Rf values,
isolation of more pure 8 from the mother liquor by recrys-
tallization and flash column chromatography was not suc-
cessful. The pure product 8 is a deep purple crystalline solid
and it is highly soluble in solvents such as hexane, ether,
CH2Cl2, CHCl3, THF, EtOAc, benzene and toluene. Once
in its pure form, solid 8 can be stored at room temperature
for long periods even exposed to air. In solution in solvents
such as hexane and diethyl ether at room temperature, no
decomposition of 8 was observed within two weeks even
exposed to air.
Converting 8 to 6a–d was achieved by treating 8 with t-
BuLi (n-BuLi gave the same results) followed by addition
of P-chlorodiphenylphosphine (6a) or P-chlorodialkyl-
phosphines (6b–d) and in all cases, good yields were
obtained. These compounds were easily purified by flash
column chromatography. In pure form, they are purple sol-
ids and are stable exposed to air. However, in solution in
CDCl3 exposed to the atmosphere, slow oxidation of the
pendant phosphorus atom to phosphorus oxide was
observed in 31P NMR spectra.
2. Results
The ligands 6a–d were prepared from the easily obtain-
able indene derivative 7 [6] in two steps (Scheme 1). To
avoid possible debromination, 7 in THF was carefully
treated with n-BuLi at ꢀ78 ꢁC and warmed to room tem-
perature slowly. Concurrently, lithium pentamethyl-cyclo-
pentadienylide (Li–Cp*), a white suspension in THF, was
prepared by treating H–Cp* with n-BuLi at room temper-
ature. The suspension of Li–Cp* was transferred to a sus-
pension of anhydrous FeCl2 in THF via a cannula. After
stirring this reaction mixture at room temperature for
1 h, the solution of lithiated 7 was added dropwise via a
cannula, and the resulting deep purple solution was stirred
at room temperature. After 2 h, TLC indicated complete
consumption of indene 7. However, after aqueous workup
(10% K2CO3) and recrystallization (cold hexane), only 63%
yield of pure product 8 could be obtained; a significant
quantity of product remained in the mother liquor. In addi-
tion to product 8, the mother liquor also contained some
starting material 7 and H–Cp*, which resulted from oxida-
The 1H and 13C NMR spectra of compound 8 are easily
resolved. The three hydrogens on the benzene ring resonate
at d 7.28, 7.19 and 6.78, and those on C-1, C-2 and C-3 of
the 5-membered ring resonate at d 4.51, 4.38 and 3.84. In
the 13C NMR spectrum, the two quaternary carbons that
are part of the 5-membered ring of the indenyl moiety res-
onate at d 90.1 and 89.8. In contrast to the simple NMR
spectra of 8, those of compounds 6a–d are relatively more
complicated. In addition to the couplings between the
hydrogen and carbon atoms and phosphorus atom in both
1H and 13C spectra, the two identical substituents on the
phosphorus atoms show different hydrogen and carbon sig-
nals because of the chiral environment created by the ferr-
1
ocenyl moiety of the molecules. For example, in the H
NMR spectrum of 6b, the two methyl groups of the ethyl
groups that are attached to the phosphorus atom resonate
FeCp*
1. n-BuLi,THF
-78 oC
FeCp*
1. t-BuLi,Et2O,-78 oC
2. Cl-PR2,-78 oC
71-74%
6a, R = Ph
6b, R = Et
6c, R = i-Pr
6d, R = Cy
2. Cp*-Li,FeCl2
THF,rt
PR2
7
Br
Br
8
63%
Scheme 1. Synthesis of monodentate phosphorus ligands 6a–d.

1958
M. Thimmaiah et al. / Journal of Organometallic Chemistry 692 (2007) 1956–1962
at d 1.28 and 0.84 as two doublets of triplets due to cou-
pling to the phosphorus atom and the methylene groups;
in the 13C NMR spectrum, the corresponding two carbons
of the methyl groups show two doublets at d 10.7
(J = 66.8 Hz) and 9.5 (J = 30.4 Hz). To further confirm
the structure of this class of compound, 6a was chosen
for X-ray crystallography analysis. The ORTEP-3 drawing
[7] of 6a in Fig. 2 shows that one phenyl group on the phos-
phorus atom points to the exo face of the indenyl ring and
the other one adopts a position that is opposite to the 5-
membered ring of the indenyl moiety. This conformation
of the two phenyl groups presumably reduces the steric hin-
drance between them and the ferrocenyl moiety of the com-
pound. As a result of such an orientation of the two phenyl
groups, the lone pair on the phosphorus atom is pointed to
the endo face of the indenyl plane, which is significant for
controlling enantioselectivity of catalytic reactions. Fig. 2
also shows that the iron atom is coordinated to the indene
in an g5 fashion, and the distances between the iron and
the five carbon atoms of the 5-membered ring are not iden-
tical as the ring is tilted upwards slightly. The carbon and
hydrogen atoms of the indenyl moiety are nearly planar
and the plane almost parallel to that of Cp* (angle between
planes is 4.42ꢁ). This rigid conformation of the benzoferro-
cene is ideal for using this class of compound as a chiral
ligand for asymmetric catalysis.
