Synthesis and characterization of a ferrocene-linked bis-fullerene[60] dumbbell

Article abstract of DOI:10.1039/c2dt12097f

A new [60]fullerene dumbbell consisting of two fulleropyrrolidines connected to a central ferrocene unit by amide linkages has been prepared and fully characterized by elemental analysis, ¹H NMR, UV/Vis, fluorescence and mass spectrometry. The electrochemical properties as determined by cyclic voltammetry show ground state electronic communication between the ferrocene and the fullerene units. In addition, the preparaton of a ferrocene building block for an alternative linking approach is presented. The Royal Society of Chemistry 2012.

Full text of DOI:10.1039/c2dt12097f

View Article Online / Journal Homepage / Table of Contents for this issue
Dynamic Article Links
Dalton
Transactions
Cite this: Dalton Trans., 2012, 41, 2374
www.rsc.org/dalton
PAPER
Synthesis and characterization of a ferrocene-linked bis-fullerene[60]
dumbbell†
Claes-Henrik Andersson,a Leif Nyholmb and Helena Grennberg*a
Received 3rd November 2011, Accepted 22nd November 2011
DOI: 10.1039/c2dt12097f
A new [60]fullerene dumbbell consisting of two fulleropyrrolidines connected to a central ferrocene unit
by amide linkages has been prepared and fully characterized by elemental analysis, 1H NMR, UV/Vis,
fluorescence and mass spectrometry. The electrochemical properties as determined by cyclic
voltammetry show ground state electronic communication between the ferrocene and the fullerene
units. In addition, the preparaton of a ferrocene building block for an alternative linking approach is
presented.
of dumbbells may hence find great use in molecular electronics,
Introduction
with C60 end groups increasing the still limited set of anchoring
Fullerenes are frequently employed as electron acceptor moieties
groups, and moreover rendering possible studies of conductivity
by direct methods.6
in donor–acceptor systems, with C60 standing out as the most
studied and well characterized member of the family. In the
ground state, C60 can accept up to six electrons in a reversible
manner,1 and numerous studies of donor–acceptor dyad systems
have shown that a large variety of C60 derivatives are acceptors for
photoinduced electron transfer.2
In a previous study, we found evidence for ground state
electronic communication between C60 and an appended arene
tricarbonyl chromium unit, with the possiblility of determining the
redox state of the C60 by IR spectroscopy.7 This property would,
in a dumbbell or larger structure, be useful as i.e. a molecular
memory. Unfortunately, the stability of the tricarbonylchromium-
arene decreases in the presence of C60 and although it was possible
to isolate and fully characterize the dyad, the corresponding
dumbbell was not stable enough.8 In contrast, the isoelectronic
ferrocene is stable under a large range of conditions, and
organometallic dyads with C60 as acceptor and ferrocene as donor
have been thorougly characterized.9 Dumbbells and higher C60
oligomers with ferrocenyl links could thus be a class of compounds
with interesting properties and potential application in molecular
devices.
In the present paper, we report synthetic studies towards
two new C60-donor-C60 dumbbells 1 and 2 (Fig. 1), as well
as characterization of 1 by NMR, UV/Vis and fluorescence
spectroscopy, elemental analysis and mass spectrometry. The
electrochemical properties of 1 have further been investigated
by cyclic voltammetry. The route towards 2, with several new
ferrocene derivatives, is also described.
Molecular derivatives consisting of more than one fullerene
unit, such as the group of dimers resembling dumbbells, are less
common. A C60-piperazine-C60 dumbbell was reported by Mart´ın
et al. already in 1996.3 Electrochemical characterization showed
that the compound behaved as two independent donor–acceptor
units with no observable interactions between the two fullerenes
in the ground state. Further examples of dumbbell systems with
donor fragments such as porphyrins, oligo(phenyleneethynylenes)
and tetrathiafulvalenes have been reported and studied in the
literature. While most of the reported C60-donor-C60 triads have
shown a lack of electronic interactions between the moieties in the
ground state, some have shown improved photophysical properties
in the excited state as compared to the corresponding dyads (i.e
stabilization of charge separated states upon electron transfer) and
photovoltaic devices based on some of these C60 dumbbells have
shown improved energy conversion efficiencies.4
It has been shown that C60 has an affinity for noble metals
and hybridizes efficiently with gold surfaces, and recent single-
molecular conductivity measurements have employed C60 as a
highly efficient anchoring group to metal electrodes.5 These types
Results and discussion
Both dumbbells have a central 1,1¢-functionalized ferrocene unit
linked to N-substituted fulleropyrrolidine units. In 1, the link
is via amides formed between a ferrocene acid chloride and
fulleropyrrolidines in the final step of the synthesis, whereas
the tertiary amine link in 2 is set up already in the ferrocene-
containing dipole precursor 12. The reason for this general
design is twofold: (i) It has been shown that there is good
aUppsala University, Dept. of Biochemistry and Organic Chemistry, Box
576, SE-751 23, Uppsala, Sweden. E-mail: Helena.Grennberg@Biorg.uu.se;
Fax: +46-18-4713818
bDept. of Materials Chemistry, Uppsala University, Box 538, SE-751 21,
Uppsala, Sweden
† Electronic supplementary information (ESI) available. See DOI:
10.1039/c2dt12097f
2374 | Dalton Trans., 2012, 41, 2374–2381
This journal is
The Royal Society of Chemistry 2012
©

View Article Online
Fig. 2 Reference compounds used in this study N-acyl-2-(3,4-bis-
(octadecyloxy)benzene)-3,4-fulleropyrrolidine 6 and 1,1-bis(pyrrolidinyl-
carbonyl)ferrocene 7.
