Synthesis and electrochemical properties of a series of ferrocene-containing alcohols

Article abstract of DOI:10.1016/j.poly.2005.04.022

A series of primary ferrocenylalcohols, Fc-(CH₂)_m-OH with m = 1-4 and Fc = ferrocenyl, was synthesised by reduction of the appropriate ferrocenylcarboxylic acids, Fc-(CH₂)_n-COOH (n = 0-3) and the ester methyl 4-ferrocenylbutanoate with LiAlH₄, the reduction of the γ-ketoacid ferrocylpropanoic acid, Fc-CO-(CH ₂)₂-COOH, with AlCl₃/LiAlH₄, and the reduction of ferrocenylcarboxaldehyde, FcCHO, with NaBH₄. The secondary ferrocenyl alcohols CpFe(C₅H₄-CH(OH)-CH ₃) and Fe(C₅H₄-CH(OH)-CH₃) ₂ were obtained by NaBH₄ reduction of acetyl and diacetyl ferrocene. The different reduction methods are compared. The electrochemistry of the alcohols was studied by cyclic voltammetry in CH₃CN/0.1 M N(ⁿBu)₄PF₆ utilising a platinum working electrode. The ferrocenyl group showed reversible electrochemistry with the formal reduction potential (E^o′ versus Fc/Fc⁺) of the ferrocenyl group inversely proportional to side chain length. The influence of the side chain length on E^o′ was more pronounced for the acids because the electron-withdrawing properties of the carbonyl group is stronger than that of the alcohol group. Ion pairing was found to play a major role in the electrochemical behaviour of ferrocenylmethanol, Fc-CH ₂-OH.

Full text of DOI:10.1016/j.poly.2005.04.022

Polyhedron 24 (2005) 1611–1616
www.elsevier.com/locate/poly
Synthesis and electrochemical properties of a series
of ferrocene-containing alcohols
*
Wade L. Davis, Rebotsamang F. Shago, Ernie H.G. Langner, Jannie C. Swarts
Department of Chemistry, University of the Free State, Nelson Mandela Drive, P.O. Box 339, Bloemfontein 9300, South Africa
Received 7 February 2005; accepted 29 April 2005
Available online 20 June 2005
Abstract
A series of primary ferrocenylalcohols, Fc–(CH2)m–OH with m = 1–4 and Fc = ferrocenyl, was synthesised by reduction of the
appropriate ferrocenylcarboxylic acids, Fc–(CH2)n–COOH (n = 0–3) and the ester methyl 4-ferrocenylbutanoate with LiAlH4, the
reduction of the c-ketoacid ferrocylpropanoic acid, Fc–CO–(CH2)2–COOH, with AlCl3/LiAlH4, and the reduction of ferrocenylcar-
boxaldehyde, FcCHO, with NaBH4. The secondary ferrocenyl alcohols CpFe(C5H4–CH(OH)–CH3) and Fe(C5H4–CH(OH)–CH3)2
were obtained by NaBH4 reduction of acetyl and diacetyl ferrocene. The different reduction methods are compared. The electro-
chemistry of the alcohols was studied by cyclic voltammetry in CH3CN/0.1 M N(nBu)4PF6 utilising a platinum working electrode.
0
The ferrocenyl group showed reversible electrochemistry with the formal reduction potential (Eo versus Fc/Fc+) of the ferrocenyl
0
group inversely proportional to side chain length. The influence of the side chain length on Eo was more pronounced for the acids
because the electron-withdrawing properties of the carbonyl group is stronger than that of the alcohol group. Ion pairing was found
to play a major role in the electrochemical behaviour of ferrocenylmethanol, Fc–CH2–OH.
ꢀ 2005 Elsevier Ltd. All rights reserved.
