Preparation of planar chiral ferrocenes containing a fluorous alcohol function

Article abstract of DOI:10.1016/j.ica.2005.05.040

Ferrocene reacts with hexafluoroacetone trihydrate in refluxing octane to afford >80% yields of [CpFe(η⁵-C₅H ₄C(CF₃)₂OH)] (X-ray), carrying out the reactions at 180°C gives an additional 5% yield of [Fe(η⁵- C₅H₄C(CF₃)₂OH)₂] (X-ray).The mono alcohol is lithiated with Bu^tOK/Bu ⁿLi/TMEDA affording partial conversion to mixtures of [CpFe(1,2-η⁵-C₅H₃C(CF₃) ₂OH)(X)] and [Fe(η⁵-C₅H₄X)(1,2- η⁵-C₅H₃C(CF₃)₂OH)(X)] (X = SMe, CPh₂OH) upon reaction with Me₂S₂ or OCPh₂.For X = CPh₂OH both structures are crystallographically characterised.Enantiopure [CpFe(1,2-η⁵- C₅H₃C(CF₃)₂OH)(SMe)] can be prepared from (R)-[CpFe(η⁵-C₅H₄S(O)C ₆H₄Me)] via [CpFe(1,2-η⁵-C ₅H₃S(O)C₆H₄Me)(C(CF ₃)₂OH)] (X-ray) or [CpFe(1,2-η⁵-C ₅H₃S(O)C₆H₄Me)(SMe)].Related procedures allow the preparation of [CpFe(1,2-η⁵-C ₅H₃CPh₂OH)(Y)] (Y = SMe, CHO (X-ray), C(CF ₃)₂OH) and[CpFe(1,2-η⁵-C₅H ₃C(CF₃)₂OH)(CHO)].

Full text of DOI:10.1016/j.ica.2005.05.040

Inorganica Chimica Acta 359 (2006) 1731–1742
www.elsevier.com/locate/ica
Preparation of planar chiral ferrocenes containing a fluorous
alcohol function
*
Victoria Albrow, Alexander J. Blake, Alexandre Chapron, Claire Wilson, Simon Woodward
School of Chemistry, The University of Nottingham, University Park, Nottingham NG7 2RD, United Kingdom
Received 27 April 2005; accepted 30 May 2005
Available online 21 November 2005
Dedicated to Professor Gerard van Koten.
Abstract
Ferrocene reacts with hexafluoroacetone trihydrate in refluxing octane to afford >80% yields of [CpFe(g5-C5H4C(CF3)2OH)] (X-ray),
carrying out the reactions at 180 ꢀC gives an additional 5% yield of [Fe(g5-C5H4C(CF3)2OH)2] (X-ray). The mono alcohol is lithiated with
ButOK/BunLi/TMEDA affording partial conversion to mixtures of [CpFe(1,2-g5-C5H3C(CF3)2OH)(X)] and [Fe(g5-C5H4X)(1,2-g5-
C5H3C(CF3)2OH)(X)] (X = SMe, CPh2OH) upon reaction with Me2S2 or O@CPh2.For X = CPh2OH both structures are crystallograph-
ically characterised. Enantiopure [CpFe(1,2-g5-C5H3C(CF3)2OH)(SMe)] can be prepared from (R)-[CpFe(g5-C5H4S(O)C6H4Me)] via
[CpFe(1,2-g5-C5H3S(O)C6H4Me)(C(CF3)2OH)] (X-ray) or [CpFe(1,2-g5-C5H3S(O)C6H4Me)(SMe)]. Related procedures allow the prep-
aration of [CpFe(1,2-g5-C5H3CPh2OH)(Y)] (Y = SMe, CHO (X-ray), C(CF3)2OH) and [CpFe(1,2-g5- C5H3C(CF3)2OH)(CHO)].
ꢁ 2005 Elsevier B.V. All rights reserved.
Keywords: Ferrocenyl ligands; Hydrogen bonding; Lithiation; Crystal structure
1. Introduction
of Kagan [2], we became interested in the preparation of
ferrocenes of type A whereby the chirality is displayed
solely due to the presence of dissimilar donor ligands
(L1, L2, Scheme 1) and particularly in the case where
L1 is a fluorinated alcohol derivative. Fluorinated alco-
hols have been little exploited as potential ligands in
asymmetric catalysis but are proving increasingly popular
solvent choices for some reactions [3]. The choice of such
solvents is driven by the unusual properties they can
offer including: strong protic effects (the pKa of hexaflu-
oroisopropanol is 9.3 compared to ca. 17.1 for PriOH),
strong hydrogen bond donor ability and stability towards
oxidation. Including such a substituent directly into a
metal ligand system might offer opportunities to exploit
similar behaviour but with only catalytic amounts of
the fluorous alcohol substituent. This paper describes
our investigations into fluorous ferrocenes containing
the C(CF3)2OH unit and comparison to species where
the fluorous substituent is replaced by CPh2OH.
The strongly displayed planar chirality of 1,2-disubsti-
tuted ferrocenes and their ready accessibility from the
low cost parent make them ideal candidates as useful
ligand additives in asymmetric catalysis. This utility has
been most elegantly demonstrated in the JOSIPHOS-pro-
moted Metolachlor process whereby a single iridium-
JOSIPHOS complex manages to produce over 800000
target molecules with a stereoselectivity in excess of
80% [1]. However, most of the current generation of fer-
rocene-based ligands (including JOSIPHOS) rely on the
asymmetric introduction of an initial chiral centre as a
route to subsequent planar chirality. Following the work
*
Corresponding author. Tel.: +44 115 9513541; fax: +44 115 9513564.
E-mail address: simon.woodward@nottingham.ac.uk (S. Woodward).
0020-1693/$ - see front matter ꢁ 2005 Elsevier B.V. All rights reserved.
doi:10.1016/j.ica.2005.05.040

1732
V. Albrow et al. / Inorganica Chimica Acta 359 (2006) 1731–1742
L1
L2
Rc PCy2
PPh2
NMe2
Rc
Fe R
p
FeR
Fe
p
Fe
(Rp)-A L2 > L1
Direct synthesis of planar chirality
(Rp,Rc)-JOSIPHOS
Induction of planar chirality
vs.
Scheme 1. Preparation of 1,2-disubstituted ferrocenes. Throughout this manuscript the subscripts ꢁpꢀ and ꢁcꢀ are used to indicate planar and
centrosymmetric chirality elements, respectively.
2. Experimental
Dichloromethane (Fisher) was freshly distilled under nitro-
gen or argon atmospheres from calcium hydride. Ferrocene
(Aldrich) was used as received. Solutions of [NBu4][BF4]
were prepared by the literature methods.
2.1. General
All reactions involving air sensitive reagents were car-
ried out under argon or nitrogen atmospheres using stan-
dard Schlenk techniques and anhydrous solvents. The
spectrometers and other instrumentation used were as pre-
viously described [4]. Fluorine nmr shifts were referenced
to CFCl3. Polarimetry studies were carried out in a ther-
mally equilibrated Jasco DIP370 polarimeter at a nominal
ambient temperature of 27 ꢀC (effective variation 25–
31 ꢀC). Hexafluoroacetone gas (TOXIC!, 97%) was sup-
plied by Aldrich and hexafluoroacetone trihydrate (98%)
by Lancaster; the former was handled in gas-tight syringes.
Column chromatography and TLC analyses were per-
formed on silica gel, Davisil and Merck Silica gel 60
2.2. Preparation of the complexes
2.2.1. 1-(1-Hydroxy-1-hexafluoromethylethyl)ferrocene
[CpFe(g5-C5H4(C(CF3)2OH))] (1a)
To a suspension of ferrocene (5.50 g, 29.6 mmol) in
octane (2 ml) was added hexafluoroacetone trihydrate
(10 ml, 71.8 mmol) and the reaction mixture heated at
reflux (bath temperature 120 ꢀC) for 5 days. The mixture
was allowed to cool to room temperature, dichloromethane
(20 ml) and water (10 ml) were added and the phases sepa-
rated. To the separated aqueous phase was added zinc dust
until the solution decolourised and then extracted with
dichloromethane. The combined organic phases were dried
(MgSO4) and then concentrated under reduced pressure.
The residue was purified by column chromatography on
silica gel (hexane to elute ferrocene then diethyl ether to
elute the product) to give the product (8.44 g, 81%) as an
orange crystalline solid. The spectroscopic properties were
identical to that of 1a prepared by the literature method [7].
M.p. 91–93 ꢀC (lit. [7] 99–101 ꢀC). Anal. Calc. for
C13H10F6FeO: C, 44.35; H, 2.86. Found: C, 44.44; H,
2.82%. mmax(CHCl3)/cmꢀ1: 3508 (br) m(OH), 3102 (w)
F254 + 366, respectively. Petrol refers to the fraction with
b.p. 40–60 ꢀC Sulfoxide 5 was prepared in both racemic
and either (Rc) or (Sc) configuration by the methodology
of Kagan [2]. The planar chirality of compounds 3 and 6
have been assigned using Schlo¨glꢀs revised procedure;
apparent inversions of the stereochemical descriptors are
a consequence of the CIP priority rules [5,6]. For conve-
nience the structures in Schemes 2 and 3 show products
that would result from the use of sulfoxide (Rc)-5 regardless
of the route (chiral or racemic) used to prepare them.
F3C
F3C
CF3
CF3
OH
i
Fe
OR
HO
+
Fe
Fe
>80%
F3C
CF3
ii
2
1a R = H
5% (at 180 oC)
86%
1b R = CH2OMe
F3C
F3C
F3C
CF3
CF3
CF3
OH
iii
OH
OH
Fe
E
E
Fe
+
Fe
see Table 1
E
1a
3a E = SMe
3b E = CPh2OH
4a E = SMe
4b E = CPh2OH
Scheme 2. Reagents and conditions: (i), O@C(CF3)2 Æ 3H2O, n-octane reflux, 110 ꢀC, 5 days or neat reagents, sealed tube 180 ꢀC, 4 days; (ii) KH followed
by MeOCH2Cl/NaI; (iii), ButOK/BunLi/TMEDA (1:1:1), for conditions see Table 2.