Fig. 1 Target structures 1 and 2.
compound 9 in a purified yield of 86%. Deprotection using K2CO3
gave alkyne 10 in 74% yield. The ferrocene component 11 for the
second Sonogashira coupling was prepared in three steps. Fer-
rocene was treated with n-butyl lithium and TMEDA followed by
reaction with tributyltin chloride to yield 1,1¢-bis(tributyl)stannyl
ferrocene,15 which was subsequently reacted with iodine to give
1,1¢-diiodoferrocene 11 in 70% purified yield.16 The assembly of
12 was achieved by a second Sonogashira coupling between an
excess of 10 and compound 11 to give compound 12 as an orange
solid in a purified yield of 39%. Hydrolysis of 12 gave compound
13 in 55% isolated yield.
electronic communication between the pyrrolidine nitrogen and
the fullerene core in fulleropyrrolidines10 and (ii) the possibility
to efficiently functionalize the fulleropyrrolidine nitrogens by
organometallic groups could allow for controlled preparation
of trimers, tetramers, pentamers and polymers by the use of
fulleropyrrolidine bis-adducts as building blocks, accessible via
established tether-based approaches.11
Synthesis
The final step to the target dumbbell 2 would be a dual Prato
reaction between compound 13, 4-methoxybenzaldehyde and C60.
The reaction was attempted several times, in refluxing
o-dichlorobenzene for up to two days without any indication
of pyrrolidine formation. Inspection of molecular models indi-
cated that, for a conformation of 13 with parallel substituents,
intersubstituent hydrogen bonding may deactivate the secondary
amine towards the initial imine condensation required for dipole
formation. This interaction is not present in diester 12. Although
pyrrolidination of C60 has been reported for i.e. ester derivatives
of sarcosine,17 no traces of any pyrrolidinated product could be
detected from our attempts using 12, 4-methoxybenzaldehyde
and C60.
In the convergent synthesis of 1 (Scheme 1), the fulleropyrrolidine
component is prepared in a 1,3-dipolar cycloadditon of an in situ
formed azomethine ylide from 3,4-bis(octadecyloxy)benzaldehyde
and glycine to one of the double bonds of C60. In order to facilitate
purification and characterization, an aldehyde with long-chain
aliphatic groups was employed. Fulleropyrrolidine 5 was isolated
in 39% yield, and in contrast to most N-unsubstituted analogues,
it was readily purified by (repeated) column chromatography
on silica. The 1,1¢-ferrocenedicarboxylic acid 3 was synthesized
according to literature procedures12 and converted to its di-acid
chloride 4 by oxalyl chloride. Fulleropyrrolidine 5 and compound
4 were then reacted in stochiometric amounts in the presence of
pyridine. Purification by precipitation in toluene/EtOH followed
by column chromatography gave the target compound in 69%
isolated yield. Elemental analysis as well as mass spectrometry
(Maldi-TOF) and 1H NMR are consistent with a pure compound.
As expected, 1 is soluble in solvents such as CH2Cl2, CHCl3,
toluene and benzene. In addition to 1, reference compounds 6
and 7 (Fig. 2) were prepared.
A route towards the related dumbbell 2 is shown in Scheme 2.
This target consists of two fulleropyrrolidines connected to a
central ferrocene moiety by ethynylbenzene bridges and thus
constitutes a more linear dumbbell, although the ferrocene unit
is farther away from the fullerenes than in 1.
The initial steps, involving construction of the modified fer-
rocene building block were successful. The dipole precursor end
10 was obtained in three steps, starting with alkylation of p-
iodoaniline with ethyl-bromoacetate in the presence of sodium
acetate to produce compound 8 in a purified yield of 48%.13 This
compound was then reacted with ethynyltrimethylsilane under
microwave-assisted Sonogashira coupling conditions14 to give
Characterization of 1
The IR spectrum of compound 1 (Fig. 3) shows signals similar
to fulleropyrrolidine 5 together with a typical amide signal at
1634 cm-1. A strong signal originating from C60 can be observed
at 526 cm-1 in both compounds. The signals from the ferrocene
unit should appear in the region between 1100–800 cm-1, but
these are probably buried beneath the strong signals from the
fulleropyrrolidines.
