Alkali-metal binding properties of simple ferrocenyl- and ruthenocenyl-substituted aza-crown ethers

Article abstract of DOI:10.1039/a703597g

Reaction of 1-chlorocarbonyl-ferrocene or -ruthenocene with aza-15-crown-5 (1,4,7,10-tetraoxa-13-aza-cyclopentadecane) afforded the respective amide-linked metallocene crowns. Nucleophilic displacement of trimethylamine from trimethylammoniomethyl-ferrocene or -ruthenocene iodides with aza-15-crown-5 yielded the corresponding methylene-linked crowns. Variable-temperature ¹³C NMR investigation of the hindered rotation about the amide links of the amide-linked crowns in the presence of guest cations indicated that cation complexation occurs exclusively within the crown ring. Extraction experiments with a range of metal-cation dye salts revealed a favourable binding capacity for the methylene-linked crowns. Detailed ¹³C NMR analysis of sodium and potassium ion binding by these systems suggests the presence of a stabilising interaction between the metallocene π system and the crown-bound cation. Cyclic voltammetric analysis of metal-ion binding by the amine-linked crowns revealed cation-induced shifts of the appropriate redox couple comparable with those reported for related systems.

Full text of DOI:10.1039/a703597g

View Article Online / Journal Homepage / Table of Contents for this issue
Alkali-metal binding properties of simple ferrocenyl- and
ruthenocenyl-substituted aza-crown ethers
Martin C. Grossel,*,† Darren G. Hamilton, Jason I. Fuller and Enrique Millan-Barios
Department of Chemistry, University of Southampton, Highfield, Southampton, UK SO17 1BJ
Reaction of 1-chlorocarbonyl-ferrocene or -ruthenocene with aza-15-crown-5 (1,4,7,10-tetraoxa-13-aza-
cyclopentadecane) afforded the respective amide-linked metallocene crowns. Nucleophilic displacement of
trimethylamine from trimethylammoniomethyl-ferrocene or -ruthenocene iodides with aza-15-crown-5 yielded the
corresponding methylene-linked crowns. Variable-temperature 13C NMR investigation of the hindered rotation
about the amide links of the amide-linked crowns in the presence of guest cations indicated that cation
complexation occurs exclusively within the crown ring. Extraction experiments with a range of metal-cation dye
salts revealed a favourable binding capacity for the methylene-linked crowns. Detailed 13C NMR analysis of
sodium and potassium ion binding by these systems suggests the presence of a stabilising interaction between the
metallocene π system and the crown-bound cation. Cyclic voltammetric analysis of metal-ion binding by the
amine-linked crowns revealed cation-induced shifts of the appropriate redox couple comparable with those
reported for related systems.
Prompted by the undeniable attraction of developing host
O
O
O
O
molecules capable of exhibiting a selective electrochemical
response to the presence of a guest cation, redox-active iono-
phores based on covalently linked crown ether–metallocene sys-
tems have received much attention in recent years.1–5 In spite of
the rather limited structural variation offered by the attachment
of simple crown ether rings to metallocene units a variety of
modes of complexation of metal cations have been identified
for such systems. Some of the first metallocenylcryptands, con-
taining amide links between the ionophore and metallocene,
were found to bind cations via their carbonyl functions rather
than within the crown ring.6 Simple metallocene–crown con-
jugates tended to behave more predictably with cation co-
ordination occurring via the crown ring system.7 Additionally,
there is evidence that transition-metal-ion binding by certain
ferrocenylcryptands involves a direct, stabilising ferrocenyl
iron–cation interaction.8 Such observations prompted the syn-
thesis and investigation of a series of simple metallocene crowns
9, 10, 15 and 16. Our aim was not to target a particular guest
cation or attempt to tune any electrochemical response, rather
we hoped to examine in some detail the factors involved in the
binding of guest cations by an investigation of the behaviour of
some constitutionally simple members of this much studied
class of materials.
N
O
N
O
O
O
O
Fe
Fe
O
N
O
O
O
O
N
O
O
O
O
1
2
N
N
O
O
O
O
O
O
O
Fe
Fe
O
O
N
O
N
3
4
into 14 proved rather less efficient (38%) than that for its ferro-
cene analogue, a reflection of the generally lower aromatic
reactivity of ruthenocene towards electrophilic species.9 Scheme
1 summarises the synthetic routes.
All new compounds were fully characterised by 1H and
13C NMR spectroscopy and either high-resolution FAB
mass spectrometry or elemental analysis (see Experimental
section).
Results and Discussion
Syntheses
Following well established methods7 reaction of 1-(chloro-
carbonyl)-ferrocene 7 or -ruthenocene 8 with aza-15-crown-5 in
the presence of NEt3 gave the respective amido-linked metallo-
cene crowns 9 and 10 after chromatographic purification. The
methylene-linked analogues 15 and 16 were prepared from the
quaternary ammonium metallocene salts 13 and 14 by nucleo-
philic displacement using aza-15-crown-5 in the presence of
K2CO3 (the ammonium salts having been obtained by quatern-
isation of dimethylaminomethylmetallocenes 11 and 12 with
MeI). Whilst the dimethylamine derivative 11 is commercially
available, its ruthenocene analogue 12 must be prepared by the
Mannich reaction of ruthenocene with Me2NCH2NMe2 using a
procedure analogous to that described for 11. The conversion
Variable-temperature (VT) 13C NMR spectroscopy
The temperature dependence of NMR spectra can be used to
obtain information about internal energy barriers to rotation
within a molecule. The restricted rotation about the C᎐N bond
of an amide linkage is the classical example of a dynamic
process that can be studied using this technique.10 Resonance
theory predicts a certain degree of double-bond character for
the C᎐N link, with an associated increase in the barrier to
rotation over that observed for strict single C᎐N bonds such as
those found, for example, in amines. On the chemical shift time-
scale it is possible to freeze this rotation and observe coales-
cence phenomena as the temperature is increased. Readily
measured parameters allow calculation of the energy barrier to
rotation (see Experimental section).