ilar conditions, the enantiomers of 6c and 6d were not well-
separated. To test the configurational stability of this class
of compound, solutions of enantiopure 6a and 6b in hex-
ane, ether and THF were stirred under nitrogen atmo-
sphere at room temperature and at reflux temperature for
12 h. In all cases, no racemization was observed as verified
by chiral HPLC analysis. In order to assign the absolute
configuration of the enantiomers, attempts were made to
grow crystals of the single enantiomers or their complexes
of palladium(II) for X-ray crystallography study but failed
at this time.
Enantiopure 6a–b and racemic 6c–d were then used for
palladium catalyzed allylic alkylation reactions between
allyl acetate 9 and dimethylmalonate (10) (Eq. (1)). The
catalysts were generated in situ by stirring [Pd(C3H5)Cl]2
and a ligand in CH2Cl2 at room temperature for 1 h. Enan-
tiopure 6a was first used to find out the optimum ratio of
ligand to palladium (L/Pd) for the reaction. At an L/Pd
ratio of 1.25/1.0, under previously established allylic alkyl-
ation reaction conditions (see Section 5) [8], complete con-
version was realized within 24 h at room temperature as
indicated by GC–MS. Aqueous workup and flash column
chromatography afforded pure 11 in 91% yield (entry 1,
Table 1). The ee of the product, determined by chiral
HPLC (see Section 5 for details), was 26%; and when 6a1
was used, the configuration of the major enantiomer of
the product was S. If the L/Pd ratio was increased to
2.5/1.0, a 100% conversion was reached within 3 h with
an isolated yield of 87%. The ee of the product was
improved to 39% (entry 2). Further increases in the L/Pd
ratio to 3.0/1.0 and 4.0/1.0 slowed down the reaction and
both the yield and the ee were reduced (entries 3 and 4).
As a result, when other ligands (6b–d) were used for the
reaction, the L/Pd ratio was set to 2.5/1.0. As shown in
Table 1, the reaction was highly efficient with 6b as ligand,
within 30 min, 100% conversion was realized and an excel-
lent yield (93%) was obtained, but the ee remained moder-
ate (51%, entry 5). When ligands 6c and 6d were used, the
reactions were found to be very slow (entries 6 and 7) prob-
ably due to the bulky ligands preventing the substrates 9
and/or 10 from approaching the palladium catalytic center.
Ligands 6a and 6b were readily resolved by chiral HPLC
using a semi-preparative CHIRALPAK AD-H column
eluting with the solvent mixture of 0.25% 2-propanol in
hexane at a flow rate of 7 mL/min. The two enantiomers
of 6a were eluted at 9.7 min (6a1) and 12.3 min (6a2) and
those of 6b at 8.3 min (6b1) and 9.1 min (6b2). Under sim-
O
O
OAc
O
O
[PdCl(C3H5)]2,Ligand
MeO
Ph
OMe
+
Ph
Ph
MeO
OMe BSA,KOAc,CH2Cl2,rt
Ph
9
10
11
ð1Þ
3. Discussion
Fig. 2. ORTEP-3 drawing [7] of 6a with selected atom numbering.
Thermo ellipsoids are drawn at the 30% probability level. Hydrogen atoms
are represented by spheres of arbitrary radii. Selected bond distances and
Benzoferrocene derivatives containing donor atoms
such as phosphorus, nitrogen, sulfur and oxygen are poten-
tial useful chiral ligands for catalytic asymmetric reactions.
However, to our knowledge, there is no report on the syn-
thesis of this type of ligand even in racemic form, not to
mention their application in transition metal catalyzed
reactions. The lack of such publications may be attributed
˚
˚
˚
angles: Fe1–C1, 2.046(9) A; Fe1–C2, 2.024(9) A; Fe1–C3, 2.050(9) A;
˚
˚
˚
Fe1–C8, 2.094(7) A; Fe1–C9, 2.094(8) A; Fe1–C10, 2.063(8) A; Fe1–C11,
˚
˚
˚
˚
˚
˚
2.044(8) A, Fe1–C12, 2.036(9) A; Fe1–C13, 2.028(8) A; Fe1–C14,
˚
2.045(8) A, P1–C4, 1.849(9) A; P1–C111, 1.824(9) A; P1–C121,
1.826(9) A; C4–P1–C111, 103.6(4)ꢁ; C4–P1–C121, 101.3(4)ꢁ; C111–P1–
C121, 101.9(4) ꢁ.