More distinct support that we have indeed obtained 1 was
obtained from Maldi-TOF analysis. Several matrices were tried
in both positive and negative detection mode. For a-cyano-4-
hydroxycinnamic acid, no molecular ion peak was observed in ei-
ther of the modes. In the positive mode, neither 9-nitroanthracene
nor dithranol gave any molecular ion peak, instead a major,
non-identified peak at m/z = 2902 together with some frag-
ment ions was obtained. However, in negative mode, both 9-
nitroanthracene and dithranol gave abundant molecular ion peaks
This journal is
The Royal Society of Chemistry 2012
Dalton Trans., 2012, 41, 2374–2381 | 2375
©

View Article Online
Scheme 1 Synthesis of dumbbell 1. Conditions: i) a) TMEDA, n-butyl lithium, n-pentane, RT, stirring, 12 h, b) THF, CO2(g), purified yield 41%. ii)
Oxalyl chloride, CH2Cl2, DMF, stirring, 1.5 h. iii) glycine, o-dichlorobenzene, reflux, 12 h, purified yield: 39%. iv) CH2Cl2, pyridine, stirring, 12 h, purified
yield: 69%.
1.25 and 0.88 ppm. In addition, there are several unresolved broad
signals in the region between 6.5 and 4.5 ppm (Fig. 4) as well as
in the aromatic region, indicating that some dynamic process is
taking place on the NMR timescale. This is not too surprising,
since both the rotation around the central ferrocene unit as well
as the rotation around the two amide bonds can be slow on the
NMR timescale and lead to such broadening of signals.18 It is
also important to consider that since the dumbbell 1 has two
stereogenic centers and is prepared from a racemic mixture of
fulleropyrrolidine 5 as the starting material, a mixture of four
diasteromeric products is expected. Based on this we attempted
some low temperature 1H NMR measurements to determine
whether we could observe individual signals for the different
diastereomers. Unfortunately, precipitation of the compound at
temperatures below –20 ◦C (CD2Cl2) prohibited us from slowing
down the dynamics sufficiently to get any resolution of the signals
and consequently to carry out this study.
However, by changing the solvent to benzene-d6 and increasing
the temperature to 70 ◦C we managed to get a 1H NMR spectrum
of improved quality with significantly sharper signals (Fig. 4
and ESI†). At this temperature the spectrum shows signals for
a single product only, with two sets of aromatic signals (6H),
four pyrrolidine doublets at 6.64, 6.37, 5.80, 5.78 ppm (4H), four
resolved multiplets (J = 1.2 Hz) at 5.39, 5.31, 5.18, 5.12 ppm and
four unresolved, partly overlapping multiplets at 4.41, 4.38 and
Fig. 3 IR spectrum (neat) of compound 5 (upper) and compound 1
(lower).
at m/z = 2989 (M-) as well as strong signals at m/z = 1468 and
1404 corresponding to different fragment ions of compound 1,
generated by the strong laser required to ionize the compound.
◦
1
The H NMR spectrum of the compound in CDCl3 at 25 C
(ESI†) has several signals for the octadecyl chains at 4.02, 1.80,
2376 | Dalton Trans., 2012, 41, 2374–2381
This journal is
The Royal Society of Chemistry 2012
©

View Article Online
Scheme 2 Synthesis of precursors for dumbbell 2. Conditions: i) NaOAc, ethanol, reflux 12 h (purified yield: 48%). ii) Ethynyltrimethylsilane (1.15 eq.),
Pd(PPh3)2Cl2, CuI, diethylamine, DMF, MW heating 120 ◦C, 45 min. (purified yield: 69%). iii) K2CO3, CH2Cl2, ethanol, RT, 12 h (purified yield: 74%).
iv) a) Et2O, n-BuLi (2.5 M), TMEDA, 0 ◦C, stirring 12 h, tributyltin chloride, RT, 12 h. b) CH2Cl2, -79 ◦C, I2, stirring 12 h (purified yield: 70%). v)
Pd(PPh3)2Cl2, CuI, DIPA, 90 ◦C, 2 days (purified yield: 39%). vi) NaOH, ethanol, reflux, 5 h (yield: 55%). vii) C60 (3 eq.), 4-methoxybenzaldehyde (excess),
o-dichlorobenzene, reflux, 2 days (yield: 0%).
4.29 ppm. The latter eight multiplets correspond to the ferrocene
protons in the dumbbell which are all non-equivalent. The methine
protons situated at the stereogenic centers in dumbbell 1 appear as
a multiplet at 5.47 ppm. Also, there are two multiplets at 4.17 and
3.89 ppm originating from the eight methylene protons closest
to the oxygen atoms in the octadecyl chains. The other signals
from the octadecyl chains (140H) appear at 1.81, 1.75, 1.34 and
0.92 ppm.