Keywords: Synthesis; Ferrocene; Alcohol; Carboxylic acid; Electrochemistry; Cyclic voltammetry
1. Introduction
of carboxylic acids, Fc–(CH2)n–COOH, we have previ-
ously demonstrated that the antineoplastic properties
of the ferrocenyl group changes with increasing value
of ‘‘n’’ [4]. In particular, more beneficial cytotoxic prop-
erties for potential cancer treatment were identified for
compounds with the largest n-value. The explanation
for this observation was related to the formal reduction
potential of the ferrocenyl group, which became smaller
with increasing chain length. In our ongoing research on
the cytotoxic properties of ferrocene-containing com-
pounds, it became important to establish how the for-
mal reduction potential of the ferrocenyl group change
by a systematic increase in alkyl chain length between
the Fc and OH moieties in the alcohol series
Fc–(CH2)m–OH. In this regard, we were surprised about
the limited amount of physical studies that were per-
formed on alcohol derivatives of ferrocene. We report
with this paper the results of an electrochemical study
The synthesis of ferrocenylmethanol via the reduction
of ferrocenecarboxaldehyde or by trimethylamine
displacement from the methyl iodide salt of dimethylam-
inomethylferrocene by –OH [1], and of 2-ferrocenyleth-
anol from acetyl ferrocene via the intermediate
ferrocenylacetic acid [2] have been described. We have
reported on the use of 4-ferrocenylbutanol as precursor
for introducing the ferrocenyl group as high-burning
rate catalyst in solid propellants for the rocket industry
[3]. These three alcohols are members of a series of fer-
rocenyl-containing alcohols where the OH group is sep-
arated from the ferrocenyl (Fc) group by alkyl chains,
–(CH2)m–, of increasing chain length. In a related series
*
Corresponding author. Tel.: +27 51 4012781; fax: +27 51 4307805.
E-mail address: swartsjc.sci@mail.uovs.ac.za (J.C. Swarts).
0277-5387/$ - see front matter ꢀ 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.poly.2005.04.022

1612
W.L. Davis et al. / Polyhedron 24 (2005) 1611–1616
of the alcohol series Fc–(CH2)m–OH with m = 1–4, Fc–
CH(OH)–CH3 and Fe(C5H4–CH(OH)–CH3)2. Results
are expressed relative to ferrocene itself and are com-
pared with the related acid series, Fc–(CH2)n–COOH,
n = 0–3.
OH
OH
OH
Fe
Fe
13
14
2. Experimental
Fig. 1. Structures of secondary alcohols 13 and 14.
2.1. Equipment and materials
NMR spectrometer at 300 MHz with chemical shifts
presented as d values referenced to SiMe4 at 0.00 ppm
utilising CDCl3 that was made acid free by passing it
through basic alumina immediately before use. IR spec-
tra (cmꢀ1) were recorded on a Hitachi spectrophotome-
ter Model 270-50 with data processor. For solid state
spectra a KBr matrix was used, while NaCl plates were
used for thin films of liquid samples. Elemental analysis
was conducted by the Canadian Microanalytical Ser-
vice, Ltd.
Ferrocene (1, Strem), LiAlH4 (Merck), AlCl3 and
other solid reagents (Aldrich) were used without further
purification. Dry acetonitrile was obtained by refluxing
under nitrogen over CaH2, distilled onto alumina for
storage, and redistilled just prior to use. THF and ether
was dried by refluxing under a nitrogen atmosphere over
sodium wire and distilled directly before use. Organic
solvents were distilled prior to use, water was double dis-
tilled. Ferrocene carboxylic acids 2–5 [5], 6 [3], the ester
7 [3] and the aldehyde, 8 [6] were synthesised as de-
scribed before and used as precursors to the alcohols
9–12, Scheme 1. Alcohol 9 was synthesised by NaBH4
reduction of 8 in 90% yield as described by Broadhead
[1] but on a 10 times larger scale. The secondary alcohols
1-hydroxyethylferrocene (13) and 1,10-bis(hydroxy-
ethyl)ferrocene (14) Fig. 1, was obtained utilising the lit-