† E-Mail: mcg1@soton.ac.uk
J. Chem. Soc., Dalton Trans., 1997, Pages 3471–3477
3471

View Article Online
O
O
N
O
O
O
CO2H
COCl
(i )
(ii )
M
M
M
7, M = Fe
8, M = Ru
5, M = Fe
6, M = Ru
9, M = Fe
10, M = Ru
O
O
NMe3+I–
(iv )
N
NMe2
O
O
(iii )
M
M
M
13, M = Fe
14, M = Ru
11, M = Fe
12, M = Ru
15, M = Fe
16, M = Ru
Scheme 1 (i) (COCl)2, CH2Cl2, dimethylformamide (dmf) (catalyst); (ii) aza-15-crown-5 (1,4,7,10-tetraoxa-13-azacyclopentadecane), NEt3,
benzene; (iii) MeI, MeOH; (iv) aza-15-crown-5, K2CO3, MeCN
The activation barrier for amide rotation in the ferrocene
bis(crown) 1 has been reported elsewhere (∆G‡ = 65.0 kJ molϪ1
,
Table 1 Summary of 13C VT NMR experimental data
no error quoted),7 a value typical of a ferrocene-linked amide
bond.11 Additional experiments performed in the presence of
stoichiometric amounts of metal salts revealed that the pres-
ence of sodium ions lowered this barrier by around 5 kJ molϪ1
whereas in the case of lithium ions it was increased by 2.5 kJ
molϪ1. These observations were presented as supporting evid-
ence for different modes of complexation of the bis(crown) with
various metal cations. From solid-state complexation studies,
sodium ions are known to reside within the crown ether rings of
1 where they will attract electron density from the amide nitro-
gen thus weakening the bond and lowering the rotation energy
barrier. In contrast, lithium ions are known to co-ordinate
externally to the two amide carbonyl groups in such structures,
further polarising the C᎐O bond, thereby increasing the double-
bond character of the C᎐N bond and raising the activation
energy to rotation.
Crown
Tc a/K
∆ν/Hz
∆G‡ b/kJ molϪ1
9
10
310 ± 5
314 ± 5
65
123
63.2 ± 1.0
62.4 ± 1.0
Error involved in determining Tc from spectral inspection. b Error
associated with coalescence temperature accuracy of ±5 K.
a
Table 2 Results of 13C VT NMR experiments in the presence of metal
ions
Experiment
∆G‡ a/kJ molϪ1
9
63.4 ± 1.0
60.0 ± 1.0
59.9 ± 1.0
b
9 ϩ LiClO4
9 ϩ NaClO4
b
a
Error involved in determining Tc from spectral inspection, ±5 K.
b Addition of 1 molar equivalent of metal salt.
The VT NMR technique has now been applied to metallo-
cene amido-crowns 9 and 10 in order to determine whether a
single binding mode was operating. Variable-temperature 13C
NMR spectra were recorded using CD3CN solutions of the
crowns at 91 MHz. The coalescence temperatures (Tc), frozen-
condition frequency separations (∆ν) and activation energy
barriers (∆G‡) obtained from the variable-temperature experi-
ments are summarised in Table 1. These results show reasonable
agreement with the barriers to rotation observed for ferrocene
bis(crown) 1 (65.0 kJ molϪ1) and its ruthenocene analogue (64.1
kJ molϪ1). They support previously reported observations11
that an amido substituent on one metallocene ring has neg-
ligible effect on a substituent on the other, and that the nature
of the metallocene has little effect on the value of the rotation
barrier (values for ferrocene derivatives are consistently slightly
higher than those obtained for the equivalent ruthenocene
compounds, again perhaps reflecting the more electron-rich
nature of the ferrocene unit).
metallocene bis(crowns) such as 1. This result suggests that a
necessary prerequisite for metal-ion binding outside the crown
cavity is the presence of two carbonyl groups and that for the
simple monocrowns described here only the crown cavity itself
provides a suitable metal-ion binding site.‡
Assessment of metal-ion extraction ability using oxonol dyes
The picrate-salt extraction technique devised by Cram has
become a widely employed method for the screening of ligands
for their ability to bind metal ions.12–15 The method is based on
monitoring the transfer of picrate anion from an aqueous to an
organic phase. This transfer accompanies the extraction of a
metal cation from the aqueous phase by a host compound, in
order to preserve electrical neutrality. Thus, a measure of the
ability of a host to bind a given metal ion can be obtained
through simple spectrophotometric measurements. The use of
oxonol dyes in place of picrate salts offers several advantages
when applied as a rapid screening technique.4 The dyes are
straightforward to synthesize for a range of alkali, alkaline-
earth and transition metals, and are far safer to handle than the
Further 13C VT NMR experiments were conducted with ferro-
cene crown 9 in the presence of lithium and sodium ions (added
as their perchlorate salts). The results are summarised in Table 2.
The addition of both metals is seen to reduce the activation
energy barrier to rotation by around 3 kJ molϪ1. Following the
argument discussed previously, this behaviour is consistent with
the binding of each of the metal cations within the crown cavity
where the electrostatic attraction between the amide nitrogen
and the metal cation lowers the electron density associated with
the C᎐N bond, thus lowering the rotational energy barrier. The
observation that the rotation barrier is lowered in the presence
of lithium ions is in direct contrast with the situation found in
‡ Infrared spectroscopy reveals that in acetonitrile solution the amide
carbonyl stretching frequency of compound 9 (1604 cmϪ1) is subject to
only a small change on the addition of 1 equivalent of either lithium
(1602 cmϪ1) or sodium (1610 cmϪ1) perchlorate. It could perhaps be
argued that the change in the sodium case, although small, is consistent
with a slight strengthening of the C᎐N bond resulting from metal-ion
binding within the crown cavity.