M. Thimmaiah et al. / Journal of Organometallic Chemistry 692 (2007) 1956–1962
1959
Table 1
Application of ligands 6a–d in allylic alkylation reactions
Entry
Ligand
Ligand/Pd
Time (h)
Conversiona
%
Yieldb
%
eec % (config.d)
1
2
3
4
5
6
7
a
6a1
6a1
6a2
6a1
6b2
6c
1.25/1.0
2.5/1.0
3.0/1.0
4.0/1.0
2.5/1.0
2.5/1.0
2.5/1.0
24
3
100
100
80
50
100
56
91
87
75
47
93
23
16
26 (S)
39 (S)
27 (R)
26 (S)
51 (R)
–
44
44
0.5
43
48
6d
20
–
Determined by GC.
Isolated yield.
Determined by analytical chiral HPLC.
b
c
d
Assigned from the optical rotation by comparison with literature data [9].
to the perceived chemical and configurational instability of
this type of compound and perhaps their demonstrated
incompatibility with transition metal catalyzed reactions.
Indeed, it is reasonable to believe that benzoferrocene
derivatives are less stable than ferrocene because energy
is required to partially break the aromaticity of the benzene
ring in the starting indene during its formation. Although
Fu and co-workers reported the preparation of a number
of ferrocene derivatives that are fused to a pyridine ring
(pyroferrocene), this does not imply that the benzoferro-
cene derivatives will be equally stable because pyridine is
electron deficient and allows for electron back donation
from the iron center to its p* orbital. Even in those cases,
oxidation of Fe(II) to Fe(III) followed by decomposition
was observed during the isolation of the product in the
reactions for preparing pyroferrocene and its derivatives
[10]. In order to use benzoferrocene derivatives as chiral
ligands, one or more electron donating atoms such as phos-
phorus and oxygen atoms must be attached to the benzene
ring; this will increase the electron density on the indene
ring and the iron center, making the iron center possibly
more susceptible to oxidation. During our preparation
and characterization of the phosphorus ligands 6a–d,
although these compounds are indeed less stable than fer-
rocene, they were easily handled and were stable at room
temperature for at least one year in capped vials; these
observations negate any concern over their chemical stabil-
ity and their use as chiral ligands for asymmetric catalysis.
In terms of configurational stability, our concerns on
using benzoferrocene derivatives as chiral ligands arose
from several reports by the Curnow’s research group [11].
Here, the authors documented facile meso to rac isomeriza-
tion of bis(1-(diphenylphosphino)-g5-indenyl)iron(II) in
THF through ring-flipping of the indenyl ligands via a
phosphorus atom coordinated or less likely an sp2 carbon
atom coordinated g1-intermediate. If such a process also
pertained with our ligands 6a–d, this project of developing
a new type of benzoferrocene ligand for enantioselective
catalysis would have to be reconsidered. Fortunately, our
results showed that enantiopure 6a and 6b were configura-
tionally stable both in solid and in solution states even at
elevated temperature for an extended period of time. The
third concern on using benzoferrocene ligands in catalysis
is their stability in the presence of other transition metals.
The iron(II) center of the less strongly coordinated benzof-
errocenyl compounds (compared with ferrocene) may par-
ticipate in redox reactions with transition metals such as
palladium required for the catalysis. If such redox commu-
nication between Fe(II) and Pd(II) occurred and the result-
ing cationic benzoferrocenium complexes were not stable,
the benzoferrocene ligands would not be able to control
the stereochemistry of the catalytic reactions. In order to
address this concern, the enantiopure 6a–b and the racemic
6c–d were used in palladium catalyzed allylic alkylation
reaction between allyl acetate 9 and dimethylmalonate
(10). Gratifyingly, when enantiopure 6a and 6b were used
as ligands, excellent chemical yields were obtained
although the enantioselectivities were moderate and these
indicated that this new type of chiral ligand was compatible
with palladium catalyzed reaction conditions. We did
anticipate moderate enantioselectivity because in the enan-
tio-determination step of allylic alkylation reactions the
chiral ligand and the incoming nucleophile are on different
sides of the intermediate p-allyl complex. In addition, the
monodentate nature of ligands 6a and 6b may result in dif-
ferent catalyst conformations, which further contribute to
the moderate enantioselectivity. Compared with known
monodentate ligands, 6a and 6b gave lower ees for allylic
alkylation. For example, Nelson’s group used monodentate
phosphine ligands based on chiral g6-Cr[arene] templates
to catalyze allylic alkylation, up to 90% ee was obtained
[12]; Hamada’s group used 9-PBN, up to 98% ee was
obtained [13]; Tsarev’s group used the P-chiral monoden-
tate diamidophophites giving up to 97% ee [14] and Bur-
gess’ group used C3-symmetric triarylphosphines giving
up to 82% ee [15]. According to these reports, monodentate
ligands generally gave much lower ees than bidentate
ligands in allylic alkylation reactions. It is reasonable to
expect that derivatization of 6a and 6b to bidentate ligands
should increase the enantioselectivity of the alkylation
reaction. The slower rates of reaction when 6c and 6d were
used as ligands may be attributed to the difficulty for the
substrates 9 and/or 10 in approaching the catalytic palla-
dium center due to the bulkiness of these ligands.