Diffusion measurements carried out on the aliphatic signals
(by the LED-PGSE pulse sequence) showed a significantly slower
diffusion (9.9 ¥ 10-11 m2 s-1 in CDCl3) for 1 as compared to
fulleropyrrolidine 5 (1.2 ¥ 10-10 m2 s-1 in CDCl3) as expected based
on the larger size of the dumbbell.
The UV/Vis spectra for toluene solutions of 1, fulleropyrro-
lidine 5 and reference compound 6 are shown in Fig. 5. The
UV/Vis spectrum of the dumbbell is relatively similar to the
spectra of compound 5 and 6 with major absorption bands around
284 nm, ~312 nm and weak absorptions at 430 and around 700 nm
characteristic for fulleropyrrolidines.19 Some differences in the area
between 430–750 nm are however present that may indicate some
ground-state interactions between the fullerenes and the ferrocene
unit in compound 1.
The fluorescence spectra of 1, fulleropyrrolidine 5 and fulleropy-
rrolidine 6 (Fig. 6) show the weak fullerene fluorescence in the
region between 650–850 nm. Interestingly, the emission maxima
for the dumbbell situated at 726 nm is red shifted as compared to
This journal is
The Royal Society of Chemistry 2012
Dalton Trans., 2012, 41, 2374–2381 | 2377
©

View Article Online
Fig. 6 Fluorescence spectra of compound 1 (solid), compound 5 (dashed)
and compound 6 (dotted) in degassed toluene at 25 ◦C. The concentrations
of the analytes were 0.10 mM, and the excitation wavelength was 470 nm.
are the result of an intramolecular excimer, formed by the two
fullerenes in the dumbbell. However, based on the short life time of
the fullerene singlet excited state, self-quenching via the formation
of an intramolecular singlet excimer is not likely. Besides, no self-
quenching has been found in C60 solutions in o-dichlorobenzene
at nearly saturated concentrations, and is also absent in a reported
fullerene dimer in which the two covalently bonded fullerene cages
are very close to each other.20 Hence, the quenching of the fullerene
emission is due to intramolecular energy or electron transfer
between the fullerene and the ferrocene parts, which also has been
observed in several related C60-ferrocene systems.9
Fig. 4 Expansion of the 3.8–6.6 ppm region in the 1H NMR spectra of
compound 1. (Upper): In CDCl3, at 25 ◦C. (Lower): In C6D6, at 70 ◦C.
Cyclic voltammetry of the dumbbell 1 (Fig. 7 and Table 1)
shows one quasi-reversible oxidation wave at +0.32 V (vs. Fc/Fc+)
corresponding to the one electron oxidation of the central
ferrocene unit in the dumbbell. This oxidation peak is shifted
to higher potential as compared to the reference compound 1,1¢-
bis(pyrrolidinylcarbonyl)ferrocene 7 by +60 mV. On the reduction
side, four (quasi) reversible waves corresponding to the reduction
of the C60 part of the dumbbell can be observed. The intensity of
these reduction waves are 2.0 0.1 as large as the single oxidation
wave at +0.32 V and correspond to two electrons each. These
observations are consistent with the expected behaviour of the
dumbbell 1, and are further evidence for its structure. Moreover,
the four C60 reductions indicate a lack of electronic interactions
between the two fullerene cores in the ground state, which is
consistent with previous studies of other C60 dumbbells.4 The
somewhat irregular shape of the two first reduction waves for
dumbbell 1 is most likely due to some adsorption phenomenon
taking place on the working electrode as similar results were
obtained with fulleropyrrolidine 5.
Fig. 5 UV/Vis spectra of compound 1 (solid), compound 5 (dashed) and
compound 6 (dotted). Spectra are normalized and recorded in toluene.
both compound 5 and 6 which have emission maximas at 715 nm
and 712 nm respectively. Also, at the same concentrations one
could expect the fluorescence yield to be twice as high in compound
1 relative to the two reference compounds since it contains two
fulleropyrrolidines.
However, there is significant quenching of the fullerene fluores-
cence in the dumbbell relative to the reference compounds under
identical experimental conditions and sample concentrations.
Similar results were also obtained in 1,2-dichloroethane in which
the quenching of the fluorescence for dumbbell 1 was slightly more
pronounced.
A comparison of the reduction potentials of dumbbell 1 with
those of fulleropyrrolidine 5 shows an anodic shift of between
80–170 mV for the four reversible or quasi-reversible reductions
in dumbbell 1. Also, the difference in E0 increases by each
red
reduction step. For a better comparison, the voltammogram of
compound 6 which is a more appropriate reference compound
was also recorded (Fig. 7), and its electrochemical data can be
found in Table 1. For this compound, the reduction potentials are
anodically shifted by 70–160 mV as compared to compound 1.