3472
J. Chem. Soc., Dalton Trans., 1997, Pages 3471–3477

View Article Online
O
Table 3 Percentage extraction abilities for metallocene crowns with a
range of metal-ion oxonol dyes
O
O
N
+
O
Metal-ion oxonol dye
ϩ
Crown
Liϩ
Naϩ Kϩ
NH4
Agϩ
Mg2ϩ Ca2ϩ
9
15
10
16
1
25
99
20
99
97
32
97
19
98
95
30
99
18
97
100
44
99
34
99
97
40
100
27
100
68
0
54
0
58
35
0
48
0
55
28
M
Fig. 1 Proposed mode of cation complexation in compounds 15ؒNaϩ
and 16ؒNaϩ
Me
Me
assignments are given in Table 4. The data for both crowns are
entirely consistent with the formation of 1:1 stoichiometry
complexes; no further changes are observed once a 1:1 ratio
of crown :Naϩ has been reached. Considerable changes in the
chemical shifts for all the methylene resonances of the crown are
observed together with changes in the resonances associated
with the carbons of the substituted metallocene ring. However,
the signal arising from the unsubstituted cyclopentadienyl ring
remains essentially unaffected (±0.05 ppm) throughout the
titration experiment suggesting that it is unperturbed by the
sodium-ion co-ordination process. This observation suggests
that this ring is not included in whatever conformational
processes occur upon the addition of sodium ions, whilst the
substituted cyclopentadienyl ring and the crown ring operate
in concert to effect binding. A possible explanation accounting
for these observations is outlined in schematic form (Fig. 1).
Although the term π complex is normally associated with
transition-metal species, like ferrocene itself, many compounds
that feature π-facial complexation of a ligand to a main-group
element are now known.19 There are examples of cation π inter-
actions within structures containing both π-electron-rich sys-
tems and polyether chains,20 benzene rings are known to be
effective donors for metal cations in a variety of complexes,21–26
and potassium-ion transport has been observed in biological
systems through channels lined with aromatic residues by a
mechanism thought to involve cation π interactions.27 Of
particular relevance to the present discussion is the crystal
structure of [Li(12-crown-4)]ϩ[η5-C5H5]Ϫ (12-crown-4 = 1,4,7,
NC
CN
–
O
N
O
O
N
O
+
Et
Et
M
17, M+ = Li+, Na+, K+, NH4+, Ag+, Ba2+, Ca2+
corresponding picrates. As photographic dyes they have very
high absorption coefficients (≈ 105 dm3 molϪ1 cmϪ1 in water and
organic solvents) thus facilitating simple spectrophotometric
assays. Full experimental detail concerning the extraction tech-
nique is given in the Experimental section. Percentage extrac-
tions for a range of oxonol dyes 17 by metallocene crowns 9, 10,
15, 16 and, for comparative purposes, bis(crown) 1 are given in
Table 3.
Unsurprisingly, the amino-linked crowns 15 and 16 are far
more effective complexing agents than are their amido-linked
analogues 9 and 10. The amide link would be expected both to
limit the flexibility of the crown, and reduce the electron density
associated with the nitrogen donor atom. Tertiary amine
crowns 15 and 16 effect near total transfer of all the mono-
valent cations studied and transfer approximately 50% of the
divalent cations, an extraction efficiency comparable with that
of ferrocene bis(crown) 1. Silver ion binding is far stronger with
these simple crowns, a direct consequence of the presence in
these structures of tertiary amine donor sites as opposed to the
amido nitrogen centres present in 1.
10-tetraoxacyclododecane) which has
a pseudo-sandwich
It is interesting that the extraction behaviour of compounds
15 and 16 is virtually identical throughout the range of metal
ions studied. Of their amido-linked analogues 9 and 10, the
ferrocene derivative 9 consistently extracts a slightly higher per-
centage of metal ion than does the ruthenocene analogue 10.
This may result from greater communication of the more
electron-rich nature of ferrocene9 through the conjugated
amide link to the nitrogen donor atom, thus increasing the
binding strength through increased electrostatic interactions.
arrangement with the Liϩ ion bound on one face by the oxygen
atoms of the crown ether, and on the other by the π face of the
aromatic cyclopentadienyl anion.28 The simple hosts 15 and 16
are the first examples of metallocene-derived crowns which
possess the structural flexibility required to achieve such a dual
binding interaction. The requirement of coplanarity between
the cyclopentadienyl ring and the amide links present in the
majority of reported systems ensures that such a binding mode
is structurally impossible.
The only available data 8 with which to compare our studies is
that for metal-ion binding by the ferrocenylcryptand 4, the
accurate spectral assignment of which was reported by Plenio et
al.1 The 13C NMR binding data presented for cryptand 4 were
used to demonstrate the high level of symmetry present in both
the free cryptand and its metal-ion complexes. The results sug-
gest that this host presents a single well defined binding cavity.