1960
M. Thimmaiah et al. / Journal of Organometallic Chemistry 692 (2007) 1956–1962
4. Conclusions
10% K2CO3 (25 mL) and water (25 mL). After drying over
anhydrous Na2SO4, the deep purple solution was concen-
trated to about 15 mL. Cooling the solution to ꢀ20 ꢁC
gave the pure product 8 as deep purple crystals (6.2 g,
63%). The mother liquor contained more product but
was contaminated with Cp*–H and 7, which resulted from
oxidation of product 8, further purification by recrystalli-
zation could not afford pure 8; purification by flash column
chromatography (SiO2, hexane) also failed because Cp*–H,
7 and 8 had very similar Rf values, all were highly non-
polar and had very short retention time on the column:
Rf = 0.70 (hexane); mp 126–8 ꢁC; 1H NMR (CDCl3,
400 MHz) d 7.28 (d, 1H, J = 8.4 Hz, H-5 or 7), 7.19 (d,
1H, J = 6.8 Hz, H-5 or 7), 6.78 (dd, 1H, J = 8.4, 6.8 Hz,
H-6), 4.51 (dt, 1H, J = 2.8, 0.8 Hz, H-1 or 3), 4.38 (dd,
1H, J = 2.4, 1.2 Hz, H-1 or 3), 3.84 (t, 1H, J = 2.4 Hz,
H-2), 1.66 (s, 15H, Cp*–CH3); 13C NMR (CDCl3,
400 MHz) d 127.1 (C-5, 6 or 7), 124.3 (C-5, 6 or 7), 123.1
(C-4), 122.5 (C-5, 6 or 7), 90.1 (C-30 or 70), 89.8 (C-30 or
70), 78.5 (C-2), 76.6 (Cp*–C), 67.2 (C-2 or 3), 66.7 (C-2
or 3), 10.2 (Cp*–CH3); EI-MS calcd for C19H21BrFe
(m/z) [M+] 384; found 384; FAB-HRMS (NBA) calcd for
C19H21BrFe (m/z) [M+] 384.0176; found 384.0180.
In conclusion, we have shown that new benzoferrocenyl
phosphorus chiral ligands 6a–d, the simplest of a new fam-
ily of ligand, can be prepared conveniently in good yields.
These ligands are chemically and configurationally stable
and are compatible with palladium catalyzed allylic alkyl-
ation reactions. Asymmetric syntheses of this new type of
ligand, derivatizing them to produce multidentate ligands
and using them to solve challenging problems in asymmet-
ric catalysis are ongoing.
5. Experimental
5.1. General procedures
All reactions were performed under a nitrogen atmo-
sphere using standard Schlenk techniques. Reagents and
solvents available from commercial sources were used as
received unless otherwise noted. THF and Et2O were dis-
tilled from Na/benzophenone ketyl. CH2Cl2 was distilled
over CaH2. Thin layer chromatography (TLC) was per-
formed using Sigma–Aldrich TLC plates, silica gel 60F-
254 over glass support, 0.25 lm thickness. Flash column
chromatography was performed using Selecto Scientific sil-
5.3. 4-(Diphenylphosphino)indenyl-
pentamethylcyclopentadienyliron (6a)
1
ica gel, particle size 32–63. H, 13C and 31P NMR were
measured on a Varian UNITYINOVA spectrometer at
400 Hz; chemical shifts (d) were reported to residue CHCl3
The deep purple solution of 8 (1.0 g, 2.6 mmol) in Et2O
(25 mL) in a round-bottomed flask was cooled to ꢀ78 ꢁC.