V. Albrow et al. / Inorganica Chimica Acta 359 (2006) 1731–1742
1733
O
S
O
S
E2
i
ii
E1
4-Tol
4-Tol
E1
Fe
Fe
Fe
(Rp-3a E1 = SMe; E2 = C(CF3)2OH
(Sp)-3b E1 = CPh2OH; E2 = C(CF3)2OH
(Sp)-3c E1 = CPh2OH; E2 = SMe
(Rp)-3d E1 = CPh2OH; E2 = CHO
(Rp)-3e E1 = C(CF3)2OH; E2 = CHO
(Sp,Rc)-6a E = SMe
(Sp,Rc)-6b E = C(CF3)2OH
(Sp,Rc-6c E = CPh2OH
(Rc)-5
(ex. ref. 2)
Scheme 3. Reagents and conditions: (i), LDA (1.1 eq.), THF, ꢀ78 ꢀC, 40 min, then either Me2S2, O@C(CF3)2 or O@CPh2 for 6a–c, respectively; (ii), ButLi
(1.1–2.1 eq.), THF, ꢀ78 ꢀC, 30 min, then O@C(CF3)2 for 3a–b, Me2S2 for 3c or DMF followed by HCl (aq.) for 3d–e.
m(CH), 1296 m (CF), 1149, 1110 m(CO), 964; dH (400 MHz,
CDCl3) 3.36 (1H, s, OH), 4.32 (5H, s, C5H5), 4.40 (2H, app
t, J 2.0, C5H4), 4.48 (2H, s, br, C5H4); dC (101 MHz,
CDCl3) 67.8, br, 69.4, 69.4, 75.4 (septet, JCF 30), 82.6,
122.4 (q, JCF 288); dF (282 MHz, CDCl3) ꢀ74.91 (s, br);
m/z (ES) 353 (M+H+, 15%), 352 (M+, 100%), 335
(M+ꢀOH, 10%); [m/z (ES) found [M]+ 351.9966;
C13H10Fe1F6O1 requires 351.9985].
2 M HCl until the solution turned cloudy. The aqueous
phase was then repeatedly extracted with dichloromethane
(50 ml portions), dried (MgSO4) and concentrated under
reduced pressure to give yellow solid (2.27 g). Recrystallisa-
tion from heptane gave golden needles (202 mg, 5%) of 2.
The remaining yellow solid (2.05 g, 81%) attained on evap-
oration of the mother liquors, was identified as 1-(1-
hydroxy-1-hexafluoromethylethyl)ferrocene (1a). M.p.
200–201 ꢀC. Anal. Calc. for C16H10F12FeO2: C, 37.09; H,
2.2.2. 1-(1-Methoxymethylether-1-hexafluoromethylethyl)
ferrocene [CpFe(g5-C5H4(C(CF3)2OCH2OMe))] (1b)
In flame-dried glassware under argon was added oil free
potassium hydride (248 mg, 6.20 mmol), tetrahydrofuran
(7 ml) and 1-(1-hydroxy-1-hexafluoromethylethyl)ferro-
cene 1a (730 mg, 2.07 mmol). After 5 min chloromethyl-
methylether (TOXIC!) (0.20 ml, 2.69 mmol) and sodium
iodide (155 mg, 50 mol%) were added. After stirring at
ambient temperature (3 h) the mixture was poured onto
ice. The layers were separated and the organic layer was
washed with KOH (aq), dried over NaSO4 and evaporated
to dryness to give an orange oil (819 mg, 100%). Purifica-
tion by chromatography on silica gel (6:1 petrol:diethyl
ether) gave an orange oil (707 mg, 86%). Anal. Calc. for
C15H14F6FeO2: C, 45.48; H, 3.56. Found: C, 45.39; H,
3.41%. mmax(CHCl3)/cmꢀ1: 2906 m(CH), 1304 m(CF), 1288
m(CF), 1156, 1142, 1092 m(CO), 1040, 986, 955; dH
(400 MHz, CDCl3) 3.42 (3H, s, OMe), 4.27 (5H, s,
C5H5), 4.41 (4H, s, C5H4), 4.83 (2H, s, CH2OMe); dC
(101 MHz, CDCl3) 56.4, 68.8 (app t, JCF, J ꢁ2), 69.2,
70.0, 75.0, 81.4 (septet, JCF 29), 92.8, 122.5 (q, JCF 290);
dF(282 MHz, CDCl3) ꢀ71.86 (s); m/z (ES) 397 (M+H+,
20%), 396 (M+, 100%), 335 (M+ꢀOCH2OMe, 25%); [m/z
(ES) found [M]+ 396.0261; C15H14FeF6O2 requires
396.0247].
1.95. Found: C, 36.84; H, 1.95%. mmax(CHCl3)/cmꢀ1
:
3697 m(OH), 3484 (br) m(OH), 1602, 1291 m(CF), 1150,
1104 m(CO), 1053, 965; dH (500 MHz, CDCl3) 4.47 (4H,
s, C5H4), 4.53 (4H, s, C5H4), 4.63 (2H, s, OH); dC
(67.9 MHz, d4-MeOH) 69.1, 70.3, 81.8, 122.8 (q, JCF 288)
the C(CF3)2 carbon was not apparent as its highly coupled
nature and the low solubility of 2 prevented the attainment
of adequate signal-to-noise levels even in extended runs; dF
(282 MHz, CDCl3) ꢀ75.30 (s); m/z (ES) 519 (M+H+,
15%), 518 (M+, 100%), 352 (M+ꢀHOC(CF3)2, 40%); [m/z
(ES) found [M]+ 517.9866; C16H10FeF12O2 requires
517.9838].
2.2.4. ( )-1-(Methylsulfanyl)-2-(1-hydroxy-1-hexafluoro
methylethyl)ferrocene [CpFe(g5-C5H3(SMe)(C(CF3)2-
OH))] (3a) and ( )-1,10-(methylsulfanyl)-2-(1-hydroxy-
1-hexafluoromethylethyl) ferrocene [Fe(g5-C5H4SMe)-
(g5-C5H3(SMe)(C(CF3)2OH))] (4a)
The result of Table 2, run 7 is representative. A solution
of BunK ꢁsuperbaseꢀ was prepared by the method of Brand-
sma [8]. A suspension of ButOK (97 mg, 0.85 mmol) and
hexane (0.70 ml) was cooled to ꢀ60 ꢀC (cryostat). To the
solution was added BunLi (0.37 ml, 0.85 mmol) and
TMEDA (0.13 ml, 0.85 mmol). The mixture was stirred
for 5 min before the addition of alcohol 1a (100 mg,
0.28 mmol) and additional stirring for 2 h at this tempera-
ture before adding methyl disulfide (0.10 ml, 1.14 mmol).
An instant colour change from dark to light orange was
seen and the reaction was then treated with excess NH4Cl
(aq). The organic phase was filtered through Na2SO4 and
evaporated to dryness. Purification by chromatography
on silica gel (19:1 petrol:diethyl ether) gave ( )-3a as an
orange crystalline solid (Rf 0.33, 9:1 petrol:diethyl ether)
(33.4 mg, 30%). M.p. 47–49 ꢀC. Anal. Calc. for
C14H12F6FeOS: C, 42.23; H, 3.04. Found: C, 42.20; H,
2.2.3. 1,10-Di-(1-hydroxy-1-hexafluoromethylethyl)
ferrocene [Fe(g5-C5H4(C(CF3)2OH)2)] (2)
Ferrocene (1.34 g, 7.18 mmol) and hexafluoroacetone
trihydrate (2 ml, 14.4 mmol) were charged into a Carius
tube, sealed under a nitrogen atmosphere and heated at
180 ꢀC for 4 days. After cooling, the tube was opened
and the product extracted in dichloromethane (50 ml).
The organic phase was extracted with NaOH (aq) (50 ml)
and resulting aqueous phase separated and acidified with