Closer inspection of those data reveals similar trends in the
changes of the OCH2 resonances (1–1.5 ppm upfield) on metal-
ion binding in comparison with those observed with 15 and 16
(1–2.5 ppm upfield). The diazacrown unit of 4 is known to be
rather constrained and the larger shifts observed with the pres-
ent systems are likely to be a reflection of a better binding fit
between the cation and crown. There are also notable differ-
ences in the behaviour of the ferrocenyl carbon resonances. The
quaternary carbons of 4 shift upfield by ≈1.3 ppm in the 4ؒNaϩ
complex whilst larger upfield shifts of ≈1.9 and 2.3 ppm are
observed for 15ؒNaϩ and 16ؒNaϩ. A more dramatic difference
may be noted in the shifts of the resonances of the o-carbons
13C NMR spectroscopic studies of alkali-cation binding
The binding of a metal ion by a host molecule is accompanied
by conformational and electronic changes which, in turn, are
1
reflected by variations in the H and 13C NMR chemical shifts
of those centres close to the site of complexation.16 Since the 1H
NMR spectra of cyclic polyethers are found to be complex and
rather difficult to interpret it has proved more convenient to
follow the binding behaviour of our systems using 13C NMR
spectra. The 13C chemical shifts are known to be sensitive to
conformational changes but relatively less sensitive to solvent
effects, and offer the possibility of using distortionless
enhancement of polarisation transfer (DEPT) experiments to
aid in assignment.17,18
Having identified tertiary amine crowns 15 and 16 as the more
effective ionophores, we performed NMR binding titrations
with Naϩ ions by monitoring variations in the 13C NMR chem-
ical shifts in the presence of varying quantities of NaClO4 in
CD3CN solutions of the crowns. The chemical shifts and peak
J. Chem. Soc., Dalton Trans., 1997, Pages 3471–3477
3473

View Article Online
Table 4 Carbon-13 NMR shifts (δ) of compounds 15 and 16 with added sodium perchlorate
b
c
a
d
e
O
O
N
5
O
O
1
4
f
3
2
M
6
Compound 15
Ratio *
C1
Cb
C2,5
Ca
Cc
Cd
C6
C3,4
Cf
Ce
0.00:1
0.25:1
0.50:1
1.00:1
1.50:1
82.83
82.12
81.54
80.90
80.93
71.45
71.06
70.87
69.62
69.64
71.13
71.43
71.56
71.80
71.79
70.87
70.64
70.04
69.62
69.65
70.72
70.56
70.23
69.03
69.10
70.27
69.52
68.43
67.59
67.59
69.45
69.51
69.48
69.54
69.53
68.81
68.96
69.01
69.25
69.20
56.47
54.78
53.87
52.65
52.66
54.71
54.32
53.06
53.06
53.00
Compound 16
Ratio *
C1
C2,5
C6
C3,4
Ca
Cb
Cc
Cd
Cf
Ce
0.00:1
0.25:1
0.50:1
1.00:1
1.50:1
2.00:1
83.20
82.65
81.54
80.98
80.90
80.91
73.33
73.44
73.62
73.82
73.87
73.88
71.50
71.52
71.55
71.62
71.55
71.55
71.50
71.52
71.22
71.44
71.28
71.30
71.01
71.08
71.52
69.76
69.56
69.57
70.92
70.56
70.25
69.76
69.56
69.57
70.79
70.56
69.99
69.15
69.01
69.02
70.07
69.47
68.65
67.43
67.59
67.57
56.22
55.53
53.55
53.11
52.63
52.63
54.62
54.29
53.82
53.11
53.04
52.04
* Molar ratio of NaClO4 to crown.
(C2,5). For the 4ؒNaϩ complex an upfield shift of 2.3 ppm is
observed; the 15ؒNaϩ and 16ؒNaϩ complexes display downfield
shifts of 0.6–0.7 ppm.
O
O
O
N
O
The 13C NMR spectral changes observed in the conversion of
compound 11 into the quaternary ammonium salt 13 provide a
model for the effect of the introduction of a positive charge at
the nitrogen centre of 15 and 16; this effect may be considered
similar in nature, but far greater in magnitude, to the partici-
pation of the nitrogen centre in cation binding. In the conver-
sion of 11 into 13 the C5H4CH2 resonance shifts around 8 ppm
downfield. This is at variance with the situation observed for the
complexes 15ؒNaϩ and 16ؒNaϩ where upfield shifts of the same
resonance by 3.5–4 ppm are recorded. Such a contrast indicates
that it is unlikely that the cation-induced changes in the spectra
of 15 and 16 are the result of an inductive effect arising from
the binding of a cation in the crown ring.§ To rationalise the
difference in behaviour of the relative C5H4CH2 shifts it is pro-
posed that the nitrogen lone pairs of 15 or 16 do not directly
participate in the binding of the metal cation. Corey–Pauling–
Koltun (CPK) models support this view by demonstrating that
our proposed binding conformation for 15ؒNaϩ (Fig. 1) requires
the nitrogen lone pair to point almost directly away from the
binding cavity.
M
+
M
O
O
N
O
O
< 1 equivalent
> 1 equivalent
+
+
14
O
O
O
O
+
N
Whilst the results of a 13C NMR titration of compound 16
with KClO4 initially seemed inconsistent with the effects
observed in the Naϩ case, more detailed analysis indicated the
presence of a rather delicate binding equilibrium. Unlike the
Naϩ titration, at less than 1 equivalent of added cation no
appreciable shifts of any of the ferrocenyl resonances are appar-
ent whereas the shifts for the methylene carbons are broadly
consistent with those reported for the potassium complex of 4.
These observations may indicate that at crown :Kϩ ratios of
<1:1 co-ordination is dominated by 2:1 sandwich host–guest
M
Fig. 2 Equilibria for the binding of Kϩ by compound 15
complexation (Fig. 2). Behaviour of this kind has previously
been noted with other metallocene monocrowns, especially
those derived from 15-crown-5 (1,4,7,10,13-pentaoxacyclo-
pentadecane),29 a crown ether which has a well documented
ability to form 2:1 sandwich type complexes with potassium
salts. No appreciable shifts of the ferrocenyl resonances would
be expected for a complex of this type since, unlike cryptand
4 where the binding site is ‘tied’ to both ferrocene rings, no
conformational reorganisation involving the ferrocene unit is
required. On the addition of an excess of KClO4 (crown :Kϩ
>1:1) the shifts become consistent with the model proposed for
15ؒKϩ, the host simply switching binding modes in response to
the increased concentration of Kϩ ions present. The proposed
binding equilibrium is presented schematically in Fig. 2.
§ The UV/VIS spectral investigation of compound 15, its cation com-
plexes and the model compounds 11 and 13 was undertaken in
acetonitrile solution. In the case of the model compounds a hypso-
chromic shift of around 7 nm was observed in the conversion of 11
(λmax = 404 ± 1 nm) into 13 (398 ± 1 nm). In contrast, no cation-
induced shifts were observed upon the addition of lithium, sodium
or potassium perchlorate to solutions of ferrocene crown 15
(λmax = 404 ± 1 nm in all cases).
3474
J. Chem. Soc., Dalton Trans., 1997, Pages 3471–3477

View Article Online
crown is increased the oxidation potential of the free crown
gradually gives way to that of the complex. The measured shifts
Table 5 Electrochemical oxidation potentials (mV) for compound 15
with added NaPF6 or KPF6
؉
؉
(∆EN a = 100, ∆EK = 80 mV) are comparable with those
obtained for the ferrocene analogue 15. The invariable value of
the second oxidation potential (E2) may be rationalised by
recognising that the first oxidation process generates a formally
charged metallocene centre [RuIII(C5H5)2]ϩ, and, paralleling the
behaviour reported for ferrocene systems,35,36 the positively
charged metallocene derivative then ejects the bound cation as a
result of unfavourable electrostatic interaction. Consequently,
since the cation is no longer associated with the host, the value
of the second oxidation potential remains constant throughout
the experiment.