erature methods [3]. Chromatography was performed on
Fluka silica gel 60 grade 60741. The electrolyte tetrabu-
tylammonium tetrakis(pentafluorophenyl)borate was
prepared as described [7]. Melting points were deter-
mined with a Reichert Thermopan Microscope with a
Koffler hot-stage and are uncorrected. 1H NMR spectra
at 20 ꢁC were recorded on a Bruker Advance DPX 300
2.2. Synthesis
2.2.1. Fc–(CH2)m–OH, with m = 1 (9), 2 (10), 3 (11)
and 4 (12), utilising LiAlH4 reduction of acids 2–5
The synthesis of 3-ferrocenylpropanol (11) may serve
as an example. To a LiAlH4 suspension in dry ether
(3.42 g, 90 mmol in 100 cm3) was slowly added a solu-
tion of 3-ferrocenylpropanoic acid (4, 3 g, 12.6 mmol)
in dry ether (110 cm3) while slight boiling was main-
tained. After refluxing the mixture for 13 h at 50 ꢁC, a
1:1 ethanol:ether mixture (50 cm3) was very carefully
added to destroy the excess LiAlH4. The reaction mix-
ture was slowly poured into an excess of ice cold 2 M
O
O
COOH
COOCH3
OC
CO
Zn/HgCl2
,
Fe
Fe
Fe
HCl , CH3OH
AlCl3
,
CH2Cl2
1
6
7
LiAlH4
,
Zn/HgCl2
HCl , CH3COOH
(to obtain 5)
,
AlCl3
CH2Cl2
(to obtain 12)
,
LiAlH4
ether
(to obtain 12)
,
O
COOH
OH
(CH2)n
(CH2)m
LiAlH4
ether
,
NaBH4 ,
H
Fe
Fe
Fe
ethanol or
CH3OH
(to obtain 9)
9: m = 1
10: m = 2
11: m = 3
12: m = 4
8
2: n = 0
3: n = 1
4: n = 2
5: n = 3
m = n + 1
Scheme 1. Different routes that may be followed to obtain ferrocene-containing alcohols 9–12.

W.L. Davis et al. / Polyhedron 24 (2005) 1611–1616
1613
NaOH solution (1 l) and the ether and aqueous phases
2.2.3. Fc–(CH2)4–OH (12) from LiAlH4 reduction of
ester 7
separated. The aqueous phase was washed with ether
after which the combined ether fractions were washed
with 1 M NaOH solution (400 cm3) and twice with equal
volumes of water. Drying of the ether layer with MgSO4
and removal of the solvent under reduced pressure gave
11 as a reddish-brown oil (2.86 g, ꢁ100%) that was clean
enough for most purposes. A purification column may
be employed with ethyl acetate–hexane (1:4) as eluent
(Rf = 0.36). 1H NMR (300 MHz, CDCl3): dH 4.14
(5H, s, C5H5), 4.11 (2H, s, C5H4), 4.08 (2H, s, C5H4),
3.69 (2H, s, –CH2–OH), 2.44 (2H, t, Fc–CH2–), 1.80
(2H, m, –CH2–). IR (NaCl-discs), mmax/cmꢀ1: 3352
(–OH), 2926 (–CH2–), 1104 (C–O). C13H16OFe requires:
C, 64.0; H, 6.6. Found: C, 63.9; H, 6.3%.
Solid LiAlH4 (2.7 g, 71 mmol) was added to a solu-
tion of 7 (6.32, 22.1 mmol) in 80 cm3 dry ether over 20
min. The reaction mixture was refluxed for 16 h before
the excess LiAlH4 was destroyed by careful addition of
a 1:1 ethanol:ether mixture (100 cm3). The ether layer
was washed successively with HCl (2 · 200 cm3, 2 mol
dmꢀ3), NaOH (200 cm3, 2 mol dmꢀ3) and water, and
dried with MgSO4, to give after solvent removal 12
(5.31 g, 93%).
2.3. Cyclic voltammetry
Cyclic voltammetric (CV) experiments were per-
formed on ca. 1 mM solutions of 9–14 in dry CH3CN/
0.100 M N(nBu)4PF6, utilising a standard three-elec-
trode cell with a Pt-working electrode of surface area
2.01 mm2 and pre-treated by polishing on a Beuhler
microcloth first with 1 micron and then with 1/4 micron
diamond paste, a Pt-wire counter electrode and a Ag/
Ag+ reference electrode (a Ag wire immersed in 0.010
M AgNO3 in CH3CN utilising a Luggin capillary with
vicor tip) under argon at 25.0 ꢁC in a Faraday cage con-
nected to a BAS CV-27 cyclic voltammograph inter-
faced with a personal computer. Some experiments
were also conducted on a glassy carbon electrode of sur-
face area 7.07 mm2. Obtained formal reduction poten-
tials were essentially the same. Data, uncorrected for
junction potentials, were collected with an Adalab-
PCꢂ and Adaptꢂ data acquisition kit (Interactive
Microwave, Inc.) with locally developed software, and
analysed with Hyperplot (JHM International, Inc.).