1734
V. Albrow et al. / Inorganica Chimica Acta 359 (2006) 1731–1742
3.00%. mmax(CHCl3)/cmꢀ1: 3273 (br) m(OH), 2925 m(CH),
1267 m(CF), 1150, 1123 m(CO), 1078, 954; dH (500 MHz,
CDCl3) 2.29 (3H, s, SMe), 4.35 (5H, s, C5H5), 4.41–4.43
(1H, m, C5H3), 4.47 (1H, app t, J 2.5, C5H3) 4.57 (1H,
dd, J 2.5, 1.5, C5H3) 6.84 (1H, br and unresolved JHF to
CF3, OH); dC (67.9 MHz, CDCl3) 22.0 (q, JCF 2.5), 69.7
(app septet, JCF 1.5), 70.5, 71.7 (6C), 76.3, 76.7, 78.6 (sep-
tet, JCF 30), 122.3 (q, JCF 288), 122.6 (q, JCF 288) one CH
in the substituted Cp is coincident with the unsubstituted
C5H5 signal; dF (282 MHz, CDCl3) ꢀ77.30 (q, JFF 10.5),
ꢀ73.86 (q, JFF 10.5); m/z (ES) 399 (M+H+, 15%), 398
(M+, 100%), 381 (M+ꢀOH, 10%); [m/z (ES) found
[M+H]+ 398.9910; C14H12FeF6OSH requires 398.9940].
Continued elution afforded ( )-4a (Rf 0.25, 9:1 pet-
rol:diethyl ether) (33.4 mg, 26%). Anal. Calc. for
C15H14F6FeOS2: C, 40.56; H, 3.18. Found: C, 40.92; H,
3.06%. mmax(CHCl3)/cmꢀ1: 3272 (br) m(OH), 2931 m(CH),
1286 m(CF), 1272 m(CF), 1150, 1123 m(CO), 1078, 954; dH
(500 MHz, CDCl3) 2.30, (3H, s, SMe), 2.32 (3H, s, SMe),
4.35–4.40 (2H, M, Cp), 4.40–4.42 (1H, m, Cp), 4.43 (1H,
dd, J 2.5, 1.5, Cp), 4.44–4.48 (1H, m, Cp) overlapping with
4.46 (1H, dd, J 5.5, 2.5, Cp), 4.54 (1H, dd, J 2.5, 1.5, Cp),
6.73 (1H, d, br, J ꢁ1, OH); dC (67.9 MHz, CDCl3) 19.1,
21.7 (q, JCF 2.5), 70.3 (app septet, JCF 1.5), 71.8, 71.8,
72.4, 72.8, 73.0, 73.0, 74.2, 78.3 (septet, JCF 30), 79.8,
87.7, 122.3 (q, JCF 289), 122.6 (q, JCF 289); dF (282 MHz,
CDCl3) ꢀ77.26 (q, JFF 10.5), ꢀ73.89 (q, JFF 10.5); m/z
(ES) 445 (M+H+, 15%), 444 (M+, 100%), 398 (MꢀSMe+,
5%); [m/z (ES) found [M]+ 443.9703; C15H14FeF6OS2
requires 443.9740].
through a plug of MgSO4 and evaporated to dryness to
give an orange oil. Purification by chromatography on sil-
ica gel (9:1 petrol:diethyl ether) gave (Sp)-3a as an orange
oil that solidified on standing (259 mg, 64%). For (Sp)-3a:
[a]D +166 (c 1.05 in CHCl3, 27 ꢀC). Other properties were
the same as racemic 3a. The conditions used for the chiral
HPLC analysis of the ee of 3a were hexane 100%, flow rate
0.5 ml minꢀ1, kdet = 258 nm (Sp)-3a 20.1 min, (Rp)-3a
22.1 min.
2.2.6. ( )-1-(1-Hydroxy-1-hexafluoromethylethyl)-2-
(1-hydroxydiphenylmethyl)ferrocene [CpFe(g5-C5H3-
(C(CF3)2OH)(CPh2OH))] (3b) and ( )-1-(1-hydroxy-
1-hexafluoromethylethyl)-2,10-(1-hydroxydiphenyl
methyl)ferrocene [Fe(g5-C5H4(CPh2OH)(g5-C5H3-
(C(CF3)2OH)(CPh2OH)))] (4b)
Prepared in an identical manner to ( )-3a/4a from 1a
(100 mg, 0.28 mmol) and benzophenone (208 mg,
1.14 mmol) as a solution in diethyl ether (0.7 ml). Purifica-
tion by chromatography on silica gel (9:1 petrol:diethyl
ether) gave ( )-3b as an orange crystalline solid (Rf 0.14,
4:1 petrol:diethyl ether) (26.4 mg, 17%). M.p. 217–220 ꢀC;
m
max(CHCl3)/cmꢀ1: 3474 (br) m(OH), 3234 (br) m(OH),
2882 (w) m(CH), 1600 m(C@C), 1494 m(C@C), 1450
m(C@C), 1274 m(CF), 1146, 1134 m(CO), 1108 m(CO),
1005, 953; dH (400 MHz, CDCl3) 4.08 (1H, s, br, C5H4),
4.22 (5H, s, C5H5), 4.32 (1H, s, OH), 4.50 (1H, app t, J
2.5, C5H4), 4.58 (1H, s, br, C5H4), 7.10–7.15 (2H, m,
2 · Ph-o), 7.16–7.27 (3H, m, Ph-m + p), 7.38–7.34 (1H,
m, Ph-p), 7.39 (2H, app t, J 7.5, 2 · Ph-m), 7.42–7.47
(2H, m, 2 · Ph-o), 7.60 (1H, s, OH); dC (68 MHz, CDCl3)
69.1, 70.9, 71.1 (app t, br, J 2.5), 72.7, 77.6, 79.7, 95.2,
122.4 (q, JCF 288), 123.0 (q, JCF 288), 126.6, 126.8, 127.5,
127.6, 127.6, 127.9, 128.4, 130.2, 132.5, 137.7, 144.4,
145.0 the C(CF3)2 carbon was not apparent as its highly
coupled nature prevented the attainment of adequate sig-
nal-to-noise levels even in extended runs; dF (282 MHz,
CDCl3) ꢀ76.17 (q, JFF 8.5), ꢀ75.68 (q, JFF 8.5); m/z
(ES) 557 (M+Na+, 100%), 534 (M+, 33%), 517
(M+ꢀOH, 100%); [m/z (ES) found [M+Na]+ 557.0570;
C26H20F6FeO2Na requires 557.0614].
Continued elution afforded ( )-4b (Rf 0.06, 4:1 pet-
rol:diethyl ether) (18.7 mg, 9%). M.p. 174–176 ꢀC with
decomposed; mmax(CHCl3)/cmꢀ1: 3598 (br) m(OH), 3480
(br) m(OH), 3223 (br) m(OH), 2925 (w) m(CH), 1600
m(C@C), 1492 m(C@C), 1448 m(C@C), 1330, 1273 m(CF),
1134 m(CO), 1108 m(CO), 1008, 954; dH (270 MHz, CDCl3)
2.25 (1H, s, OH), 3.86 (1H, s, br, C5H4 or C5H3), 4.01 (1H,
m, br, C5H4 or C5H3), 4.06 (1H, app t, J 3.0, C5H4 or
C5H3), 4.23 (1H, s, OH), 4.27 (1H, s, br, C5H4 or C5H3),
4.31 (1H, s, br, C5H4 or C5H3), 4.47 (1H, s, br, C5H4 or
C5H3), 4.53 (1H, s, br, C5H4 or C5H3), 7.01–7.15 (5H, m,
Ph), 7.15–43 (5H, m, Ph), 7.60 (1H, s, OH); dC
(67.9 MHz, CDCl3) 69.8, 70.6, 70.9, 71.5, 71.6, 71.9 (br),
72.0, 73.6, 76.9 (septet, JCF 30), 77.8, 79.8, 94.9, 98.8,
122.3 (q, JCF 290), 122.6 (q, JCF 290), 126.0, 126.4, 126.7,
126.7, 126.8, 127.2, 127.3, 127.6, 127.7, 127.8, 127.8,
2.2.5. (+)-(Sp)-1-(Methylsulfanyl)-2-(1-hydroxy-1-hexa-
fluoromethylethyl)ferrocene [CpFe(g5-C5H3(SMe)-
(C(CF3)2OH))] (3a)
To a solution of (Sp,Rc)-(6a) (73 mg, 0.20 mmol)
[derived from (Rc)-5] in THF (1 ml) cooled to ꢀ78 ꢀC
was added ButLi (0.13 ml, 0.22 mmol). The reaction stirred
for 30 min and the solution colour was seen to change from
pale orange to deep red. Hexafluoroacetone gas (10 cm3,
0.44 mmol) was added and the solution stirred for 15 min
in which the colour changed from deep red to yellow.
The reaction was quenched with water (1 ml) and allowed
to warm to room temperature. The organic phase was fil-
tered through a plug of silica and evaporated to dryness.
Purification by chromatography on silica gel (19:1 pet-
rol:diethyl ether) gave (Rp)-3a as an orange oil that solidi-
fied on standing (51 mg, 65%).
Alternatively (Sp,Rc)-(6b) (500 mg, 1.02 mmol) [derived
from (Rc)-5] in THF (10 ml) was treated with NaH (60%
in mineral oil, 54 mg, 1.22 mmol). The solution was stirred
for 30 min until the evolution of gas had ceased then cooled
to ꢀ78 ꢀC before addition of ButLi (0.76 ml, 1.22 mmol).
The solution instantly turned very deep red in colour. After
stirring for 40 min methyl disulfide (0.11 ml, 1.22 mmol)
was added to the reaction. After a further hour the reaction
was quenched with NH4Cl (aq) (3 ml) then allowed to
warm to room temperature. The organic phase was filtered