Oxidation potential (mV)
؉
؉
Molar ratio *
EN a
EK
0.0:1
0.5:1
1.0:1
480
520
560
480
510
530
* Molar ratio MPF6 to crown.
Table 6 Electrochemical oxidation potentials (mV) for compound 16
with added NaPF6 or KPF6
Oxidation potential (mV)
Conclusion
Naϩ
Kϩ
The work described in this paper allows two observations to
be made concerning the properties of simple metallocene
azacrown ethers: (i) the presence of an amino-, rather than
an amido-, link between the metallocene and azacrown ether
greatly enhances cation binding ability, and (ii) in solution,
simple amino-linked metallocene–azacrown ethers may have
available a binding mode involving a stabilising π interaction
between a guest cation and the π system of the metallo-
cene. Few previously reported metallocene crowns possess
the conformational or geometrical freedom to adopt such a
co-ordination mode.
Molar ratio a
E1
E1Ј
E1
E1Ј
0.00:1
0.25:1
0.50:1
0.75:1
1.00:1
420
420
420
420
b
b
b
520
520
520
420
420
b
b
b
b
490
490
490
490
E2 = 620 mV in each case. Molar ratio MPF6 to crown. b Not
a
observed.
The possibility of π participation in cation binding by com-
pounds 15 and 16 is an intriguing addition to the known
modes of cation complexation adopted by ferrocene crowns
and cryptands. The original ferrocene-based cryptand 3 is
known to bind cations both internally, within the crown res-
idues, and externally, via two cis-oriented amide-carbonyl
groups.6 Both situations are also observed with ferrocene bis-
(crowns) such as 2.7 Gokel and co-workers8 have shown that
the amino-linked cryptand 4 has only one well defined binding
cavity, and in addition may allow direct d-orbital interaction
between the ferrocenyl iron centre and a guest cation such as
silver. The present work suggests that in sufficiently flexible
systems the π system of the metallocene may also act as a
donor group for a bound cation, though in the absence of
solid-state structural data it is not possible to provide absolute
proof for this hypothesis, or to discount alternative interpret-
ations. Further investigation of the nature and importance of
this interaction in these and related systems is currently in
progress.37
Cyclic voltammetry
The electrochemical response of metallocenylcrowns and cryp-
tands to the presence of a guest species has always been the
primary motivation behind their design and synthesis.5 The
groups of Saji30 and Hall31 were the first to report a correlation
between the shift of the ferrocene–ferrocenium redox couple
and the charge–radius ratio of the guest cation. More recently
Plenio and co-workers32,33 have demonstrated
a relation
between the spatial separation of the ferrocenyl iron centre
from the bound guest and the redox potential of the complex.
Such analysis holds for redox switching of sodium ion binding32
and protonation of a number of ferrocene amines, crowns and
cryptands.33
Cyclic voltammetric analysis (see Experimental section) of
compound 15 in acetonitrile reveals the expected one-electron
reversible oxidation process at 480 mV. The changes observed
on the addition of increasing amounts of NaPF6 and KPF6 are
shown in Table 5. The addition of stoichiometric equivalents of
these salts to 15 resulted in anodic shifts (∆E) of the ferrocene–
ferrocenium redox couple of 80 and 50 mV, respectively.
These shifts appear rather small when compared with those
Experimental
؉
؉
reported for ferrocenylcryptand 4 (∆EN a = 188, ∆EK = 124
mV),8 where the cation is of necessity held in close proximity
to the iron centre, but are similar to the electrochemical
؉
General
Column chromatography was performed on flash silica
(Rhône-Poulenc, Sorbsil C60-H, 40–60 mesh) or alumina
(Fluka, Type 507 C Neutral, Brockmann Grade 1, 100–125
mesh). All melting points were determined in open capillary
tubes using a Gallenhamp Electrothermal melting point appar-
atus and are uncorrected. Combustion analyses were performed
by Medac Ltd., Brunel University. Fast atom bombardment
(FAB) mass spectra were recorded on a VG Analytical normal
geometry double-focusing mass spectrometer, UV/VIS spectral
measurements on Perkin-Elmer 550S and 551S spectro-
photometers, IR spectra using a 1600 Series Perkin-Elmer
Fourier-transform spectrophotometer and NMR spectra on
JEOL 270 MHz and Bruker 360 MHz spectrophotometers.
Cyclic voltammetric electrochemical measurements were made
using unstirred solutions of the compounds in dry acetonitrile
responses reported for the ferrocenylbis(crowns) 1 (∆EN a
=
70 mV)7 and 2 (∆EN a = 90, ∆EK = 60 mV). The responses
obtained for 15 indicate that there is significant cation redox
interaction and that, if our proposed binding model is valid
under the conditions of the electrochemical studies, then metal
ion co-ordination to the π face of the metallocene provides
effective communication of the binding event to the redox
centre.
3
؉
؉
Ruthenocene derivatives are known to undergo two distinct,
irreversible one-electron oxidation processes.34 Such a lack of
reversibility has ensured that no electrochemical investigation
of cation binding by a ruthenocene-derived crown has been
reported. Accordingly, the electrochemical behaviour of com-
pound 16 in the presence of varying amounts of NaPF6 and
KPF6 has been studied and the results are summarised in Table
6. The first and second oxidation potentials of the free crown
are represented by E1 and E2, the first oxidation potential of the
cation complex being denoted by E1Ј. As the ratio of cation to
(5 m) containing NBun PF6 (30 m) as the supporting electro-
4
lyte using a custom-built potentiostat with platinum working
and auxiliary electrodes, and a standard calomel reference
electrode.