All potentials in this study, however, are presented refer-
Characterisation data for 9: Chromatography on sil-
ica with ethyl acetate–hexane (1:4) as eluent (Rf = 0.28)
gave 9 in 89% yield as a yellow solid, m.p. 76 ꢁC. H
1
NMR (300 MHz, CDCl3): dH 4.35 (2H, d, C5H4), 4.27
(2H, t, C5H4), 4.21 (5H, s, C5H5), 4.20 (2H, s, Fc–
CH2–OH). IR (KBr), mmax/cmꢀ1: 3227 (–OH), 2932
(–CH2–), 1102 (C–O). C11H12OFe requires: C, 61.2; H,
5.6. Found: C, 61.4; H, 5.4%.
Characterisation data for 10: Chromatography on sil-
ica with ethyl acetate–hexane (1:4) as eluent (Rf = 0.25),
gave 10 as a low melting solid in 93% yield, m.p. 32–33
ꢁC. 1H NMR (300 MHz, CDCl3): dH 4.13 (9H, m,
–C5H4 and C5H5), 3.75 (2H, m, –CH2–OH) and 2.62
(2H, t, Fc–CH2–). IR (NaCl discs), mmax/cmꢀ1: 3408
(–OH), 2935 (C–H), 1038 (C–O). C12H14OFe requires:
C, 62.6; H, 6.1. Found: C, 62.6; H, 5.8%.
Characterisation data for 12: Chromatography on sil-
ica with ethyl acetate–hexane (1:4) as eluent (Rf = 0.49),
1
gave 12 as a thick dark yellow liquid in 94% yield. H
NMR (300 MHz, CDCl3): dH 4.12 (5H, s, C5H5), 4.06
(4H, d, C5H4), 3.68 (2H, t, CH2–OH), 2.37 (2H, t, Fc–
enced against Fc/Fc+ as recommended by IUPAC [8].
0
The Fc/Fc+ couple exhibited Eo = 0.087 V versus Ag/
Ag+, ipc/ipa = 0.98, DEp = 74 mV under our experimental
conditions. Successive experiments under the same
experimental conditions have shown that potentials
were reproducible within 3 mV. CV experiments with
9 in CH2Cl2/0.100 M [N(nBu)4][B(C6F5)4], utilising a
glassy carbon electrode of surface area 0.79 mm2, were
experimentally referenced against a Ag/AgCl reference
electrode.
CH2), 1.60 (4H, m, 2 · CH2). IR (NaCl-discs), mmax
/
cmꢀ1: 3346 (–OH), 2800–2900 (C–H), 1104 (C–O).
C14H18OFe requires: C, 65.1; H, 7.0. Found: C, 65.3;
H, 6.9%.
2.2.2. Fc–(CH2)4–OH (12) utilising LiAlH4/AlCl3
reduction of the c-ketoacid (6)
To a suspension of LiAlH4 (0.267 g, 6.99 mmol) in
dry ether (5 cm3) was added AlCl3 (0.228 g, 1.63 mmol),
also suspended in dry ether (45 cm3) under nitrogen.
3-Ferrocenoylpropanoic acid (6, 0.208 g, 0.699 mmol)
was added and additional AlCl3 (0.189 g, 1.39 mmol)
suspended in dry ether (9 cm3) was added over a 5
min interval. The reaction mixture was stirred for 30
min at room temperature and refluxed for 3 h. To the
cooled reaction mixture was added water (8 cm3) and
6 mol dmꢀ3 H2SO4 (5 cm3). The mixture was extracted
with ether (3 · 80 cm3), dried with MgSO4 and solvent
removal afforded a yellowish oil of 4-ferrocenylbutanol
(12) (0.18 g, 97%).