V. Albrow et al. / Inorganica Chimica Acta 359 (2006) 1731–1742
1735
127.8, 127.9, 144.5, 144.8, 146.9, 147.0; dF (282 MHz,
CDCl3) ꢀ76.28 (q, JFF 9.5), ꢀ75.70 (q, JFF 9.5); m/z
(FAB) 717 (M+H+, 40%), 716 (M+, 100%), 699
(M+ꢀOH, 15%); [m/z (FAB) found [M]+ 716.1470;
C39H30F6Fe1 requires 716.1449].
27 ꢀC). Racemic (3c) was prepared by an analogous proce-
dure but starting with ( )-(5).
2.2.9. One-pot method to (+)-(Sp)-1-(1-
hydroxydiphenylmethyl)-2-formyl ferrocene [CpFe(g5-
C5H3(CPh2OH)(CHO))] (3d)
2.2.7. (+)-(Rp)-1-(1-hydroxydiphenylmethyl)-2-(1-
hydroxy-1-hexafluoromethylethyl)ferrocene [CpFe(g5-
C5H3(CPh2OH)(C(CF3)2OH))] (3b)
A solution of (Sc)-5 (485 mg, 1.50 mmol) in THF (9 ml)
under argon was cooled to ꢀ78 ꢀC and pre-chilled lithium
diisopropylamide (0.90 ml of 1.8 M hexane/THF solution,
1.65 mmol) added. After 20 min benzophenone (300 mg,
1.65 mmol) was added. After a further 10 min ButLi
(1.94 ml of 1.7 M hexane solution, 3.3 mmol) was added
and the reaction stirred for 60 min. To the red solution was
added dried DMF (256 ll, 3.30 mmol) and after stirring
for 1 h at ꢀ30 ꢀC water (1 ml) was added. The layers were
separated and the organic phase was dried (MgSO4) and con-
centrated under reduced pressure to give an orange solid.
Purification by chromatography on silica gel (9:1 pet-
rol:diethyl ether) gave (Sp)-(3d) as a red solid (279 mg,
47%). M.p. 229–230 ꢀC. Anal. Calc. for C24H20FeO2: C,
72.74; H, 5.09. Found: C, 72.54; H, 5.13%. mmax(CHCl3)/
cmꢀ1: 3338 (br) m(OH), 2830 (w) m(CH), 1652 m(C@O),
1490 m(C@C), 1450 m(C@C), 1346, 1108 m(C–O) 1047; dH
(400 MHz, CDCl3) 4.20 (1H, m, C5H4), 4.47 (5H, s, C5H5),
4.69 (1H, app t, J 2.5, C5H4), 4.85 (1H, dd, J 2.5, 1.5,
C5H4), 6.75 (1H, s, OH), 7.18–7.32 (5H, m, Ph), 7.35–7.40
(1H, m, Ph-p), 7.40–7.47 (2H, m, 2 · Ph-m), 7.57–7.62 (2H,
m, 2 · Ph-o), 9.83 (1H, s, RC(O)H); dC (67.9 MHz, CDCl3)
71.0, 71.6, 74.0, 74.7, 77.0, 78.6, 101.8, 127.0 (2C), 127.3,
127.4, 127.6, 127.7, 145.5, 148.8, 196.3; m/z (ES) 419
(M+Na+, 90%), 396 (M+, 10%), 379 (M+ꢀOH, 100%);
[m/z (ES) found [M+Na]+ 419.0722; C24H20FeO2Na
requires 419.0710]. For (Sp)-(3d): [a]D +276 (c 1.00 in CHCl3,
18 ꢀC), +993 (c 1.00 in CHCl3, 27 ꢀC).
To a solution of (Sp,Rc)-(6b) (300 mg, 0.61 mmol) in
THF (6 ml) was added NaH (60% in mineral oil, 32 mg,
0.73 mmol). The solution was stirred for 30 min until the
evolution of gas had ceased then cooled to ꢀ78 ꢀC before
addition of ButLi (0.38 ml, 0.73 mmol). The solution
instantly turned very deep red in colour. After stirring for
40 min benzophenone (134 mg, 0.73 mmol) was added to
the reaction. After a further hour the reaction was
quenched with NH4Cl (aq) (2 ml) then allowed to warm
to room temperature. The organic phase was filtered
through a plug of MgSO4 and evaporated to dryness to
give an orange oil. Purification by chromatography on sil-
ica gel (4:1 petrol:diethyl ether) gave (Rp)-3b as an orange
crystalline solid (120 mg, 37%). Continued elution afforded
recovered (Sp,Rc)-6b (120 mg, 37%). For (Rp)-3b: [a]D +112
(c 1.02 in CHCl3, 27 ꢀC).
2.2.8. One-pot method to (ꢀ)-(Sp)-1-(methylsulfanyl)-2-
(1-hydroxydiphenylmethyl) ferrocene [CpFe(g5-
C5H3(SMe)(CPh2OH))] (3c)
A solution of (Sc)-5 (500 mg, 1.54 mmol) in THF (8 ml)
under argon was cooled to ꢀ78 ꢀC and pre-chilled lithium
diisopropylamide (0.94 ml of 1.8 M hexane/THF solution,
1.69 mmol). After 40 min methyl disulfide (156 ll,
1.69 mmol) was added. After a further 10 min ButLi
(2 ml of 1.7 M hexane solution, 3.40 mmol) was added
and the reaction stirred for 70 min. To the deep red solu-
tion was added solid benzophenone (616 mg, 3.40 mmol)
and after 5 min, when the solution had turned yellow,
water (1 ml) was added. The layers were separated and
the organic phase was dried (MgSO4) and concentrated
under reduced pressure to give an orange solid. Purification
by chromatography on silica gel (9:1 petrol:diethyl ether)
gave the product (Sp)-(3c) as a orange solid (364 mg,
57%). M.p. 145–147 ꢀC. Anal. Calc. for C24H22FeOS: C,
69.57; H, 5.35. Found: C, 69.76; H, 5.32%. mmax(CHCl3)/
cmꢀ1: 3428 (br) m(OH), 2924 m(CH), 1600, 1490 m(C@C),
1448 m(C@C), 1360, 1107 m(CO), 1033, 1002; dH
(400 MHz, CDCl3) 1.52 (3H, s, SMe), 3.67 (1H, dd, J
2.5, 1.5, C5H4), 4.17 (1H, app t, J 2.5, C5H4), 4.26 (5H, s
C5H5), 4.41 (1H, dd, J 2.5, 1.5, C5H4), 5.44 (1H, s, OH),
7.15–7.27 (6H, m, Ph), 7.29–7.34 (2H, m, 2 · Ph-p), 7.50–
7.54 (2H, m, 2 · Ph-o); dC (101 MHz, CDCl3) 20.4, 67.9,
70.8, 72.8, 75.6, 76.6, 78.9, 100.9, 127.2, 127.3, 127.7,
127.9, 127.9, 128.0, 145.1, 149.7; m/z (ES) 437 (M+Na+,
85%), 414 (M+, 90%), 397 (M+ꢀOH, 100%); [m/z (ES)
found [M+Na]+ 437.0616; C24H22FeOSNa requires
437.0638]. For (Sp)-(3c) [a]D ꢀ31 (c 1.00 in CHCl3,
2.2.10. (ꢀ)-(Rp)-1-(1-hydroxydiphenylmethyl)-2-formyl
ferrocene [CpFe(g5-C5H3(C(CF3)2OH)(CHO))] (3e)
Solid (Sp,Rc)-(6b) (100 mg, 0.20 mmol) [prepared from
(Rc)-(5)] in THF (2 ml) was treated with NaH (60% in min-
eral oil, 11 mg, 0.24 mmol) at room temperature. The solu-
tion was stirred for 30 min until the evolution of gas had
ceased then cooled to ꢀ78 ꢀC before addition of ButLi
(0.13 ml, 0.24 mmol). The solution instantly turned very
deep red in colour. After stirring for 40 min DMF
(0.02 ml, 0.24 mmol) was added to the reaction and the col-
our turned green. After a further hour the reaction was
quenched with NH4Cl (aq) (1 ml) then allowed to warm
to room temperature. The organic phase was filtered
through a plug of MgSO4 and evaporated to dryness. Puri-
fication by chromatography on silica gel (4:1 petrol:diethyl
ether) followed by recrystallisation (diethyl ether/petrol)
gave (Rp)-(3e) as red crystals (45 mg, 59%). M.p. 155–
156 ꢀC. mmax(CHCl3)/cmꢀ1: 3161 (br) m(OH), 2862 m(CH),
1646 m(C@O), 1279 m(CF), 1121 m(CO), 954; dH
(270 MHz, CDCl3); 4.46 (5H, s, C5H5), 4.81 (1H, dd, br,
J 3.0, 1.5, C5H3), 4.82 (1H, app t, J 3.0, C5H3), 5.02 (1H,
s, br, C5H3), 8.56 (1H, s, OH), 9.74 (1H, s, RC(O)H); dC