To this solution was added t-BuLi (1.7 M in pentane,
3.3 mL, 5.7 mmol) or n-BuLi (2.0 M in pentane, 2.8 mL,
5.7 mmol) slowly via syringe and the mixture was stirred
at the same temperature for 30 min giving a deep red solu-
tion. P-Chlorodiphenylphosphine (1.1 mL, 5.7 mmol) was
next added dropwise via syringe and the color of the solu-
tion returned to deep purple immediately. After stirring at
ꢀ78 ꢁC for 1 h, the reaction was quenched with water, and
the mixture partitioned between 10% K2CO3 (25 mL) and
ether (50 mL) and the isolated organic layer further washed
with 10% K2CO3 (25 mL) and water (25 mL). After drying
over anhydrous Na2SO4, the deep purple solution was con-
centrated to dryness. Flash column chromatography (SiO2,
pretreated with 5% Et3N in hexane; hexane) gave product
6a as deep purple crystals (0.92 g, 72%): Rf = 0.25 (hex-
ane); mp 159–61 ꢁC; 1H NMR (CDCl3, 400 MHz) d
7.40–7.37 (m, 5 H, Ph), 7.31 (dt, 1H, J = 8.8, 0.8 Hz, H-
7), 7.27–7.22 (m, 5H, Ph), 6.85 (ddd, 1H, J = 8.4, 6.8,
1.2 Hz, H-6), 6.42 (ddd, 1H, JHH,HP,HH = 6.4, 5.2,
0.8 Hz, H-5), 4.35–4.33 (m, 1H, H-1 or 3), 4.30 (dt, 1H,
J = 2.8, 1.2 Hz, H-1 or 3), 3.72 (t, 1H, J = 2.8 Hz, H-2),
1.71 (s, 15H, Cp*–CH3); 13C NMR (CDCl3, 400 MHz) d
136.6 (d, JCP = 30.4 Hz, C · 2, Ph), 135.0 (Ph), 134.8
(Ph), 134.3 (Ph), 134.1 (Ph), 129.1 (C-5, 6 or 7), 128.7
(Ph), 128.6 (d, JCP = 54.4 Hz, C-4), 128.6 (Ph), 127.3 (C-
5, 6 or 7), 122.1 (C-5, 6 or 7), 91.0 (d, JCP = 97.2 Hz, C-
30), 87.9 (C-70), 78.2 (C-2), 76.5 (Cp*–C), 66.3 (d,
JCP = 42.8 Hz, C-3), 66.1 (C-1), 10.5 (Cp*–CH3); 31P
1
(d = 7.27 ppm for H and 77.23 ppm for 13C) and H3PO4
(d = 0 ppm). GC–MS were measured on GCMS-
QP5050A, Shimadzu; column, DB-5MS, 0.25 lm thick-
ness, 0.25 mm diameter, 25 m length; MS, positive EI.
Melting points were determined using a MEL-TEMPꢂ
melting point apparatus and are uncorrected. HPLC was
performed on a JASCO LC-2000Plus System, Pump PU-
2080Plus, Detector UV-2075Plus.
5.2. 4-Bromoindenyl-pentamethylcyclopentadienyliron (8)
7-Bromo-1(H)-indene [6] (7, 5.0 g, 25.7 mmol) in THF
(150 mL) was cooled to ꢀ78 ꢁC, n-BuLi (2.0 M in pentane,
12.8 mL, 25.7 mmol) was added via syringe along the wall
of flask in a dropwise manner. After addition, the solution
was warmed to rt slowly. Pentamethylcyclopentadiene
(Cp*–H, 3.7 mL, 25.7 mmol) in THF (100 mL) in another
round-bottomed flask was treated with n-BuLi (2.0 M in
pentane, 12.8 mL, 25.7 mmol) at rt giving a white suspen-
sion. This white suspension was added to a suspension of
anhydrous FeCl2 (3.3 g, 25.7 mmol) in THF (100 mL) with
vigorous stirring via a cannula. After stirring at rt for 1 h,
the solution of lithiated 7 was added dropwise via a can-
nula, and the resulting deep purple solution was stirred at
rt. After 2 h, TLC (SiO2, hexane) indicated quantitative
conversion. THF was removed under reduced pressure on
a rotary evaporator, the deep purple residue was parti-
tioned between 10% K2CO3 (100 mL) and hexane
(200 mL) and the organic layer was further washed with

M. Thimmaiah et al. / Journal of Organometallic Chemistry 692 (2007) 1956–1962
1961
NMR (CDCl3, 400 MHz) d ꢀ11.25; EI-MS calcd for
C31H31FeP (m/z) [M+] 490, found 490; Anal. Calc. for
C31H31FeP: C, 75.92; H, 6.37. Found: C, 76.03; H,
6.24%. The two enantiomers of 6a were resolved with
semi-preparative chiral HPLC using a CHIRALPAK
AD-H column eluting with the solvent mixture of 0.25%
2-propanol in hexane at a flow rate of 7 mL/min. The
two enantiomers were eluted at 9.7 min (6a1) and
(d, JCP = 60.8 Hz, PCH), 21.4 (d, JCP = 66.8 Hz, PCH),
20.4 (d, JCP = 42.4 Hz, CH3), 20.1 (d, JCP = 42.8 Hz,
CH3), 19.7 (d, JCP = 73.2 Hz, CH3), 18.2 (CH3), 10.5
(Cp*–CH3); 31P NMR (CDCl3, 400 MHz) d ꢀ1.61; EI-
MS calcd for C25H35FeP (m/z) [M+] 422, found 422;
FAB-HRMS (NBA) calcd for C25H35FeP (m/z) [M+]
422.1826; found 422.1829.
25
12.3 min (6a2, ½aꢁD ꢀ1895.5, c 0.045, hexane), respectively.
5.6. 4-(Dicyclohexylphosphino)indenyl-
pentamethylcyclopentadienyliron (6d)
In enantiopure form, the two enantiomers are deep purple
sticky solids.