3. Results and discussion
3.1. Synthesis
The alcohol series 9–12 were synthesised according to
Scheme 1. Ferrocenylmethanol (9) was synthesised from
ferrocenecarboxaldehyde (8), utilising NaBH4 as reduc-
ing agent in high (ꢁ90%) yield. NaBH4 is somewhat
safer to use in large quantities than LiAlH4. However,
NaBH4 is not a strong enough reducing agent to reduce

1614
W.L. Davis et al. / Polyhedron 24 (2005) 1611–1616
Table 1
Selected potentials (V) and current (lA) data for ca. 1.0 mM solutions
carboxylic acids or esters. Hence, for the reduction of
2–5 and 7, LiAlH4 was used. All these reactions were
essentially quantitative.
of the indicated compounds in CH3CN/0.100 M N(nBu)4PF6 at 25.0 ꢁC
0
Compound
Epa
DEp
b
Eo
ipa
b
ipc/ipa
The yield-limiting step in obtaining alcohols 9–12
from ferrocene is centred on the success rate in obtaining
the acid or other precursors. Hence the preparation of 9
from aldehyde 8 gave a much better overall yield (71% in
a two-step syntheses) based on the initial amount of fer-
rocene used, than the preparation utilising the acid, 2
(ca. 49% yield at least in a three-step syntheses). Because
LiAlH4 reduction of the present ferrocene acid series is
so facile, there is nothing gained in terms of yields by
synthesising alcohols 9–12 utilising ester precursors.
This is demonstrated for the synthesis of 4-ferrocenyl-
butanol (12) from ester 7. It is also unnecessary to ob-
tain 5 for the synthesis of 12, since this alcohol can
directly be obtained by the reduction of the c-ketoacid
(6). However, in this latter reaction, LiAlH4 does not
liberate the alcohol 12 because LiAlH4 does not reduce
the c-keto group of 6 to a CH2 group. Rather, the for-
mation of an a-alcohol intermediate possessing the
molecular fragment Fc–CH(OH)– and follow-up degra-
dation occurs. To obtain 12 from 6, the use of the mixed
reducing system AlCl3/LiAlH4 leads to simultaneous
Clemenson reduction of the c-keto group to –CH2–
and the carboxylic acid group to –CH2–OH. This reac-
tion is also essentially quantitative. The long term stabil-
ity of the alcohols with spacers (CH2)m and m = 2, 3 and
4 is excellent. However, the shelf-life of Fc–CH2–OH (9)
and also the secondary alcohols 13 and 14 (Fig. 1), is less
impressive. This is the result of the high electron donat-
ing power of the ferrocenyl group (group electronegativ-
ity of Fc is vFc = 1.89 [9,10þ]), which tend to stabilise
ferrocyl cations ð¼ Fc–CH2 Þ [11,12]. This promotes
decomposition of 9 by water elimination under very
mild conditions as follows:
b
9 (peak 1)
9 (peak 2)
ꢀ0.017a
0.093
0.016c
0.060c
ꢀ0.046
ꢀ0.052
ꢀ0.054
ꢀ0.011
ꢀ0.021
b
b
b
10
11
12
13
14
ꢀ0.011
ꢀ0.016
ꢀ0.017
0.022
0.070
0.070
0.075
0.066
0.066
8.78
8.28
7.61
6.47
8.23
0.96
0.97
0.99
0.98
0.95
0.012
a
Epc value.
A poorly defined oxidation wave in peak 1 and reduction wave in
b
peak 2 made current and DEp measurements impossible.
0
c
Eo were estimated by assuming DEp = 66 mV, the value obtained
for 13 and 14.