1736
V. Albrow et al. / Inorganica Chimica Acta 359 (2006) 1731–1742
(101 MHz, CDCl3) 72.3, 74.2, 74.4, 75.3, 84.1, 122.6 (q, JCF
290), 122.8 (q, JCF 290), 195.4 the C(CF3)2 carbon was not
apparent as its highly coupled nature prevented the attain-
ment of adequate signal-to-noise levels even in extended
runs; dF (282 MHz, CDCl3) ꢀ78.09 (q, JFF 10.5), ꢀ72.96
(q, JFF 10.5); m/z (FAB) 381 (M+H+, 20%), 380 (M+,
40%), 69 ðCF3þ; 54%Þ, 65 ðC5H5þ; 14%Þ; [m/z (ES) found
[M+H]+ 381.0027; C14H10F6Fe1O2H requires 381.0013].
For (ꢀ)-(Rp)-(3e): [a]D ꢀ902 (c 0.51 in CHCl3, 27 ꢀC).
ous phase extracted with diethyl ether (2 · 25 ml). The
combined organic phases were washed with brine (20 ml)
and dried (MgSO4) and concentrated under reduced pres-
sure to give an orange solid. Purification by chromatogra-
phy on silica gel (4:1 petrol:diethyl ether) gave (ꢀ)-(Sp,Rc)-
6b as a yellow foam (1.84 g, 81%). M.p. 114–115 ꢀC. Anal.
Calc. for C20H16F6FeO2S: C, 49.00; H, 3.29. Found: C,
48.90; H, 3.22%. mmax(CHCl3)/cmꢀ1: 3128 (br) m(OH),
2870 (br) m(CH), 1272 m(CF), 1150, 1134 m(CO), 1087,
1005, 955; dH (400 MHz, CDCl3) 2.36 (3H, s, p-Me), 4.59
(5H, s, C5H5), 4.62 (1H, s, br, C5H3), 4.70 (1H, app t, J
2.5, C5H3), 4.91 (1H, dd, J 2.5, 1.5, C5H3), 7.22 (2H, d, J
8.0, C6H4), 7.40 (2H, d, J 8.0, C6H4) 8.63, (1H, s, OH);
dC (100.6 MHz, CDCl3) 21.3, 71.5, 71.8, 72.1 (br), 72.5,
76.6 (septet, JCF 30), 81.1, 87.8, 122.1(q, JCF 290), 124.4
(q, JCF 290), 124.2, 129.6, 139.1, 141.1; dF (282 MHz,
CDCl3) ꢀ77.68 (q, JFF 10.5), ꢀ74.32 (q, JFF 10.5); m/z
(ES) 513 (M+Na+, 55%), 491 (M+H+, 100%), 352
(M+HꢀS(O)PhMe+, 10%); [m/z (ES) found [M+Na]+
512.9971; C20H16FeF6O2SNa requires 513.0022]. For (ꢀ)-
(Sp,Rc)-(6b): [a]D ꢀ146 (c 0.99 in CHCl3, 27 ꢀC). The con-
ditions used for the chiral HPLC analysis of the ee of 6b
2.2.11. (Rp,Rc)-1-(toluene-4-sulfinyl)-2-
(methylsulfanyl)ferrocene [CpFe(g5-C5H3(S(O)-4-
Tol)(SMe))] (6a)
An LDA solution (0.98 ml, 1.77 mmol in THF) was
added to a cooled (ꢀ78 ꢀC) solution of sulfoxide (Rc)-(5)
(500 mg, 1.61 mmol) in THF (10 ml). The reaction was stir-
red for 40 min before the addition of methyl disulfide
(163 ll, 1.77 mmol). The solution was stirred for 50 min
before being quenched with water (5 ml) and being allowed
to warm to room temperature. The organic phase was dried
(MgSO4) and concentrated under reduced pressure. The
residual oil was dissolved in diethyl ether (10 ml). Drop-
wise addition of petrol (10 ml) gave crude (Rp,Rc)-6a as
an orange crystalline solid (441 mg, 74%). Crude (Rp,Rc)-
were hexane:PriOH 95:5, flow rate 0.5 ml minꢀ1
det = 258 nm (Sp,Rc)-6b 20.7 min, (Rp,Sc)-6b 23.1 min.
,
k
1
6a (ꢁ90–92% pure by H NMR spectroscopy) could not
be separated from the (Rc)-(5) recovered from these reac-
tions, however, the crude material could be used directly
in subsequent syntheses with no effect on enantiomeric
purities and only marginal loss in yield for subsequent
2.2.13. (+)-(Rp,Sc)-1(1-hydroxydiphenylmethyl)-2-
(toluene-4-sulfinyl)ferrocene [CpFe(g5-
C5H3(CPh2OH)(S(O)-4-Tol))] (6c)
To a solution of (Sc)-5 (1.00 g, 3.08 mmol) in THF (20 ml)
cooled to ꢀ78 ꢀC was added LDA (1.96 ml, 3.54 mmol). The
reaction was stirred for 40 min before the addition of benzo-
phenone (655 mg, 3.6 mmol). The solution was stirred for
50 min before quenching with water (5 ml) and allowing to
warm to room temperature. The layers were separated and
the organic phase dried (MgSO4) and concentrated under
reduced pressure. Purification by chromatography on silica
gel (4:1 petrol:diethyl ether) gave (+)-(Rp,Sc)-6c as a yellow
foam (1.16g, 74%). M.p. 196–198 ꢀC. Anal. Calc. for
C30H26FeO2S: C, 71.15; H, 5.12. Found: C, 70.75; H,
5.25%. mmax(CHCl3)/cmꢀ1: 3252 (br) m(OH), 2996 (w)
m(CH), 1600 m(C@C), 1492 m(C@C), 1450 m(C@C), 1108
m(CO), 1049 m(S–O), 1010; dH (270 MHz, CDCl3) 2.26 (3H,
s, p-Me), 3.70 (1H, dd, J 2.5, 1.5, C5H3), 4.36 (1H, app t, J
2.5, 2.5, C5H3), 4.56 (5H, s, C5H5), 4.80 (1H, dd, J 2.5, 1.5,
products. For 91% pure (Rp,Rc)-6a: mmax(CHCl3)/cmꢀ1
:
2983 m(CH), 2924 m(CH), 1108, 1084, 1040 m(S–O), 1004;
dH (270 MHz, CDCl3) 2.32 (3H, s, p-Me), 2.46 (3H, s,
SMe), 3.88 (1H, dd, J 3.0, 1.5, C5H3), 4.23 (5H, s, C5H5),
4.30 (1H, app t, J 3.0, C5H3), 5.58 (1H, dd, J 3.0, 1.5,
C5H3), 7.38 (2H, d, J 8.5, C6H4), 7.75 (2H, d, J 8.5,
C6H4); dC (67.9 MHz, CDCl3) 21.6, 21.7, 69.4, 69.9, 71.1,
76.7, 84.3, 97.2, 125.6, 129.6, 140.0, 141.6; m/z (ES) 429
ðMþ MeCN þ NH4þ; 90%Þ, 371 (M+H+, 70%), 232
(M++Hꢀ S(O)tolyl, 60%); [m/z (ES) found [M+H]+
371.0203; C18H18Fe1OS2H requires 371.0227]. The remain-
ing mass balance was (Rc)-(5): dH (270 MHz, CDCl3) 2.38
(3H, s, p-Me), 4.32 (1H, m, C5H4), 4.36–4.38 (2H, m,
C5H4), 4.38 (5H, s, C5H5), 4.61 (1H, m, C5H4), 7.26 (2H,
d, J 8.0, C6H4), 7.53 (2H, d, J 8.0, C6H4); dC (67.9 MHz,
CDCl3) 21.5, 65.4, 68.0, 70.0 (6C), 70.1, 94.6, 124.4,
129.7, 143.1, 143.0.
0
C5H3), 6.70–6.93 (8H, m, Pha, H3,3 of C6H4 and OH),
0
7.15–7.45 (7H, m, Phb and H2,2 of C6H4); dC (67.9 MHz,
CDCl3) 21.3, 68.4, 71.1, 71.4 (6C), 75.3, 88.1, 98.9, 123.1,
125.9, 126.8, 127.2, 127.3 (2C), 127.9, 129.3, 139.6, 139.7,
146.1, 148.2; m/z (ES) 529 (M+Na+, 75%), 506 (M+, 15),
489 (M+ꢀOH, 100%); [m/z (ES) found [M+Na]+
529.0862; C30H26FeO2SNa requires 529.0900]. For (Rp,Sc)-
6c: [a]D +252 (c 1.00 in CHCl3, 27 ꢀC).
2.2.12. (ꢀ)-(Sp,Rc)-1-(toluene-4-sulfinyl)-2-(1-hydroxy-1-
hexafluoromethylethyl)ferrocene [CpFe(g5-C5H3(S(O)-4-
Tol)(C(CF3)2OH))] (6b)
To a solution of (Rc)-5 (1.50 g, 4.62 mmol) in THF
(30 ml) cooled to ꢀ78 ꢀC was added LDA (2.84 ml,
5.09 mmol in THF). The reaction was stirred for 30 min
before the addition of hexafluoroacetone gas (130 cm3,
5.78 mmol). The solution was stirred for 1 h 15 min before
being allowed to warm to room temperature and quench-
ing with water (10 ml). The layers were separated and aque-
2.3. Electrochemistry
Cyclic voltammetry (CV) was carried out using an Auto-
lab PGSTAT20 potentiostat under an atmosphere of argon