Following the procedure for the preparation of 6a, com-
pound 6d was synthesized in 74% yield as a deep purple
solid: Rf = 0.30 (hexane); mp 83–6 ꢁC; H NMR (C6D6,
5.4. 4-(Diethylphosphino)indenyl-
pentamethylcyclopentadienyliron (6b)
1
400 MHz) d 7.20 (d, 1H, J = 8.4 Hz, H-7), 7.02 (dd, 1H,
J = 6.4, 3.2 Hz, H-5), 6.88 (dd, 1H, J = 8.8, 6.8 Hz, H-6),
4.95 (brs, 1H, H-1 or 3), 4.15–4.1.3 (m, 1H, H-1 or 3),
3.71 (t, 1H, J = 2.4 Hz, H-2), 2.27–2.18 (m, 2H,
PCH · 2), 0.59–1.99 (m, 35H, two cyclohexyls (Cy) and
Cp*–CH3); 13C NMR (C6D6, 400 MHz) d 134.8 (d,
JCP = 90.8 Hz, C-4), 125.59 (C-5, 6 or 7), 125.56 (C-5, 6
or 7), 121.5 (C-5, 6 or 7), 94.6 (d, JCP = 103.2 Hz, C-30),
88.5 (d, JCP = 30.4 Hz, C-70), 78.0 (C-2), 75.8 (Cp*–C),
66.0 (d, JCP = 48.4 Hz, C-3), 65.7 (C-1), 35.5 (d,
JCP = 66.8 Hz, Cy), 31.2 (d, JCP = 48.4 Hz, Cy), 30.3 (d,
JCP = 54.8 Hz, Cy), 29.5 (d, JCP = 60.8 Hz, Cy), 28.9 (d,
JCP = 12.4 Hz, Cy), 28.1 (d, JCP = 42.4 Hz, Cy), 27.8 (d,
JCP = 18.4 Hz, Cy), 27.1 (d, JCP = 42.4 Hz, Cy), 26.8
(Cy), 26.3 (Cy), 22.9 (Cy), 14.2 (Cy), 10.5 (Cp*–CH3); 31P
NMR (C6D6, 400 MHz) d ꢀ10.30; EI-MS calcd for
C31H43FeP (m/z) [M+] 502, found 502; FAB-HRMS
(NBA) calcd for C31H43FeP (m/z) [M+] 502.2452; found
502.2449.
Following the procedure for the preparation of 6a,
compound 6b was synthesized in 71% yield as a deep pur-
ple sticky solid: Rf = 0.25 (hexane); 1H NMR (CDCl3,
400 MHz) d 7.27 (d, 1H, J = 7.2 Hz, H-7), 6.99–6.93 (m,
2H, H-5 and 6), 4.65 (brs, 1H, H-1 or 3), 4.31 (brs, 1H,
H-1 or 3), 3.81 (t, 1H, J = 2.4 Hz, H-2), 1.97–1.88 (m,
1H, PCH2), 1.83–1.73 (m, 1H, PCH2), 1.65 (s, 15H,
Cp*–CH3), 1.63–1.55 (m, 2H, PCH2), 1.28 (dt, 3H,
JHP,HH = 16.4, 7.6 Hz, CH3), 0.84 (dt, 3H, JHP,HH = 12.0,
7.6 Hz, CH3); 13C NMR (CDCl3, 400 MHz) d 137.9 (d,
JCP = 72.8 Hz, C-4), 127.7 (C-5, 6 or 7), 123.5 (C-5, 6
or 7), 121.9 (C-5, 6 or 7), 91.8 (d, JCP = 90.8 Hz, C-30),
88.1 (C-70), 78.1 (C-2), 76.1 (Cp*–C), 66.1 (C-1), 65.2
(d, JCP = 42.8 Hz, C-3), 20.2 (d, JCP = 54.8 Hz, PCH2),
15.5 (d, JCP = 42.8 Hz, PCH2), 10.7 (d, JCP = 66.8 Hz,
CH3), 10.4 (Cp*–CH3), 9.5 (d, JCP = 30.4 Hz, CH3); 31P
NMR (CDCl3, 400 MHz) d ꢀ 24.53; EI-MS calcd for
C23H31FeP (m/z) [M+] 394, found 394; FAB-HRMS
(NBA) calcd for C23H31PFe (m/z) [M+] 394.1513; found
394.1510. The two enantiomers were resolved using chiral
HPLC under the same conditions used for resolution of
5.7. General procedure for palladium catalyzed allylic
alkylation reactions
25
6a, and were eluted at 8.3 min (6b1, ½aꢁD þ2061.1, c
0.035, hexane) and 9.1 min (6b2), respectively. In enantio-
pure form, the two enantiomers are deep purple sticky
solids.
The pre-catalyst [Pd(C3H5)Cl]2 (0.02 mmol) and a ligand
with a specific ligand to palladium ratio were combined in a
two-necked round-bottomed flask and flushed with nitro-
gen. Freshly distilled CH2Cl2 (2 mL) was then added via
a syringe and the mixture was stirred at rt for 1 h. Allyl ace-
tate 9 (1 mmol) in CH2Cl2 (2 mL) was then added via a
cannula, and dimethylmalonate (10, 3.0 mmol) and BSA
(3.0 mmol) were added via syringes, and finally KOAc
(0.5 mmol) was added under positive nitrogen pressure.