J. Chem. Soc., Dalton Trans., 1997, Pages 3471–3477
3475

View Article Online
Preparations
(2 × 50 cm3) and dried in vacuo (1.41 g, 38%), m.p. 155–170 ЊC
(decomp.) (Found: C, 38.77; H, 4.61; N, 3.26. C14H20INRu
requires C, 38.98; H, 4.68; N, 3.22%); ν max/cmϪ1 (CDCl3) 2974w,
2888w, 2739w, 1814w, 1787w, 1477s, 1461s, 1381s, 1295w,
1247w, 1210w, 1167w, 1092m, 1044m and 980w; δC[75 MHz,
(CD3)2SO] 76.83 (C1), 74.07 (C2,5), 72.23 (C3,4), 71.43 (C5H5),
65.00 (CH2) and 44.34 (CH3); δH (270 MHz, CDCl3) 4.89 (2 H,
d, J = 2, H2,5), 4.70 (2 H, d, J = 2 Hz, H3,4), 4.66 (5 H, s, C5H5),
4.54 (2 H, s, CH2) and 3.36 (9 H, s, CH3); FAB mass spectrum:
m/z 430 (Mϩ, 100%).
1,4,7,10-Tetraoxa-13-azacyclopentadec-13-ylcarbonylferro-
cene 9. A suspension of ferrocenecarboxylic acid 38 5 (500 mg,
2.17 mmol) in dry CH2Cl2 (15 cm3) containing dry dmf (2
drops) was treated under N2, with stirring, with a solution of
oxalyl chloride (2.21 g, 1.52 cm3, 17.4 mmol) in dry CH2Cl2 (15
cm3) and the resulting mixture stirred overnight. The solvent
was then removed under reduced pressure and the residue
recrystallised from dry hexane to yield dark red plates of chloro-
carbonylferrocene 7 (520 mg, 95%). A solution of aza-15-
crown-539,40 (132 mg, 0.6 mmol) and dry triethylamine (66 mg,
0.09 cm3, 0.66 mmol) in dry benzene was treated dropwise with
a solution of chlorocarbonylferrocene (150 mg, 0.6 mmol) in
dry benzene (10 cm3) over 10 min. The resulting mixture was
stirred for a further hour and then filtered and the solvent
removed under reduced pressure. The residual oil was purified
by column chromatography (neutral Al2O3; CH2Cl2 then 2%
MeOH in CH2Cl2; Rf = 0.60, 1% MeOH in CH2Cl2) to afford
an orange oil (254 mg, 98%); ν max/cmϪ1 (CDCl3) 2900m, 1603s,
1474s, 1410m, 1381m, 1295w, 1264w, 1216w, 1127s and 1098s;
N-Ferrocenylmethyl-1,4,7,10-tetraoxa-13-azacyclopenta-
decane 15. A solution of trimethylammoniomethylferrocene
iodide43 13 (1.03 g, 2.68 mmol) in dry acetonitrile (40 cm3)
was added dropwise over 1 h to a refluxing solution of aza-15-
crown-5 (420 mg, 1.92 mmol) in dry acetonitrile (40 cm3) con-
taining anhydrous potassium carbonate (874 mg, 6.32 mmol)
and the mixture refluxed for 6 h. After cooling the mixture was
filtered and the solids washed with hot acetonitrile (2 × 20 cm3).
The solvent was removed from the organic extracts under
reduced pressure, the residue dissolved in CH2Cl2 (50 cm3) and
washed with water (2 × 100 cm3). The organic extracts were
then dried over MgSO4 and evaporated to yield an orange oil
which was purified by column chromatography (neutral Al2O3;
2% MeOH in CH2Cl2; Rf = 0.40, 2% MeOH in CH2Cl2) to
afford an orange oil (640 mg, 80%); ν /cmϪ1 (CDCl3) 3097w,
2875s, 1656m, 1561w, 1468s, 1381s, 1357m, 1254m, 1264m,
1126vs, 1042m and 1001m; δC(75 MHz, CDCl3) 71.61
(C2,5), 70.12 (C3,4), 69.97, 69.69, 69.51, 69.34 (OCH2), 69.22
(C5H5), 64.70 (OCH2), 55.50, 53.60 (CH2N) and 52.82
(C5H4CH2N); δH (270 MHz, CDCl3) 4.53 (2 H, s, H2,5), 4.50
(2 H, s, C5H4CH2N), 4.29 (2 H, s, H3,4), 4.25 (5 H, s, C5H5), 3.68
(16 H, m, OCH2) and 3.49 (4 H, br s, NCH2); FAB accurate
mass: m/z 418.1673 (Found), C21H32FeNO4 ([M ϩ H]ϩ) requires
418.1681.
δ (75 MHz, CDCl ) 170.80 (C᎐O), 77.72 (C1), 71.77 (CH O),
᎐
C
3
2
70.72 (C2,5), 70.34, 70.24 (C3,4 and OCH2), 69.79 (C5H5), 50.80,
50.08 (NCH2); δH (270 MHz, CDCl3) 4.67 (2 H, s, H2,5), 4.29 (2
H, s, H3,4), 4.21 (5 H, s, C5H5), 3.80–3.60 (20 H, m, CH2); FAB
accurate mass: m/z 432.1432 (Found), C21H30FeNO5 ([M ϩ H]ϩ)
requires 432.1473.