V. Albrow et al. / Inorganica Chimica Acta 359 (2006) 1731–1742
1737
using a three-electrode arrangement in a single compart-
ment cell. A glassy carbon working electrode, a Pt wire sec-
ondary electrode and a saturated calomel reference
electrode, chemically isolated from the test solution via a
bridge tube containing electrolyte solution and fitted with
a porous Vycor frit, were used in the cell. The solutions
were 10ꢀ3 M in test compound and 0.4þM in [NBu4][BF4]
molecules. All atoms were refined as described above except
two of the 4-Tol groups for which the carbon atoms were
refined with isotropic displacement parameters. Crystallo-
graphic data for all compounds are summarised in Table 1.
3. Results and discussion
ꢂ
ꢂ
as supporting electrolyte. The ½FeCp2 ꢃ =½FeCp2 ꢃ couple
was used as the internal reference to avoid overlapping
redox couples and the potential of the redox process was
referenced to that of the Fc+/Fc (E1/2 = 0.54 V versus
SCE) couple by an independent calibration. The DE1/2
3.1. Racemic complex preparation through lithiation of
fluorous alcohol 1a
Although the initial starting material (1a, R = H) had
been known since 1974 its literature preparation called
for reaction of gaseous O@C(CF3)2 with ferrocene in a
sealed tube at 120–180 ꢀC [7]. We find that its preparation
can be more conveniently carried out using the inexpensive
liquid trihydrate (HO)2C(CF3)2 Æ 2H2O (Scheme 2). Ther-
molysis, under otherwise identical conditions, affords excel-
lent (>80%) yields of 1a, which can be isolated by
chromatography, crystallisation, or acid base extraction
procedures due to facile deprotonation of 1a at pH values
above 10. It is noteworthy that oxidation of 1a to its ferroc-
inium cation did not occur under either strong acid or base
conditions and we suspected that the presence of the
strongly electron withdrawing group had a stabilising effect
on 1a. CV experiments indicated reversible electrochemical
behaviour for 1a with E1/2 = 190 mV versus Fc+/Fc in
dichloromethane indicating that 1a is significantly more
resistant to oxidation than ferrocene.
þ
+
ꢂ
ꢂ
between Fc /Fc and ½FeCp2 ꢃ =½FeCp2 ꢃ recorded under
identical conditions was 0.52 V. Redox potentials are
quoted versus the ferrocenium-ferrocene. Compensation
for internal resistance was not applied.
2.4. Molecular modelling
Calculations were carried out using Spartan [9] for
Macꢀ02 version 1.0.8 running on a IMac G4 1.25 GHz
768Mo RAM Mac with OS X 10.4.1 Lithiated 1a–b were
built as usual using the model kit and the energies mini-
mised using the default molecular mechanics option and
the file saved. From the calculation box (set up menu)
the ꢁequilibrium geometry at ground stateꢀ using the semi-
empirical PM3 level of theory was selected. Calculations
were started from the initial geometry, subject to symme-
try, requesting computation of the frequencies and thermo-
dynamics. Control calculations on genuine molecules, e.g.,
replication of the X-ray structure of 1a through calcula-
tion, indicated the parameterisation of PM3 is adequate
for the required simulations.
Use of temperatures above 180 ꢀC and long reaction times
(P4 days) led to small (ꢁ5%) additional yields of the disub-
stituted species 2, which had not been attained under the con-
ditions of Stone [7]. Isolation of 2 is facilitated by its low
solubility but attempts to increase its yield by use of longer
reaction times or by direct treatment of 1a with excess
(HO)2C(CF3)2 Æ 2H2O under high temperatures were not
successful. Because of the exceptionally electron-rich nature
of ferrocene we speculated that direct Friedel–Crafts substi-
tution should be possible without the use of sealed tube
experiments. In line with the findings of Masciadri on meth-
oxyarenes [10], direct reflux of ferrocene and (HO)2C(C-
F3)2 Æ 2H2O in n-octane lead to a >80% yield of 1a and an
added advantage of this approach is that, as no sealed tube
heating apparatus is required, the reaction can run on
multi-gram scales. Because of our interest in using the –
OH functions of fluorous alcohols of this type in hydro-
gen-bond donor roles, and to gain a better insight into the
steric demands of the bulky C(CF3)2OH group at the ferro-
cene compounds 1a and 2 were subjected to crystallographic
analysis. The molecular structures are shown in Figs. 1 and 2
together with selected hydrogen bond distances and angles.
The ferrocene cores of 1a and 2 are unremarkable, however,
in both compounds it is the OH function that is driven below
the ferrocenyl Cp-plane to accommodate the bulky trifluo-
romethyl groups. In 2, this results in a long intramolecular
2.5. Crystallography
Yellow-orange crystals of 1a, 3b–d, 4b and 6b were
grown from dichloromethane/petrol and 2 from heptane.
All single crystal diffraction data were collected using
graphite monochromated Mo Ka X-radiation on either a
Bruker SMART APEX (1a, 3c and 6b) or SMART1000
(2, 3b, 3d and 4b) CCD area detector diffractometer
equipped with an Oxford Cryostream cooling device. All
data were collected at 150 K. Details of the individual data
collections and refinements are given in Table 1. All struc-
tures were solved using direct methods using SIR-92 for 1a,
3b, 3d and SHELXS-97 for 2, 3c, 4b and 6b. All structures were
refined by full-matrix least-squares refinement against F2
using SHELXL-97 and all non-H atoms refined with aniso-
tropic atomic displacement parameters (adps). Hydrogen
atoms, except those of OH groups, were geometrically
placed and refined as part of a riding model. OH hydrogen
atoms were located in difference Fourier syntheses and
refined as rigid rotors. Data for 1a showed non-merohedral
twinning and data for both components were integrated
and used in the refinement with twin fraction refining to
0.3540(10). Compound 6b crystallised with six independent
˚
O(2)–Hꢄ ꢄ ꢄO(1) interaction (2.02 A).
One potential rapid route to racemic 1,2-disubstituted
ferrocenes containing a –C(CF3)2OH function is through

Table 1
Crystal data for the complexes
Complex
1a
2
3b
3c
3d
4b
6b
Formula
Mr
C
352.06
13H10F6FeO
C16H10F12FeO2
518.09
C26H20F6FeO2
534.27
C24H22FeOS
414.33
C24H20FeO2
396.25
C39H30F6FeO3 Æ CH2Cl2
801.41
C20H16F6FeO2S
490.24
Cell setting
Space group
monoclinic
P21/n
9.857 (2)
7.491 (2)
16.841 (4)
90.00
96.167 (4)
90.00
1236.3 (5)
4
orthorhombic
Fdd2
26.309 (2)
36.877 (2)
7.1358 (6)
90.00
90.00
90.00
6923 (2)
16
1.988
monoclinic
P21/c
10.2165 (5)
15.6461 (7)
14.2827 (7)
90.00
108.120 (1)
90.00
2169.84 (18)
4
monoclinic
P21
11.8528 (10)
10.6270 (9)
15.9857 (13)
90.00
96.340 (1)
90.00
2001.2 (3)
4
orthorhombic
P212121
8.6054 (14)
13.742 (2)
15.444 (2)
90.00
90.00
90.00
1826.3 (5)
4
1.441
triclinic
orthorhombic
P212121
8.8252 (8)
29.212 (2)
44.406 (4)
90.00
90.00
90.00
11447.9 (17)
24
1.707
ꢀ
P1
˚
a (A)
9.5774 (11)
11.2269 (12)
17.788 (2)
102.671 (2)
104.232 (2)
100.770 (2)
1749.2 (6)
2
˚
b (A)
˚
c (A)
a (ꢀ)
b (ꢀ)
c (ꢀ)
3
˚
V (A )
Z
Dx (Mg mꢀ3
)
1.891
1.29
150 (2)
1.635
0.77
150 (2)
1.375
0.87
150 (2)
1.522
0.66
150 (2)
l (mmꢀ1
T (K)
)
1.01
150 (2)
0.84
150 (2)
0.97
150 (2)
Crystal form
lath
tablet
block
block
plate
column
column
Colour
orange
yellow
orange
orange
orange
orange
orange
Crystal size (mm)
Tmin, Tmax
0.55 · 0.13 · 0.03
0.776, 1.000
7315, 3572, 2962
0.27 · 0.22 · 0.07
0.507, 0.613
7496, 2578, 2104
0.59 · 0.53 · 0.47
0.573, 0.700
19125, 5121, 4555
0.53 · 0.43 · 0.32
0.294, 0.760
11864, 6831, 6414
0.25 · 0.16 · 0.03
0.709, 1.000
11330, 4133, 3499
0.22 · 0.08 · 0.07
0.824, 0.905
15740, 7867, 4290
0.27 · 0.09 · 0.08
0.594, 1.00
77535, 20163, 13534
Number of measured,
independent and observed
(I > 2r(I)) reflections
Rint
R [F2 > 2r(F2)], wR(F2), S
Number of reflections
0.030
0.042, 0.095, 1.01
3572
0.034
0.029, 0.058, 0.92
2578
0.012
0.026, 0.071, 1.04
4933
0.017
0.044, 0.123, 1.04
6831
0.038
0.037, 0.079, 1.19
4112
0.051
0.044, 0.095, 0.87
7867
0.079
0.095, 0.221, 1.17
20163
Number of paramꢀet3ers
192
0.52, ꢀ0.35
283
318
0.38, ꢀ0.30
489
246
472
0.56, ꢀ0.68
1564
˚
Dqmax, Dqmin (e A
)
0.27, ꢀ0.20
0.92, ꢀ0.74
0.42, ꢀ0.22
0.87, ꢀ1.74
Flack parameter
0.00 (2)
ꢀ0.010 (15)
0.000 (16)
0.02 (3)