The reaction mixture was stirred at rt and monitored by
GC–MS. Aqueous workup (Et2O/4% HCl) and flash col-
umn chromatography (SiO2, hexane/ether 8:1) gave pure
product 11 with isolated yields ranging from 16% to 91%.
Enantiomeric excess was determined by analytical chiral
HPLC using CHIRALPAK AD-H column eluting with
2-propanol at a flow rate of 0.6 mL/min by integration of
the two peaks with retention times of 13 and 21 min. Con-
figuration of the major enantiomer of the product 11 was
assigned from the optical rotation by comparison with lit-
erature data [9].
5.5. 4-(Diisopropylphosphino)indenyl-
pentamethylcyclopentadienyliron (6c)
Following the procedure for the preparation of 6a,
compound 6c was synthesized in 78% yield as a deep
purple solid: Rf = 0.19 (hexane); mp 68–74 ꢁC; 1H
NMR (CDCl3, 400 MHz) d 7.30 (d, 1H, J = 8.4 Hz, H-
7), 6.98 (brs, 1H, H-5), 6.94 (t, 1H, J = 7.6 Hz, H-6),
4.77 (brs, 1H, H-1 or 3), 4.30 (brs, 1H, H-1 or 3), 3.79
(brs, 1H, H-2), 2.41–2.28 (m, 1H, PCH), 1.92–1.79 (m,
1H, PCH), 1.66 (s, 15H, Cp*–CH3), 1.31–1.21 (m, 6H,
CH3), 0.88–0.84 (m, 6H, CH3); 13C NMR (CDCl3,
400 MHz) d 134.2 (d, JCP = 91.2 Hz, C-4), 127.7 (C-5,
6 or 7), 126.0 (C-5, 6 or 7), 121.6 (C-5, 6 or 7), 94.3
(d, JCP = 97.2 Hz, C-30), 88.3 (C-70), 78.3 (C-2), 75.6
(Cp*–C), 65.8 (d, JCP = 48.4 Hz, C-3), 65.5 (C-1), 25.3

1962
M. Thimmaiah et al. / Journal of Organometallic Chemistry 692 (2007) 1956–1962
Table 2
html, or from the Cambridge Crystallographic Data Cen-
tre, 12 Union Road, Cambridge CB2 1EZ, UK; fax:
(+44) 1223-336-033; or e-mail: deposit@ccdc.cam.ac.uk.
Supplementary data associated with this article can be
found, in the online version, at doi:10.1016/j.jorganchem.
2007.01.004.
Summary of crystallographic data for 6a.
6a
Formula
C31H31FeP
490.38
Monoclinic
P 21/c
Formula weight (g molꢀ1
Crystal system
)
Space group
˚
a (A)
11.646(1)
16.568(6)
13.493(2)
90
92.61(1)
90
2600.8(10)
4
1.252
References
˚
b (A)
˚
c (A)
[1] (a) R. Noyori, Asymmetric Catalysis in Organic Synthesis, Wiley-
Interscience, New York, 1994;
a (ꢁ)
b (ꢁ)
c (ꢁ)
(b) B.M. Trost, Proc. Natl. Acad. Sci. USA 101 (2004) 5348–5355;
(c) E.N. Jacobsen, A. Pfaltz, H. Yamamoto, Comprehensive Asym-
metric Catalysis, Springer, Berlin, 1999;
(d) I. Ojima, Catalytic Asymmetric Synthesis, second ed., Wiley
VCH, New York, 2000;
(e) H.-U. Blaser, E. Schmidt, Asymmetric Catalysis on Industrial
Scale: Challenges, Approaches, and Solutions, Wiley-VCH, Wein-
heim, 2004.
3
˚
V (A )
Z
dcalc (g cmꢀ3
)
l (mmꢀ1
h (ꢁ)
)
0.658
1.75–22.48
0.71073
293(2)
˚
k (A)
T (K)
[2] (a) L.-X. Dai, T. Tu, S.-L. You, W.-P. Deng, X.-L. Hou, Acc. Chem.
Res. 36 (2003) 659–667;
GOF
R1a, wR2 (I > 2r(I))b
1.01
0.073, 0.202c
P
P
(b) T.J. Colacot, Chem. Rev. 103 (2003) 3101–3118;
(c) R.C.J. Atkinson, V.C. Gibson, N.J. Long, Chem. Soc. Rev. 33
(2004) 313–328;
a
R1 ¼ kF oj ꢀ jF ck= jF oj.
P
P
2
2 1=2
b
c
wR2 ¼ ½ ½wðF 2o ꢀ F c2Þ ꢁ= ½wðF 2oÞ ꢁꢁ
.
2
2
2
2
2
w ¼ 1=½ ðF oÞ þ ð0:1653PÞ ꢁ where P ¼ ðF o þ 2ðF cÞ Þ=3.