Due to the high rotational flexibility around the central
ferrocene unit, compound 1 can adopt a conformation in which
the two fullerenes are in close proximity. It is therefore possible
that the red shifted emission maxima and lower fluorescence yield
2378 | Dalton Trans., 2012, 41, 2374–2381
This journal is
The Royal Society of Chemistry 2012
©

View Article Online
Table 1 Electrochemical data for compounds 1, 5, 6 and 7a
Compound
E0 [V]
DEp [mV]
E0
[V]
DEp [mV]
E0
[V]
DEp [mV]
E0
[V]
DEp [mV]
E0
[V]
red,4
DEp [mV]
ox
red,1
red,2
red,3
1
5
6
7
+0.32
114
-1.19
-1.27
-1.09
200
280
120
-1.60
-1.72
-1,53
170
290
110
-2.20
-2.34
-2.10
190
260
120
-2.70
-2.87
-2.54
210
320
180
+0.26
146
a Potentials are reported against the ferrocene/ferrocenium couple. Working electrode: Glassy carbon electrode (3 mm); counter electrode: Pt; quasi-
reference electrode: Pt. Toluene/acetonitrile (4 : 1) was used as solvent together with 0.1 M TBAPF6. Concentrations of the analytes were 1 mM. All
measurements were carried out at 25 ◦C under nitrogen atmosphere in degassed solvents. E0 = (Epc + Epa)/2, in which Epc and Epa correspond to cathodic
and anodic peak potentials respectively. DEp = (Epc - Epa). Scan rate: 25 mVs-1.
Fig. 7 (Left): Cyclic voltammograms of dumbbell 1 (upper) and fulleropyrrolidine 6 (lower). (Right): oxidation waves of ferrocene (dashed),
1,1-bis(pyrrolidinylcarbonyl)ferrocene 7 (dotted) and dumbbell 1 (solid).
Prato et al. have previously reported the electrochemical
characterization of some ferrocenyl fulleropyrrolidines by cyclic
voltammetry.21 In that study the authors found no apparent
ground-state electronic interation between the fullerene and the
ferrocene parts in a dyad in which the ferrocene was connected to
the fulleropyrrolidine nitrogen by an amide group (as it is also in
compound 1). The small differences in redox potentials observed
for compound 1 as compared to the two reference compounds 6
and 7 indicate however some ground-state electronic interaction
between the ferrocene unit and the fullerenes in the dumbbell.
Thus, a reasonable rationale for this is that the introduction of
a secondary fulleropyrrolidine to the ferrocene unit leads to more
pronounced interactions between the two electroactive moieties
which become measureable by cyclic voltammetry.
no evidence for interactions between the two individual fullerenes
in the dumbbell were observed, and the compound thus behaves
as two independent donor–acceptor systems.
Experimental
General
Thin-layer chromatography analyses were performed on pre-
coated Merck silica gel plates (60 F254) and visualized with UV
light. Melting points were recorded on a Stuart scientific SMP10
1
melting point apparatus and are uncorrected. H and 13C NMR
spectra were recorded on a Varian 500 MHz spectrometer and
chemical shifts are given in ppm (d) using the residual CHCl3 or
C6D6 peak as internal standard. Diffusion measurements were
made using the LED-PGSE pulse sequence. z-gradients were
employed and 16 scans were acquired. A relaxation delay of
0 s, 10 ms gradient pulse duration, 100 ms diffusion delay, 5 ms
storage delay were used and the gradient pulse strength was
arrayed between 0 and 20 gauss cm-1 (30 steps). The diffusion
coefficients were calculated based on the diffusion coefficient of
CHCl3 (5.2 ¥ 10-6 cm2 s-1). Solid state IR spectra were recorded
on a Perkin–Elmer Spectrum-100 FT-IR spectrometer with an
ATR accessory. The samples were analyzed by placing neat
samples directly on the ATR crystal. UV spectra were recorded
on a Varian Cary 3 Bio UV/Vis spectrometer. Fluorescence
spectra were recorded on a Varian Cary Eclipse fluorescence
spectrometer using a 10 mm quartz cell. The slits were 10 nm.
0.1 mM solutions of the analytes were made in degassed toluene
Conclusion
We have successfully synthesized a highly soluble organometallic
fullerene dumbbell in which two fulleropyrrolidines are connected
to a central ferrocene unit. The dumbbell has been fully charac-
terized by standard methods such as 1H NMR, mass spectroscopy
and elemental analysis.
Fluorescence spectroscopy showed a significant quenching
of the fullerene emission that was red-shifted as compared to
reference compounds 5 and 6, indicating that energy or electron
transfer occurs in the dumbbell in its excited state. Moreover,
UV/Vis and in particular cyclic voltammetry studies revealed
some degree of electronic communication between the central
ferrocene unit and the fullerenes in the ground state. In contrast,
This journal is
The Royal Society of Chemistry 2012
Dalton Trans., 2012, 41, 2374–2381 | 2379
©

View Article Online
and kept under nitrogen atmosphere during the measurements.