V. Albrow et al. / Inorganica Chimica Acta 359 (2006) 1731–1742
1739
Fig. 1. View of molecular structure of 1a, ellipsoids drawn at 30%
probability level, H atoms other than OH omitted. Hydrogen bond details,
˚
˚
O1–H1oꢄ ꢄ ꢄO2, d(Hꢄ ꢄ ꢄA) 1.77 A, d(Dꢄ ꢄ ꢄA) 2.584(3) A, \(DHA) 164ꢀ O2–
H2oꢄ ꢄ ꢄO3i, d(Hꢄ ꢄ ꢄA) 1.97 A, d(Dꢄ ꢄ ꢄA) 2.794(2) A, \(DHA) 165ꢀ,
˚
˚
i = x ꢀ 1, y, z.
Fig. 2. View of molecular structure of 2, ellipsoids drawn at 30%
probability level, H atoms other than OH omitted. Hydrogen bond details,
ortho lithiation of 1a directed off either the OH or CF func-
tions. Preliminary trials using n-, sec- or t-butyl lithium on
either 1a or its MOM protected analogue 1b failed to lead
to any appreciable lithiation (typical conditions in Table 2,
Runs 1 and 2). Semi-empirical PM3 calculations were car-
ried out to provide evidence as to whether a viable ortho
lithiation geometries can be attained for 1a–b considering
the large steric demand of the C(CF3)2OR (R = Li,
MOM) units (Fig. 3). These suggested that while lithiated
1a was viable (and potentially stabilised by a Liꢄ ꢄ ꢄF–C)
that ortho lithiation directed off of the –OLi function in
deprotonated 1b is a high energy process. The presence of
the OMOM group in 1b further exacerbates the congestion
at the ferrocenyl core, worse still the usual MOM chelation
coordination leads only to very high energy CF3ꢄ ꢄ ꢄFe con-
tacts. These calculations suggested to us that lithiation of
1a would only be viable through the use of highly reactive
bases (Schlosserꢀs ꢁsuper-baseꢀ) at low temperatures and
that lithiation of 1b is not viable (see Fig. 3).
i
˚
˚
O1–H1ꢄ ꢄ ꢄO2 , d(Hꢄ ꢄ ꢄA) 2.02 A, d(Dꢄ ꢄ ꢄA) 2.738(3) A, \(DHA) 142.7ꢀ O2–
˚
˚
H2ꢄ ꢄ ꢄ.O1, d(Hꢄ ꢄ ꢄA) 2.02 A, d(Dꢄ ꢄ ꢄA) 2.792(3) A, \(DHA) 152.1ꢀ,
i = ꢀx + 1, ꢀy, z.
The ꢁsuper-baseꢀ ButOK/BunLi/TMEDA mixture was
prepared in hexane at ꢀ60 ꢀC as described by Brandsma
[8]. The fluorous alcohol 1a was subjected to lithiation
under varying conditions and the intermediates quenched
with Me2S2 leading to recovered 1a, the desired mono func-
tionalised product 3a and difunctionalised 4a (Scheme 2,
Table 2). Control experiments indicated that reaction of
the intermediate anions with Me2S2, above ꢀ80 ꢀC, was
rapid. Additionally, it was found that little lithiation of
1a occurred below ꢀ60 ꢀC. Under optimal conditions for
the deprotonation (Run 6) a mono:disubstituted ratio of
3:1 was attained in 40% combined yield. Unfortunately,
good chemoselectivity for the monolithiation could not
be attained in combination with high conversions of 1a
Table 2
Lithiation of ferrocene 1a using the ꢁsuper-baseꢀ mixture ButOK/BunLi/TMEDA
Run
Base
Conditions
Electrophile
1a (%)a
Mono (%)a
Di (%)a
1
2
3
4
5
6
7
8
a
BunLi/TMEDA in Et2O
ButLi in THF on 1b
2.2 eq., 0 ꢀC, 5 h
Me2S2 ꢀ60 ꢀC
Me2S2 ꢀ60 ꢀC
Me2S2 ꢀ60 ꢀC
Me2S2 ꢀ60 ꢀC
Me2S2 ꢀ60 ꢀC
Me2S2 ꢀ60 ꢀC
Me2S2 ꢀ60 ꢀC
O@CPh2 ꢀ60 ꢀC
>90
>90b
61
56
57
36
32
ꢁ54
0
0
0
0
1.4 eq., ꢀ78 ꢀC, 2 h
2.0 eq., ꢀ60 ꢀC, 2 h
2.0 eq., ꢀ60 ꢀC, 5 h
2.0 eq., ꢀ60 ꢀC, 8 h
2.4 eq., ꢀ60 ꢀC, 16 h
3.0 eq., ꢀ60 ꢀC, 2 h
3.0 eq., ꢀ60 ꢀC, 2 h
ButOK/BunLi/TMEDAc
ButOK/BunLi/TMEDA
ButOK/BunLi/TMEDA
ButOK/BunLi/TMEDA
ButOK/BunLi/TMEDA
ButOK/BunLi/TMEDA
18 (3a)
21 (3a)
22 (3a)
30 (3a)
30 (3a)
17 (3b)
6 (4a)
10 (4a)
6 (4a)
10 (4a)
26 (4a)
9 (4b)
Isolated yields of chromatographically separated 1a, mono (3a–b) and disubstituted (4a–b) products.
Recovered 1b.
Reaction in n-hexane.
b
c

1740
V. Albrow et al. / Inorganica Chimica Acta 359 (2006) 1731–1742
and 5 along with selected hydrogen bond distances and
angles. The hydrogen bonding motifs observed provide
pointers to the locations of potential metal binding sites
in the utilisation of these ligands in asymmetric catalytic
chemistry. Compound 4b co-crystallises with a molecule
of CH2Cl2 but this is not associated with any hydrogen
bonding to 4b in the solid state.
3.2. Enantiopure complex preparation through lithiation of
sulfoxide 5
The sulfoxide 5 (shown in the Rc configuration at sulfur
for convenience in Scheme 3) is a versatile unit, introduced
by Kagan [2], for the preparation of 1,2-planar chiral fer-
rocenes. Deprotonation of (Rc)-5 with LDA proceeds with
very high enantioselectivity to the Rp (R planar) lithio spe-
cies leading to the essentially stereospecific formation of the
1,2-ferrocenyl sulfoxides upon reaction with electrophiles
(E1) to afford the geometry shown in Scheme 3. Subsequent
cleavage of the sulfoxide with ButLi allows stereospecific
introduction of a second electrophilic group (E2). Because
of the CIP assignment procedure for planar chirality [5,6]
the nature on the substituents often causes inversion of
the stereochemical descriptor (Rp) versus (Sp) but this is
of no significance regarding the reaction mechanism.
As Kaganꢀs studies [2] had concentrated on the inclusion
of non-protic donor groups (E1,E2), especially phosphines,
we were keen to extend its use to O@C(CF3)2 allowing
chemo- and stereospecific preparation of ligands (3). Expo-
Fig. 3. Calculated (PM3) products of the lithiation of 1a–b. Coordination
of the OCH2OMe group of the MOM to lithium is unattainable as this
generates very high energy Feꢄ ꢄ ꢄCF3 contacts. Selected bond distances, for
1
B: Liꢄ ꢄ ꢄF1 3.808, Li, Feꢄ ꢄ ꢄO1 2.924, Feꢄ ꢄ ꢄLiO1 2.990 A; for C: Liꢄ ꢄ ꢄF
˚
3.508, Liꢄ ꢄ ꢄF2 2.254, Feꢄ ꢄ ꢄO1 3.535, Feꢄ ꢄ ꢄO2 3.862 A.
˚
and mixtures were always isolated. The chemistry could be
extended to trapping reactions with benzophenone, leading
to isolable 3b and 4b. However, reaction of the anion mix-
ture from 1a with Br2 or I2 led to a chromatographically
inseparable mixture of mono and dihalides and recovered
1a.
To confirm the regiochemistry of the lithiation crystallo-
graphic studies were carried out on ferrocenes 3b and 4b.
The resulting molecular structures are shown in Figs. 4
Fig. 5. View of molecular structure of 4b, ellipsoids drawn at 30%
probability level, H atoms other than OH omitted. Hydrogen bond details,
˚
˚
Fig. 4. View of molecular structure of 3b, ellipsoids drawn at 30%
probability level, H atoms other than OH omitted. Hydrogen bond details,
O1–H1oꢄ ꢄ ꢄO2, d(Hꢄ ꢄ ꢄA) 1.77 A, d(Dꢄ ꢄ ꢄA) 2.584(3) A, \(DHA) 164ꢀ O2–
H2oꢄ ꢄ ꢄ.O3i, d(Hꢄ ꢄ ꢄA) 1.97 A, d(Dꢄ ꢄ ꢄA) 2.794(3) A, \(DHA) 165ꢀ,
˚
˚
˚
˚
O1–H1ꢄ ꢄ ꢄO2, d(Hꢄ ꢄ ꢄA) 1.86 A, d(Dꢄ ꢄ ꢄA) 2.6721(3) A, \(DHA) 163.6 ꢀ.
i = xꢀ1, y, z.