(d) U. Siemeling, T.C. Auch, Chem. Soc. Rev. 34 (2005) 584–594.
[3] R. Shintani, G.C. Fu, Org. Lett. 4 (2002) 3699–3702.
[4] T.C. Maier, G.C. Fu, J. Am. Chem. Soc. 128 (2006) 4594–4595.
[5] G.C. Fu, Acc. Chem. Res. 37 (2004) 542–547.
[6] (a) M. Adamczyk, D.S. Watt, D.A. Netzel, J. Org. Chem. 49 (1984)
4226–4237;
5.8. X-ray crystallography
A suitable crystal of compound 6a (obtained by crystal-
lization from hexane at ꢀ20 ꢁC) was rolled in epoxy resin
and mounted on a glass fiber. An Enraf–Nonius CAD-4
X-Ray diffractometer was the instrument used for the
determination. The windows program WINGX was used as
the interface for the solution and refinement of the model
[16]. The data were first reduced and corrected for absorp-
tion using psi-scans [17], and then solved using the program
SIR97 [18]. The model was refined using the program
SHELXL97 [19]. All non-hydrogens atoms were refined with
anisotropic thermal parameters and the hydrogen atoms
were refined at calculated positions with thermal parame-
ters constrained to the carbon atom on which they were
attached. Crystal data and final structural refinement
parameters are listed in Table 2.
(b) W.E. Bondinell, F.W. Chapin, G.R. Girard, C. Kaiser, A.J.
Krog, A.M. Pavloff, M.S. Schwartz, J.S. Silvestri, P.D. Vaidya, B.L.
Lam, G.R. Wellman, R.G. Pendleton, J. Med. Chem. 23 (1980) 506–
511;
(c) F.M. Hauser, S. Prasanna, Synthesis-Stuttgart (1980) 621–623;
(d) F. Ishikawa, H. Yamaguchi, J. Saegusa, K. Inamura, T. Mimura,
T. Nishi, K. Sakuma, S. Ashida, Chem. Pharm. Bull. 33 (1985) 3336–
3348.
[7] L.J. Farrugia, J. Appl. Cryst. 30 (1997) 565.
[8] B.M. Trost, M.L. Crawley, Chem. Rev. 103 (2003) 2921–2943, and
references cited therein.
[9] T. Hayashi, A. Yamamoto, T. Hagihara, Y. Ito, Tetrahedron Lett.
27 (1986) 191–194.
[10] J.C. Ruble, G.C. Fu, J. Org. Chem. 61 (1996) 7230–7231.
[11] (a) O.J. Curnow, G.M. Fern, M.L. Hamilton, A. Zahl, R. Van
Eldik, Organometallics 23 (2004) 906–912;
(b) O.J. Curnow, G.M. Fern, Organometallics 21 (2002) 2827–
2829;
(c) O.J. Curnow, G.M. Fern, M.L. Hamilton, E.M. Jenkins,
J. Organomet. Chem. 689 (2004) 1897–1910.
[12] S.G. Nelson, M.A. Hilfiker, Org. Lett. 1 (1999) 1379–1382.
[13] Y. Hamada, N. Seto, Y. Takayanagi, T. Nakano, O. Hara,
Tetrahedron Lett. 40 (1999) 7791–7794.
[14] V.N. Tsarev, S.E. Lyubimov, A.A. Shiryaev, S.V. Zheglov, O.G.
Bondarev, V.A. Davankov, A.A. Kabro, S.K. Moiseev, V.N. Kalinin,
K.N. Gavrilov, Eur. J. Org. Chem. (2004) 2214–2222.
[15] M.T. Powell, A.M. Porte, J. Reibenspies, K. Burgess, Tetrahedron
57 (2001) 5027–5038.
[16] L.J. Farrugia, J. Appl. Crystallogr. 32 (1999) 837–838.
[17] A.C.T. North, D.C. Phillips, F.S. Mathews, Acta Crystallogr. A 24
(1968) 351–359.
[18] A. Altomare, M.C. Burla, M. Camalli, G. Cascarano, C. Giacovazzo,
A. Guagliardi, A.G.G. Moliterni, G. Polidori, R. Spagna, J. Appl.
Crystallogr. 32 (1999) 115–119.
Acknowledgements
We thank the Department of Chemistry, Michigan
Technological University for financial support and Dr.
Dallas K. Bates for helpful discussions. The assistance
from Mr. Jerry L. Lutz (NMR), Mr. Shane Crist (compu-
tation) and Mr. Dean W. Seppala (electronics) are grate-
fully acknowledged. The authors also thank NSF for an
equipment grant (CHE-9512445).
Appendix A. Supplementary material
CCDC 624026 contains the supplementary crystallo-
graphic data for 6a. These data can be obtained free
of charge via http://www.ccdc.cam.ac.uk/conts/retrieving.
[19] G.M. Sheldrick, SHELX-97, Programs for Crystal Structure Analysis
(Release 97-2), University of Go¨ttingen, Germany, 1998.