Mass spectrometry was carried out using a Voyager-DE PRO
Maldi-TOF spectrometer in negative ion-mode. Dithranol or 9-
nitroanthracene was used as a matrix. Matrix solutions were
prepared in toluene (dithranol) or acetone (9-nitroanthracene) by
dissolving 10 mg of the matrix in 1 mL of the solvent. The analytes
were then dissolved in toluene and mixed with a small amout of
the corresponding matrix solution and placed on the target plate.
Cyclic voltammetry was carried out using a Solartron Mobrey
1285 potentiostat. Sample solutions (5 mL) were prepared from
toluene/acetonitrile (4 : 1) containing 0.1 M tetrabutylammonium
hexafluorophosphate (TBAPF6) (Fluka, electrochemical grade) as
supporting electrolyte. The working electrode was a glassy carbon
disc (3 mm) and platinum electrodes served as quasi-reference and
counter electrodes. The ferrocene/ferrocenium couple was used as
an internal reference in all experiments. All cyclic voltammograms
shown were recorded at a scan rate of 25 mV s-1. Before the
measurements, oxygen was removed by bubbling nitrogen through
the stirred solutions and samples were kept under nitrogen
atmosphere during all measurements.
7.28 (1H, dd, J = 8.3, 2.1 Hz), 6.91 (1H, d, J = 8.3 Hz), 5.72 (1H, s),
5.08 (1H, d, J = 10.3 Hz), 4.86 (1H, d, J = 10.3 Hz), 4.01 (2H, t, J =
6.8 Hz), 3.98 (2H, t, J = 6.8 Hz), 3.26 (1H, br s, NH) 1.80 (2H, m),
1.74 (2H, m), 1.43 (4H, m), 1.34–1.19 (56H, m), 0.88 (6H, m) ppm.
dC(125 MHz, CDCl3, 25 ◦C): 156.4, 153.6, 149.5, 149.1, 147.4,
147.15, 147.35, 147.30, 147.1, 146.9, 146.7, 146.5, 146.40, 146.35,
146.30, 146.2, 146.1, 145.9, 145.63, 145.60, 145.55, 145.46, 145.44,
145.42, 145.40, 145.32, 144.8, 144.7, 144.50, 144.47, 144.3, 143.1,
142.86, 142.77, 142.72, 142.67, 142.46, 142.37, 142.31, 142.29,
142.19, 142.18, 142.07, 142.02, 141.9, 141.7, 140.32, 140.28, 140.0,
139.8, 136.4, 136.3, 136.1, 136.0, 130.1, 121.0, 114.4, 113.5, 69.7,
69.3, 32.1, 29.9, 29.6, 29.50, 29.46, 29.3, 26.2, 22.9 ppm. MS
(Maldi-TOF, 9-nitroanthracene, negative mode) m/z: 1376 (M-).
1,1¢-Ferrocenedicarboxylic acid12
Ferrocene (3.0 g, 16.1 mmol) was added to a flame-dried three-
necked flask equipped with a septum. The flask was evacuated and
flushed with nitrogen. Dry pentane (45 mL) and TMEDA (6 mL)
was added to the flask via syringe. The resulting mixture was
cooled to 0 ◦C and allowed to stir for 15 min. Butyl lithium (2.5 M
in hexanes, 17 mL) was slowly added to the cooled solution using
a syringe. The solution was then allowed to stir under nitrogen
atmosphere overnight after which a precipitate had formed in
the reaction flask. Dry THF (100 mL) was added to the flask,
resulting in a red solution. CO2 (g) was then bubbled though the
solution during a 1.5 h period resulting in the formation of a yellow
precipitate. Water (150 mL) was added to the reaction mixture and
the organic solvents were then evaporated. The aqueous phase was
washed with diethyl ether (3 ¥ 100 mL) and the aqueous phase
was then acidified using HCl (1 M) resulting in an orange/yellow
precipitate. The precipitate was filtered off using a glass sinter
funnel and the solid was then dried in vacuo (purified yield: 1.81 g,
41%). dH (500 MHz, DMSO-d6, 25 ◦C): 4.69 (4H, t, J = 1.8 Hz),
4.45 (4H, t, J = 1.8 Hz) ppm.