V. Albrow et al. / Inorganica Chimica Acta 359 (2006) 1731–1742
1741
sure of (Rc)-5 to 1.1 equivalents of LDA, followed by
Me2S2 allowed the formation of (Rp,Rc)-6a in 74% yield
although this compound could not be separated by chro-
matography or crystallisation from traces of recovered
(Rc)-5 which has very similar properties to 6a. Equivalent
procedures carried out on (Rc)-5 showed no hint of alterna-
tive (Sp,Rc) diastereomer testifying to the very high selectiv-
ity in the initial lithiation. Subsequent reaction of the crude
6a (racemic or chiral) with ButLi followed by O@C(CF3)2
gas resulted in the desired 3a in 65% yield which is now eas-
ily separated from traces of CpFe(g5-C5H4SMe), derived
from the impurity (5) and the starting material 6a. Chiral
(Rp,Rc)-6a (prepared from >99% ee (Rc)-5) suffered no ero-
sion of its enantiopurity during the second lithiation as
(Rp,Rc)-3a was isolated in >99% ee (HPLC). It was possible
to carry out a one-pot procedure for the synthesis of 3c
directly from 5 by sequential reaction with LDA/Me2S2
then ButLi/O@CPh2. The yield of the one-pot procedure
is significantly increased compared to the two-stage process
(57% versus 38%) making this the most convenient route to
3c. Formation of 6b–c also proceeded in synthetically use-
ful yields (74–81%). However, subsequent reaction of these
isolated alcohols with ButLi followed by electrophilic trap-
ping led to significantly lower yields (35–36%). Similar
problems were encountered in the preparation of 3c from
isolated 6c. The main by-product in these less successful
reactions of 6b was 1a pointing to protonoylsis of the inter-
mediate litho species derived from 6b. Identical behaviour
was noted on attempted lithiation of 6c followed by elec-
trophilic trapping. To confirm that no intrinsic congestion
problems were contributing to the observed lower yields
existing in these molecules a crystallographic study of 3c
was undertaken. The resulting structure (Fig. 6) was as
expected and confirms the viability of the target product
class.
Fig. 6. View of molecular structure of one of two independent molecules
of 3c, ellipsoids drawn at 30% probability level, H atoms other than OH
˚
omitted. Hydrogen bond details, O1–H1ꢄ ꢄ ꢄS1, d(Hꢄ ꢄ ꢄA) 2.39 A, d(Dꢄ ꢄ ꢄA)
˚
˚
˚
3.148(2) A, \(DHA) 150.2ꢀ O1a–H1aꢄ ꢄ ꢄ.S1a, d(Hꢄ ꢄ ꢄA) 2.42 A, d(Dꢄ ꢄ ꢄA)
3.168(2) A, \(DHA) 148.1ꢀ.
Two possibilities remained to account for the lower
yields and the observation of sulfoxide cleavage/proton-
ation by-products in these reactions of 6b–c: either the
presence of hydrogen bonded protic impurities in 6b/6c
or inefficient deprotonation of the C(CF3)2OH/CPh2OH
functions competing with faster than expected sulfoxide
cleavage upon lithiation of 6b/6c with two equivalents of
ButLi. To provide information on the former possibility a
crystallographic study was carried out on 6b. Selected
hydrogen bond distances and angles are provided in Fig. 7.
While the presence of hydrogen bonding between the
hydroxyl function and the sulfoxide was detected, careful
analysis of the unit cell revealed no occluded molecules
(protic or otherwise) at even partial occupancies so atten-
tion was turn to the reactivity of 6b/c with ButLi. Given
the relatively low pKa of the fluorous alcohols (ca. 9–10)
it was expected that deprotonation with ButLi would be
rapid, even at ꢀ78 ꢀC, compared to sulfoxide cleavage. In
a control reaction to test this hypothesis 6b was treated,
at ꢀ78 ꢀC, with one equivalent of ButLi and after 5 min
the reaction quenched with excess D2O. Combined TLC
and NMR analysis of the crude reaction product revealed
Fig. 7. View of one of the six independent molecules of 6b, ellipsoids
drawn at 30% probability level, H atoms other than OH omitted.
˚
Hydrogen bond details for all molecules, O1–H1ꢄ ꢄ ꢄO2, d(Hꢄ ꢄ ꢄA) 1.77 A,
˚
˚
˚
d(Dꢄ ꢄ ꢄA) 2.597(12) A, \(DHA) 170.2ꢀ, O1–H1ꢄ ꢄ ꢄS1, d(Hꢄ ꢄ ꢄA) 2.51 A,
d(Dꢄ ꢄ ꢄA) 3.182(8) A, \(DHA) 138.0ꢀ, O101–H101ꢄ ꢄ ꢄO102, d(Hꢄ ꢄ ꢄA)
˚
˚
1.91 A, d(Dꢄ ꢄ ꢄA) 2.606(13) A, \(DHA) 140.0ꢀ, O201–H201ꢄ ꢄ ꢄO202,
˚
˚
d(Hꢄ ꢄ ꢄA) 1.95 A, d(Dꢄ ꢄ ꢄA) 2.623(12) A, \(DHA) 136.0ꢀ, O301–
˚
˚
H301ꢄ ꢄ ꢄO302, d(Hꢄ ꢄ ꢄA) 1.84 A, d(Dꢄ ꢄ ꢄA) 2.531(12) A, \(DHA) 138.2ꢀ,
˚
˚
O301–H301ꢄ ꢄ ꢄS301, d(Hꢄ ꢄ ꢄA) 2.79 A, d(Dꢄ ꢄ ꢄA) 3.211(10) A, \(DHA)
˚
˚
112.6ꢀ, O401–H401ꢄ ꢄ ꢄO402, d(Hꢄ ꢄ ꢄA) 1.78 A, d(Dꢄ ꢄ ꢄA) 2.615(12) A,
˚
\(DHA) 175.7ꢀ, O401–H401ꢄ ꢄ ꢄS401, d(Hꢄ ꢄ ꢄA) 2.53 A, d(Dꢄ ꢄ ꢄA)
˚
˚
3.237(9) A, \(DHA) 142.7ꢀ, O501–H501ꢄ ꢄ ꢄO502, d(Hꢄ ꢄ ꢄA) 1.99 A,
˚
d(Dꢄ ꢄ ꢄA) 2.558(14) A, \(DHA) 124.6ꢀ, O501–H501ꢄ ꢄ ꢄS501, d(Hꢄ ꢄ ꢄA)
˚
˚
2.92 A, d(Dꢄ ꢄ ꢄA) 3.191(10) A, \(DHA) 100.9ꢀ.

1742
V. Albrow et al. / Inorganica Chimica Acta 359 (2006) 1731–1742
ture dependent rising from ca. +275 (at c = 1.00, 18 ꢀC)
to over +900 at 27 ꢀC.
In conclusion, we have demonstrated new and versatile
routes to new types of fluorous containing ferrocenyl spe-
cies and compared these to species containing the CPh2OH
fragment. The use of these building blocks in further
ligands and their applications to asymmetric catalysis is
being actively investigated in our lab at present.
4. Supplementary data
Additional crystallographic details and complete listings
for the structures reported have been deposited with the
Cambridge Crystallographic Data Centre (CCDC) with
the following CCDC Nos. 272951 (1a), 272952 (2),
272953 (3b), 272954 (3c), 272955 (3d), 272956 (4b),
272957 (6b). Copies of this information may be obtained
free of charge from the Director, CCDC, 12 Union Road,
Cambridge CB2 1EZ, UK (fax: +44 1223 336 033; web:
www.ccdc.cam.ac.uk).
Fig. 8. View of molecular structure of 3d, ellipsoids drawn at 30%
probability level, H atoms other than OH omitted. Hydrogen bond details,
O1–H1ꢄ ꢄ ꢄO2, d(Hꢄ ꢄ ꢄA) 1.90 A, d(Dꢄ ꢄ ꢄA) 2.709(3) A, \(DHA) 162.3ꢀ.
Acknowledgements
˚
˚
We are grateful to The University of Nottingham for
support of a studentship to V.A. and to EPSRC for partial
contributions (DRA and Grant GR/S560054/01 and
award of X-ray diffractometers). We thank the EU for
support from the COST-D24 (Working Group 003) pro-
gramme administered by the European Science Founda-
tion. We are indebted to Dr. E.S. Davies for help with
CV experiments.
that the product of this reaction was a mixture of
[CpFe(g5-C5H3(C(CF3)2OD)(S(O)4-Tol))] and [CpFe(g5-
C5H3DC(CF3)2OD)]. It is clear that the rates of deprotona-
tion and sulfoxide cleavage are comparable in 6b. A similar
experiment was carried out using [CpFe(g5-C5H3(CPh2-
OH)(S(O)4-Tol))] (6c) leading to [CpFe(g5-C5H3(CPh2-
OD)(S(O)4-Tol))] and [CpFe(g5-C5H3DCPh2OD)]. This
evidence indicates that sulfoxide cleavage and deprotona-
tion also occur at comparable rates at ꢀ78 ꢀC in 6c. These
problems can be overcome if a base which deprotonates
6b/c but which does not cleave the sulfoxide can be identi-
fied. Experimentation revealed that NaH fulfilled this role
allowing the synthesis of 3a and 3b from 6b in good yield.
In particular, 3e could be prepared by this NaH route as it
was not accessible via initial lithiation of (Rc)-5 with LDA
followed by ButLi/O@C(CF3)2 treatments as this lead only
to extensive decomposition. The preferred synthesis of
(RP)-3d was, however, by a one-pot procedure from (Rc)-
5 to avoid the need for alcohol deprotonation of isolated
6c. Compound 3d has only previously been prepared in
racemic form [11]. This species is a promising building
block for the development of chiral Schiffꢀs base/peptidic
ligands using the approach of Hoveyda [12]. To aid in com-
putation design of future ligands an X-ray crystal structure
of 3d was determined and this is shown in Fig. 8 along with
selected hydrogen bond distances and angles. The optical
rotatory power of 3d was found to be extremely tempera-
References
[1] H.-U. Blaser, Adv. Synth. Catal. 344 (2002) 17.
[2] G. Argouarch, O. Samuel, O. Riant, J.-C. Daran, H.B. Kagan, Eur.
J. Org. Chem. (2000) 2893.
[3] S. Steenken, Pure Appl. Chem. 70 (1998) 2031.
[4] K. Biswas, C. Borner, J. Gimeno, P.J. Goldsmith, D. Ramazzotti,
A.L.K. So, S. Woodward, Tetrahedron 61 (2005) 1433.
[5] E.L. Eliel, S.H. Wilen, L.N. Mander, Stereochemistry of Organic
Compounds, Wiley, New York, 1994, pp. 119–1122 (Chapter
14.1).
[6] K. Schlo¨gl, Top. Stereochem. 1 (1967) 39.
[7] M.I. Bruce, F. Gordon, A. Stone, B.J. Thomson, J. Organomet.
Chem. 77 (1974) 77.
[8] L. Branndsma, H. VerkniijssePreparative Polar Organometallic
Chemistry, vol. 1, Springer, Berlin, 1987, p. 19.
[9] For details of Spartan, see: <www.wavefun.com>.
[10] R. Masciadri, M. Kamer, N. Nock, Eur. J. Org. Chem. 21 (2003)
4286.
[11] J.H. Peet, B.W. Rockett, J. Chem. Soc., Perkin Trans. 1 (1973) 106.
[12] A.H. Hoveyda, A.W. Hird, M.A. Kacprzynski, Chem. Commun.
(2004) 1779.