the CV of 9 hinted on two poorly resolved ferrocene-
based waves (Fig. 3) despite the fact that 9 gave satisfac-
1
torily elemental analyses and having a clean H NMR
spectrum. In an attempt to prove that their really are
two peaks associated with the CV of 9, Osteryoung
Square Wave Voltammetry (OSWV) was performed on
this complex. The OSWV peaks were certainly broad-
ened, but no clear-cut indication towards two redox
peaks could be demonstrated. It then was suspected that
the use of the coordinating solvent CH3CN and 0.100 M
N(nBu)4PF6 as supporting electrolyte masked some key
features of the CV of 9. Reports by Geiger and co-work-
ers [13] that poorly resolved peaks in CH3CN/0.100 M
N(nBu)4PF6 often becomes well resolved in the system
dichloromethane utilising the hyper uncoordinative
and poorly nucleophylic salt [N(nBu)4][B(C6F5)4] as sup-
porting electrolyte, caused us to repeat the CV of 9 in
Fc–CH2–OH þ Hþ ! Fc–CH2–ðOH2Þþ
! Fc–ðCH2Þþ þ H2O
Free Fc
70
12
The ferrocenyl group cannot stabilise cations of this
nature in 10–12.
50
11
3.2. Cyclic voltammetry
30
CV experiments in dry CH3CN/0.100 M N(nBu)4PF6
indicated reversible electrochemical behaviour for the
ferrocenyl group in compounds 10–14 with
DEp = Epa ꢀ Epc 6 80 mV at scan rates up to 250 mV
sꢀ1 in acetonitrile (Table 1, Fig. 2). Determined peak
cathodic/peak anodic current ratios for 10–14 were in
the range 0.94 < ipc/ipa < 1 and the formal reduction
10
10
9
-10
o0
-0.3
-0.1
0.1
0.3
0.5
potentials, E =1/2 (Epa + Epc), were scan rate indepen-
dent between 50 and 250 mV sꢀ1
E/Vvs Fc/Fc+
.
Fig. 2. CVÕs of 9–12 in CH3CN/0.100 M N(nBu)4PF6 at 25.0 ꢁC and a
Compound 9, though, exibited DEp = 110 mV, much
larger than for the other alcohols. Close inspection of
scan rate of 100 mV sꢀ1 at a Pt-working electrode.

W.L. Davis et al. / Polyhedron 24 (2005) 1611–1616
1615
10
6
3
1
2
-1
-3
-2
-6
-0.3
-0.2
-0.1
0
0.1
0.2
0.3
-0.50
-0.30
-0.10
0.10
0.30
E/V vs Fc/Fc+
E/V vs Fc/Fc+
Fig. 3. CV of 9 in CH3CN/0.100 M N(nBu)4PF6 (left) showing two poorly resolved peaks. In CH2Cl2/0.100 M [N(nBu)4][B(C6F5)4] (right, bold line),
two peaks with DEp ꢁ 89 mV are clearly seen. Upon addition of 0.100 M N(nBu)4PF6 to this solution, the CV collapsed into one showing two poorly
resolved peaks (right, less bold line).
the system CH2Cl2/0.100 M [N(nBu)4][B(C6F5)4] with
spectacular success.
dimerisation as it is realised for 9 is not as feasible in 13
and 14.
Fig. 3, right, clearly shows two very well-defined
reversible CV waves for 9 in this medium. In a separate
test, the stability of 9 in acetonitrile was investigated by
1H NMR. This compound did not show any degrada-
tion within 24 h. The two observed waves in Fig. 3,
right, are typical of those associated with a symmetrical
dimer, which generates a mixed valent intermediate
upon oxidation or reduction. Different formal reduction
potentials for side groups on symmetrical complexes in
which mixed-valent intermediates are generated are well
known in systems that allow electron delocalisation,
either through bridge mediated paths or from a direct
metal–metal interaction [10,14]. The inequivalence of
the ferrocenyl and ferrocenium groups of such a mixed
valent intermediate is highlighted in terms of their group
electronegativities [9,10]: vFc = 1.87, vFc+ = 2.82. The
electron withdrawing power of the ferrocenium group
is almost as high as that of the CF3 group (vCF3 = 3.01).
We therefore propose the formation of a dimeric species
of 9 in solution that electrochemically are oxidised
according to the following electron transfer processes:
The structure of the mixed valent species in Eq. (1)
above can only be speculated on at this stage, but in
view of the known hydrogen bonding capabilities of
alcohols and the fact that the ferrocenyl group are
known to interact with protons [15], a hypothetic struc-
ture as in Fig. 4 would be consistent with our results.