1,4,7,10-Tetraoxa-13-azacyclopentadec-13-ylcarbonyl-
ruthenocene 10. Treatment of a suspension of ruthenocene-
carboxylic acid 41 6 (500 mg, 1.82 mol) in dry CH2Cl2 (15 cm3)
containing dry dmf (2 drops) with a solution of oxalyl chloride
(1.84 g, 1.27 cm3, 14.5 mmol) in dry CH2Cl2 (15 cm3) and stir-
ring overnight under N2 afforded the acid chloride derivative 8
after solvent evaporation and recrystallisation from hexane
(482 mg, 90%). A solution of aza-15-crown-5 (187 mg, 0.85
mmol) and dry triethylamine (95 mg, 0.13 cm3, 0.94 mmol)
in dry benzene (10 cm3) was treated dropwise with a solution
of chlorocarbonyl ruthenocene (250 mg, 0.85 mmol) in dry
benzene (10 cm3) over 10 min and then stirred for a further
hour. The mixture was then filtered, evaporated under reduced
pressure and purified by column chromatography (neutral
Al2O3; CH2Cl2 then 2% MeOH in CH2Cl2; Rf = 0.65, 1%
MeOH in CH2Cl2) to afford a pale yellow oil (390 mg, 96%);
ν max/cmϪ1 (CDCl3) 2868m, 1600s, 1477s, 1410m, 1380m,
1354m, 1293m, 1250w, 1127s and 1100s; δC(75 MHz, CDCl3)
N-Ruthenocenylmethyl-1,4,7,10-tetraoxa-13-azacyclopenta-
decane 16. Following the procedure given above for the ferro-
cene analogue, reaction of aza-15-crown-5 (330 mg, 1.50 mmol)
with compound 14 (712 mg, 1.66 mmol) in dry acetonitrile (80
cm3) containing anhydrous potassium carbonate (684 mg, 4.95
mmol) afforded the crude crown. Column chromatography
(neutral Al2O3; 2% MeOH in CH2Cl2; Rf = 0.42, 2% MeOH
in CH2Cl2) provided the pure crown derivative (580 mg,
83%); ν max/cmϪ1 (CDCl3) 2900s, 1640w, 1561w, 1468s, 1381s,
1295w, 1248w, 1216w and 1098vs; δC(75 MHz, CDCl3)
77.40 (C1), 72.58 (C2,5), 71.07 (C3,4), 70.72 (C5H5), 70.52, 70.40,
70.26, 69.70 (OCH2), 55.78 (C5H4CH2N) and 53.60 (NCH2);
δH (270 MHz, CDCl3) 4.59 (2 H, s, H2,5), 4.50 (5 H, s, C5H5),
4.48 (2 H, s, H3,4), 3.64 (16 H, br s, OCH2), 3.40 (2 H, s,
C5H4CH2N) and 2.80 (4 H, t, J = 6 Hz, NCH2); FAB accurate
mass: m/z 464.1339 (Found), C21H32NO4Ru ([M ϩ H]ϩ)
requires 464.1375.
169.60 (C᎐O), 81.40 (C1), 72.85 (C2,5), 72.00 (C3,4), 71.79
᎐
(C5H5), 70.31, 70.21 (CH2O), 51.26, 49.90 (CH2N); δH (270
MHz, CDCl3) 4.99 (2 H, apparent t, J = 2, H2,5), 4.62 (2 H,
apparent t, J = 2 Hz, H3,4), 4.58 (5 H, s, C5H5), 3.74 (4 H, br s,
NCH2), 3.53 and 3.51 (16 H, 2 × s, OCH2); FAB accurate mass:
m/z 478.1129 (Found), C21H30NO5Ru ([M ϩ H]ϩ) requires
478.1167.
Trimethylammoniomethylruthenocene iodide 14. A mixture
of ruthenocene42 (2.00 g, 8.66 mmol), N,N,NЈ,NЈ-tetramethyl-
diaminomethane (1.49 g, 2.00 cm3, 14.0 mmol), orthophos-
phoric acid (1.5 cm3 of a 85% aqueous solution) and acetic acid
(25 cm3) was heated at 100 ЊC for 5 h with vigorous stirring
under a N2 atmosphere. The reaction was allowed to cool and
diluted with water (50 cm3); extraction with diethyl ether
(2 × 50 cm3) recovered unchanged ruthenocene (1.14 g, 57%).
The aqueous extracts were cooled in ice and made alkaline with
the addition of NaOH pellets (20.0 g, 0.5 mol). Extraction with
diethyl ether (2 × 50 cm3) afforded a small volume of a pale
yellow oil. This crude oil was dissolved in methanol (25 cm3),
methyl iodide (3.70 g, 1.62 cm3, 0.026 mol) added slowly and
the resulting solution subsequently refluxed for 30 min before
being cooled to room temperature and poured dropwise into
stirred diethyl ether (100 cm3). The precipitated pale yellow
solid was collected at the pump, washed with diethyl ether
Calculation of amide bond-rotation barriers from 13C NMR
spectra
The Gutowsky44 and Eyring45 equations may be combined to
yield equation (1) where kC = ∆ν(π/√2). Activation energies
∆G‡ = (19.148)Tc(10.319 Ϫ log10 kc ϩ log10 Tc) (1)
(∆G‡) may thus be calculated from knowledge of the coal-
escence temperature (Tc) of the exchange process and the
frequency separation (∆ν) of the individual resonances in the
frozen condition. The quoted errors reflect the inherent dif-
ficulty of accurately determining both the coalescence tem-
perature (a change of 1 K results in a shift of around 0.2 kJ
molϪ1), and the coalescence point (computational lineshape
spectral analysis would be required accurately to determine this
parameter).
3476
J. Chem. Soc., Dalton Trans., 1997, Pages 3471–3477

View Article Online
13 M. Newcomb, J. M. Timko, D. M. Walba and D. J. Cram, J. Am.
Extraction experiments with oxonol dye salts4
Chem. Soc., 1977, 99, 6392.
Oxonol dye salts were prepared according to literature
methods46 and recrystallised to constant absorption coefficient
prior to use in the extraction experiments. Typically two to
three recrystallisations of each dye salt were required. Absorp-
tion coefficients (εmax) for each dye salt in water were deter-
mined in the concentration range 10Ϫ6 to 10Ϫ4  of standard
solutions prepared by serial dilutions of a 5 × 10Ϫ4  solution
of the dye in deionised water. The average value over this range
was used in all calculations. Equal volumes (10 cm3) of equi-
molar (1.5 × 10Ϫ5 ) solutions of oxonol dye (in water) and a
CH2Cl2 solution of crown were shaken mechanically at ambient
temperature for 10 min in sealed vessels. Each two-phase system
was then allowed to separate for 10 min. Taking care to avoid
solvent evaporation, aliquots of the aqueous layer were then
removed and the absorbance of the sample at 589 nm deter-
mined; each measurement was repeated five times. The percent-
age of oxonol dye remaining in the aqueous layer was given by
(A/A0) × 100, where A0 is the absorbance of the aqueous layer
prior to extraction and A that following extraction. Each
determination was repeated five times.
14 S. S. Moore, T. L. Tarnowski, M. Newcomb and D. J. Cram, J. Am.
Chem. Soc., 1977, 99, 6398.