3,4-Bis(octadecyloxy)benzaldehyde
3,4-Dihydroxybenzaldehyde (1.0 g, 7.25 mmol), 1-octadecyl-
bromide (4.84 g, 14.5 mmol) and anhydrous potassium carbonate
(2.0 g, 14.5 mmol) was dispersed in acetone (60 mL) and the
mixture was refluxed under nitrogen atmosphere overnight. The
hot reaction mixture was filtrated and the solid material on the
filter paper was collected and subjected to column chromatography
(silica, dichloromethane). The first band was collected and dried in
vac◦uo (purified yield: 1.9 g, 41%). Mp 85 ◦C. dH (500 MHz, CDCl3,
25 C): 9.83 (1H, s), 7.41 (1H, dd, J = 8.1, 1.9 Hz), 7.39 (1H, d,
J = 1.9 Hz), 6.95 (1H, d, J = 8.1 Hz), 4.08 (2H, t, J = 6.6 Hz),
4.05 (2H, t, J = 6.6 Hz), 1.85 (4H, m), 1.48 (4H, m), 1.41–1.20
(56H, m), 0.88 (6H, m) ppm. dC(125 MHz, CDCl3, 25 ◦C): 191.2,
154.8, 149.6, 130.0, 126.8, 111.9, 111.1, 69.29, 69.27, 32.1, 29.9,
29.8, 29.5, 29.2, 29.1, 26.15, 26.10, 22.8, 14.3 ppm.
1,1¢-N,N¢-(2-(3,4-Bis(octadecyloxy)benzene)-3,4-
fulleropyrrolidinyl)ferrocenedicarboxamide (1)
2-(3,4-Bis(octadecyloxy)benzene)-3,4-fulleropyrrolidine (5)
1,1¢-Ferrocenedicarboxylic acid (10.0 mg, 0.0365 mmol) was
added to a flame-dried flask equipped with a septum. Dry
dichloromethane (1.5 mL), DMF (100 mL) and oxalyl chloride
(0.1 mL) was added via syringe and the mixture was allowed to stir
under nitrogen atmosphere for 1.5 h. The solvents were evaporated
and the residue was redissolved in dry dichloromethane (4 mL).
The resulting solution of 1,1¢-ferrocenedicarboxylic acid chloride
was added to a solution containing compound 1 (100 mg,
0.073 mmol) in dry dichloromethane (10 mL) and dry pyridine
(0.9 mL). The mixture was stirred under nitrogen atmosphere at
room temperature overnight. The solvent was evaporated and the
crude product was redissolved in toluene. Ethanol was then added
and the resulting precipitate was filtered off. The solid material
was washed with pentane and ethanol and was then subjected to
column chromatography (pentane:EtOAc 10 : 1–toluene/EtOAc
5 : 1) to afford compound 2 as a brown solid. (Purified yield:
75 mg, 69%) C220H168FeN2O6 (2989.23): calcd. C 88.33 H 5.66
C60 (50 mg, 0.07 mmol), glycine (10 mg, 0.13 mmol), 3,4-
bis(octadecyloxy)benzaldehyde (80 mg, 0.12 mmol) was dissolved
in o-dichlorobenzene (3 mL) in a 25 mL pear-shaped flask
equipped with a reflux condenser and the mixture was stirred
at 150 ◦C under nitrogen atmosphere overnight. Methanol (5 mL)
was added to precipitate the crude product which was filtered
off on a glass sinter funnel. The solid material was washed with
methanol and was then subjected to two subsequent column
chromatography purifications on silica. In the first column toluene
was used to remove unreacted C60, and the product was then eluted
together with unreacted 3,4-bis(octadecyloxy)benzaldehyde using
toluene:EtOAc (2 : 1). Final purification of the compound was
accomplished by a subsequent column using pentane:EtOAc 10 : 1
to remove non-C60 containing substances and then toluene: EtOAc
2 : 1 to elute compound 1 as a brown solid (purified yield: 37 mg,
39%). FT-IR (neat): nmax/cm-1 = 2918, 2850 (s, C–H stretch), 1604
(N–H bend), 1590, 1513, 1466, 1429, 1381, 1303, 1262, 1232, 1177,
1134, 1020, 942, 892, 842, 800, 764, 719, 693, 659, 623, 598, 574,
553, 525(s). dH(500 MHz, CDCl3, 25 ◦C): 7.36 (1H, d, J = 2.1 Hz),
N 0.94; found C 88.34 H 5.72 N 1.25. FT-IR (neat): nmax/cm-1
=
2920, 2850 (C–H), 1635, 1589, 1513, 1462, 1456, 1429, 1392, 1258,
1238, 1189, 1137, 1090, 1023, 854, 802, 770, 759, 743, 720, 699,
2380 | Dalton Trans., 2012, 41, 2374–2381
This journal is
The Royal Society of Chemistry 2012
©

View Article Online
576, 563, 552, 525. dH (500 MHz, C6D6, 70 ◦C): 7.80 (1H, d, J =
2.2 Hz), 7.78 (1H, d, J = 2.2 Hz), 7.68 (2H, m), 7.07 (1H, d, J =
8.2 Hz), 7.05 (1H, d, J = 8.2 Hz), 6.64 (1H, d, J = 12 Hz), 6.37 (1H,
d, J = 12 Hz), 5.80 (1H, d, J = 12 Hz), 5.78 (1H, d, J = 12 Hz),
5.47 (2H, m), 5.39 (1H, m), 5.31 (1H, m), 5.18 (1H, m), 5.12 (1H,
m), 4.41 (2H, m), 4.38 (1H, m), 4.29 (1H, m), 4.17 (4H, m), 3.89
(2H, t, J = 6.5 Hz), 3.87 (2H, t, J = 6.5 Hz), 1.81 (4H, m), 1.75
(4H, m), 1.34 (120H, m), 0.92 (12H, m) ppm. MS (MALDI-TOF,
dithranol, negative mode) m/z: 2989 (M-).