It would allow a symmetrical dimer to account for the
shape of the CV. It would allow communication of the
FeII and FeIII species through space and through conju-
gation, and it disallows a similar communication pattern
of longer chain alcohols in two ways. Firstly, in
–(CH2)m– chains with m P 2, electron delocalisation
through conjugation or inductive effects along the side
chains would be very weak indeed. Secondly, through
space communication of the FeII and FeIII species, or
the FeIIꢃ ꢃ ꢃH–O–CH2–Fc interaction in Fig. 4 would be
very unlikely over the distances that will be required
for the longer side chains.
The lower Eo for 10 compared to 9 (Table 1) clearly
indicates that there is much less communication be-
0
tween the hydroxyl group and the ferrocenyl moiety
0
of 10 than in 9. In particula0r, for Eo of 10 to be 46
mV more negative than Eo of ferrocene itself, the
side-chain of 10, –CH2CH2OH, has to become elec-
tron-donating relative to H. This is consistent with a
conclusion that –CH2CH2– and larger alkyl spacers
effectively isolates the –OH group from the ferrocenyl
½2FcCH2OH ꢀ ðFcCH2OHÞ ꢂ
2
ꢀ ðFcþCH2OHÞ–ðFcCH2OHÞ
ꢀ ðFcþCH2OHÞ
ð1Þ
2
It is instructive to observe that 10–14 do not show the
double peaks in their CVÕs as 9 had. This implies that
the structural effects that was discussed and interpreted
above as dimerisation for 9 is absent in 10–14. This
observation is consistent with the fact that the longer
chain of 10–12 introduces so much flexibility into the
side chains of these molecules compared to what is ex-
pected for 9, that a dimeric product may not be stable
enough to allow its detection in noticeable amounts on
a CV time scale. As for 13 and 14, it is conceivable that
the methyl groups in the side chain is bulky enough to
impose so much steric constraints on the molecules that
+
-
δ
δ
H
O
FeII
+
O
H
-
+
δ
δ
FeIII
Fig. 4. Possible mixed valent intermediate in the oxidation of 9.

1616
W.L. Davis et al. / Polyhedron 24 (2005) 1611–1616
group, but that a single –CH2– group does not. This
also means that electron-donating properties of the
–CH2CH2– part of the side-chain in 10 dominates over
the electron-withdrawing properties of the –OH moiety
of 10.
Further increases in 0the side chain length to that of 11
and 12, lowered the Eo values further, but not nearly as
dramatically as did 10, probably because the –(CH2)m–
chains became so long that they are very effective isola-
tors: the ferrocenyl group is really existing independent
of any through-space or through-bond inductive effects
that the OH group may have. The only thing that the
slight lowering of the formal reduction potentials of 11
and 12 are thought to imply is a slight increase in the
electron donating power of the alkyl chains with increas-
In conclusion, the yield-limiting step to obtain ferro-
cenylalcohols is centred in the successful high yield syn-
thesis of suitable precursors. Reduction reactions of
these precursors to obtain ferrocenyl alcohols are very
efficient. The ferrocenylalcohols under discussion have
smaller reduction potentials than the carboxylic acid
analogues, and they become easier to oxidise with
increasing alkyl chain length. Ferrocenylmethanol un-
dergo solution molecular pairing to a notable extent.
A forthcoming publication [17] will provide the details
of how more beneficial cytotoxic properties for potential
cancer treatment by 9–12 are observed with increasing
side chain length and decreasing ferrocenyl formal
reduction potential.
ing chain length.
0
The Eo of the secondary alcohols 13 and 14 is more
negative than that of 9. This is because they have addi-
tional electron-donating methyl groups in the side-
Acknowledgements
The authors acknowledge the UFS, CANSA and the
MRC for financial support.
chains compared to 9. It is also instructive to note that
0
our Eo = ꢀ0.011 V for 13 corresponds very well with
the chronopotentiometric determined E1/4 = ꢀ0.009 V
value that was reported by Kuwana and co-workers
[16]. Alcohol 14 is more easily oxidised than 13 which
is what one may expect when the number of electron
donating side chains, here two –CH(CH3)OH groups,
are doubled.
References
[1] (a) G.D. Broadhead, J.M. Osgerby, P.L. Pauson, J. Chem. Soc.