15 K. E. Koenig, G. M. Lein, P. Stuckler, T. Koneda and D. J. Cram,
J. Am. Chem. Soc., 1979, 101, 3553.
16 D. Live and S. I. Chan, J. Am. Chem. Soc., 1986, 98, 3769.
17 F. W. Wehrli, A. P. Marchand and S. Wehrli, Interpretation of
Carbon-13 NMR Spectra, Wiley, Chichester, 1988.
18 J. B. Stothers, Carbon-13 NMR Spectroscopy, Academic Press, New
York, 1972.
19 P. Jutzi, Adv. Organomet. Chem., 1986, 26, 217.
20 J. M. Larson and L. R. Souson, J. Am. Chem. Soc., 1978, 100, 1943.
21 F. Vögtle and W. Kissener, Chem. Ber., 1984, 117, 2538.
22 C. Cohen-Addad, P. Baret, P. Chautemps and J.-L. Pierre, Acta
Crystallogr., Sect. C, 1983, 39, 1346.
23 H. Schmidbauer, R. Hager, B. Huber and G. Müller, Angew. Chem.,
1987, 99, 34.
24 C. Elschenbroich, J. Schneider, M. Wünsch, J.-L. Pierre, P. Baret
and P. Chautemps, Chem. Ber., 1988, 121, 177.
25 H. Bock, K. Ruppert, Z. Haulas and D. Fenske, Angew. Chem., Int.
Ed. Engl., 1990, 29, 1042.
26 J. A. Samuels, E. B. Lobkovsky, W. E. Streib, K. Folting,
J. Huffman, J. W. Zwanziger and K. G. Caulton, J. Am. Chem. Soc.,
1993, 115, 5093.
27 C. Miller, Science, 1991, 252, 1092.
28 P. Jutzi, Organometallics, 1991, 10, 1282.
Acknowledgements
29 P. D. Beer, H. Sikanyika, C. Blackburn, J. F. McAleer and
M. G. B. Drew, J. Chem. Soc., Dalton Trans., 1990, 3295.
30 T. Saji, Chem. Lett., 1989, 275.
31 C. D. Hall, N. W. Sharpe, I. P. Danks and Y. P. Song, J. Chem. Soc.,
Chem. Commmun., 1989, 419.
32 H. Plenio and R. Diodone, Inorg. Chem., 1995, 34, 3964.
33 H. Plenio, J. Yang, R. Diodone and K. Heinze, Inorg. Chem., 1994,
33, 4098.
We thank the former SERC for funding two studentships
(J. I. F. and D. G. H.), Mrs J. M. Street and Dr. G. J. Langley
(University of Southampton) for recording NMR and mass
spectra respectively, and Johnson Matthey plc for a generous
loan of ruthenium trichloride.
34 M. Rosenblum, Chemistry of the Iron Group Metallocenes,
Interscience, New York, 1965.
35 T. Saji and I. Kinoshita, J. Chem. Soc., Chem. Commun., 1986, 716.
36 T. Saji, K. Hoshino and S. Aoyagui, J. Chem. Soc., Chem.
Commun., 1985, 865.
37 M. C. Grossel and J. I. Fuller, unpublished work.
38 P. C. Reeves, Org. Synth., 1978, 56, 28.
39 M. R. Johnson, I. O. Sutherland and R. F. Newton, J. Chem. Soc.,
Perkin Trans. 1, 1979, 357.
40 R. A. Schultz, B. D. White, D. M. Dishong, K. A. Arnold and
G. W. Gokel, J. Am. Chem. Soc., 1985, 107, 6659.
41 S. I. Goldberg, L. H. Keith and T. S. Prokopov, J. Org. Chem., 1963,
28, 850.
42 P. Pertici and G. Vitulli, Inorg. Synth., 1983, 22, 180.
43 D. Ledricer and C. R. Hauser, Org. Synth., 1960, 40, 31.
44 H. S. Gutowsky and C. H. Holm, J. Chem. Phys., 1956, 25, 1228.
45 S. Gladston, K. J. Laidler and H. Eyring, The Theory of Rate
Processes, McGraw-Hill, New York, 1941.
References
1 H. Plenio, H. El-Desoky and J. Heinze, Chem. Ber., 1933, 126, 2403.
2 P. D. Beer, Z. Chen and M. I. Ogden, J. Chem. Soc., Faraday Trans.,
1995, 295.
3 P. D. Beer, Z. Chen, M. G. B. Drew and A. J. Pilgrim, Inorg. Chim.
Acta, 1994, 225, 137.
4 M. C. Grossel, M. R. Goldspink, J. A. Hriljac and S. C. Weston,
Organometallics, 1991, 10, 851.
5 P. D. Beer, Chem. Soc. Rev., 1989, 18, 409.
6 C. D. Hall, J. H. R. Tucker and N. W. Sharpe, Organometallics, 1991,
10, 1727.
7 P.
D Beer, A. D. Keefe, H. Sikanyika, C. Blackburn and
J. F. McAleer, J. Chem. Soc., Dalton Trans., 1990, 3289.
8 J. C. Medina, T. T. Goodnow, M. T. Rojas, J. L. Atwood, B. C. Lynn,
A. E. Kaifer and G. W. Gokel, J. Am. Chem. Soc., 1992, 114, 10 583.
9 M. D. Rausch, E. O. Fischer and H. Grubert, J. Am. Chem. Soc.,
1960, 82, 76.
46 M. C. Grossel, D. J. Edwards, A. K. Cheetham, M. R. Eddy,
O. Johnson and S. R. Postle, J. Mater. Chem., 1991, 1, 223.
10 W. Kemp, NMR in Chemistry,
A Multinuclear Introduction,
Macmillan, Basingstoke, 1986, p. 158.
11 J. P. Knychala, Ph.D. Thesis, University of London, 1987.
12 E. P. Kyba, R. C. Helgeson, K. Madan, G. W. Gokel, T. L.
Tarnowski, S. S. Moore and D. J. Cram, J. Am. Chem. Soc., 1977, 99,
2564.
Received 23rd May 1997; Paper 7/03597G
J. Chem. Soc., Dalton Trans., 1997, Pages 3471–3477
3477