5 C. A. Martin, D. Ding, J. K. Sørensen, T. Bjørnholm, J. M. van
Ruitenbeek and H. S. J. van der Zant, J. Am. Chem. Soc., 2008, 130,
13198–13199.
6 Leary, M. Teresa Gonza´lez, C. van der Pol, M. R. Bryce, S. Filippone,
N. Mart´ın, G. Rubio-Bollinger and Nicola´s Agra¨ıt, Nano Lett., 2011,
11, 2236–2241.
7 C.-H Andersson, G. Berggren, S. Ott and H. Grennberg, Eur. JIC,
2011, 11, 1744–1749.
8 C. H. Andersson, PhD thesis, Uppsala Univerisity 2011.
9 D. M. Guldi, M. Maggini, G. Scorrano and M. Prato, J. Am. Chem.
Soc., 1997, 119, 974–980.
10 A. Bagno, S. Claeson, M. Maggini, M. L. Martini, M. Prato and G.
Scorrano, Chem.–Eur. J., 2002, 8, 1015–1023.
11 Z. Zhou, D. I. Schuster and S. R. Wilson, J. Org. Chem., 2006, 71,
1545–1551.
Acknowledgements
12 R. Laufer, U. Veith, N. J. Taylor and V. Snieckus, Can. J. Chem., 2006,
Financial support from the Swedish Research Council and the
Uppsala University priority program for molecular electronics is
acknowledged.
84, 356–369.
13 W. K. Anderson, D. Bhattacharjee and D. M. Houston, J. Med. Chem.,
1989, 32, 119–127.
14 M. Erde´lyi and A. Gogoll, J. Org. Chem., 2001, 66, 4165–4169.
15 M. E. Wright, Organometallics, 1990, 9, 853–856.
16 I. R. Butler, S. B. Wilkes, S. J. Mcdonald and L. J. Hobson, Polyhedron,
1993, 12, 129–131.
17 (a) P. A. Troshin, A. S. Peregudov, S. I. Troyanov and R. N.
Lyubovskaya, Russ. Chem. Bull., 2008, 57, 887–912; (b) M. Prato, M.
Maggini, C. Giacometti, G. Scorrano, G. Sandonh and G. Farnia,
Tetrahedron, 1996, 52, 5221–5234.
18 M. Maggini, A. Karlsson, L. Pasimeni, G. Scorrano, M. Prato and L.
VaIlid, Tetrahedron Lett., 1994, 35, 2985–2988.
19 M. Prato and M. Maggini, Acc. Chem. Res., 1998, 31, 519–
526.
Notes and references
1 (a) L. Echegoyen and L. E. Echegoyen, Acc. Chem. Res., 1998, 31,
593–601; (b) D. M. Guldi and M. Prato, Acc. Chem. Res., 2000, 33,
695–703.
2 For example: (a) C. J. Brabec, S. Gowrisanker, J. J. M. Halls, D. Laird,
S. Jia and S. P. Williams, Adv. Mater., 2010, 22, 3839–3856; (b) C.-F.
Lin, M. Zhang, S.-W Liu, T.-L. Chiu and J.-H. Lee, Int. J. Mol. Sci.,
2011, 12, 476–505; (c) H. Imahori and F. Fukuzumi, Adv. Funct. Mater.,
2004, 14, 525–536.
3 A. I. Lucas, N. Martin, L. Sa´nchez and C. Seoane, Tetrahedron Lett.,
1996, 37, 9391–9394.
20 B. Ma, J. E. Riggs and Y.-P. Sun, J. Phys. Chem. B, 1998, 102, 5999–
6009.
´
4 L. Sa´nchez, M. A. Herranz and N. Mart´ın, J. Mater. Chem., 2005, 15,
21 M. Maggini, A. Karlsson, G. Scorrano, G. Sandona, G. Farnia and M.
Prato, J. Chem. Soc., Chem. Commun., 1994, 589–590.
1409–1421.
This journal is
The Royal Society of Chemistry 2012
Dalton Trans., 2012, 41, 2374–2381 | 2381
©