(1958) 650;
(b) J.K. Lindsay, C.R. Hauser, J. Org. Chem. 92 (1957) 355.
[2] C. Nataro, W.M. Cleaver, C.C. Landry, C.W. Allen, Polyhedron
18 (1999) 1471.
Fig. 5 demonstrates a few final key points pertinent to
compounds of the type Fc–(CH2)n–COOH and Fc–
(CH2)m–OH.
[3] P.J. Swarts, M. Immelman, G.J. Lamprecht, S.E. Greyling, J.C.
Swarts, S. Afr. J. Chem. 50 (1997) 208.
0
Firstly, it can clearly be seen how the value of Eo
decreases as the alkyl chain of the carboxylic acid series
2–5 and the ferrocenylalcohol series 9–12 becomes long-
er. Secondly, the electron withdrawing properties of
both the carbonyl group of –COOH and the0OH group
in Fc–(CH2)n/m-compounds see to it that Eo are much
more drastically affected at short chain length than in
compounds with longer side chains. Finally, it can be
noted that ferrocenyl alcohols are more easily oxidised
than the parent carboxylic acids 2–5.
[4] J.C. Swarts, D.M. Swarts, D.M. Maree, E.W. Neuse, C. La
Madeleine, J.E. van Lier, Anticancer Res. 21 (2001) 2033.
[5] N.F. Blom, E.W. Neuse, H.G. Thomas, Transition Met. Chem.
12 (1987) 301.
[6] D.E. Bublitz, K.L. Rinehart, Org. Reactions 17 (1969) 77.
[7] R.J. LeSuer, C. Buttolph, W.E. Geiger, Anal. Chem. 76 (2004)
6395.
[8] G. Gritzner, J. Kuta, Pure Appl. Chem. 56 (1984) 461.
[9] W.C. Du Plessis, T.G. Vosloo, J.C. Swarts, J. Chem. Soc., Dalton
Trans. (1998) 2507.
[10] W.C. Du Plessis, J.J.C. Erasmus, G.J. Lamprecht, J. Conradie,
T.S. Cameron, M.A.S. Aquino, J.C. Swarts, Can. J. Chem. 77
(1999) 378.
[11] E.W. Neuse, D.S. Trifan, J. Am. Chem. Soc. 85 (1963) 1952.
[12] J.H. Richards, E.A. Hill, J. Am. Chem. Soc. 81 (1959) 3484.
[13] F. Barriere, R.U. Kirss, W.E. Geiger, Organometallics 10 (1991)
2403, and references therein.
300
(CH ) COOH
2 n
Fe
200
100
0
[14] See for example:
(a) C. Creutz, H. Taube, J. Am. Chem. Soc. 91 (1969) 3988;
(b) W.E. Geiger, N. Van Order, D.T. Pierce, T.E. Bitterwolf,
A.L. Reingold, N.D. Chasteen, Organometallics 10 (1991) 2403;
(c) N. Van Order, W.E. Geiger, T.E. Bitterwolf, A.L. Reingold, J.
Am. Chem. Soc. 109 (1987) 5680;
vs.
(CH
)
OH
2 m
Fe
(d) D.T. Pierce, W.E. Geiger, Inorg. Chem. 33 (1994) 373;
(e) J. Conradie, T.S. Cameron, M.A.S. Aquino, G.J. Lamprecht,
J.C. Swarts, Inorg. Chim. Acta 358 (2005) 2530.
[15] T.J. Curphey, J.O. Santer, M. Rosenblum, J.H. Richards, J. Am.
Chem. Soc. 82 (1960) 5249.
-100
0
1
2
3
4
n or m (nr. of C-atoms in chain)
0
Fig. 5. Graph of Eo vs. number of carbon atoms in the alkyl chain
linking the ferrocenyl group with functional groups for Fc–(CH2)n–
[16] T. Kuwana, D.E. Bublitz, G. Hoh, J. Am. Chem. Soc. 82 (1960)
5811.
[17] R.F. Shago, J.C. Swarts, Anticancer Res. (2005), accepted.
COOH ( ) and Fc–(CH ) –OH (¤). The point d refers to the second
n
couple of 9.
2 m