A simple method for desymmetrizing 1,1′-ferrocenedicarboxaldehyde

Article abstract of DOI:10.1016/j.tetlet.2014.05.040

A simple chromatography-free method for desymmetrizing ferrocene is described starting from the readily available dialdehyde. Oxidation of 1,1′-ferrocenedicarboxaldehyde in a water/acetonitrile mixture with KMnO₄ produced 1′-formyl-ferrocenecarboxylic acid. The same reaction carried out in a water/acetone mixture produced 1′-[(E)-3-oxo- but-1-enyl]-ferrocenecarboxylic acid.

Full text of DOI:10.1016/j.tetlet.2014.05.040

Tetrahedron Letters 55 (2014) 3777–3780
Contents lists available at ScienceDirect
Tetrahedron Letters
journal homepage: www.elsevier.com/locate/tetlet
A simple method for desymmetrizing 1,
10-ferrocenedicarboxaldehyde
b
b
c
c
Andrew C. Benniston a, , Dumitru Sirbu , Constantin Turta , Michael R. Probert , William Clegg
⇑
a Molecular Photonics Laboratory, School of Chemistry, Newcastle University, Newcastle upon Tyne NE1 7RU, UK
b Institute of Chemistry, Academy of Sciences of Moldova, 3, Academiei str., Chisinau, Republic of Moldova
c Crystallography Laboratory, School of Chemistry, Newcastle University, Newcastle upon Tyne NE1 7RU, UK
a r t i c l e i n f o
a b s t r a c t
Article history:
A simple chromatography-free method for desymmetrizing ferrocene is described starting from the read-
ily available dialdehyde. Oxidation of 1,10-ferrocenedicarboxaldehyde in a water/acetonitrile mixture
with KMnO4 produced 10-formyl-ferrocenecarboxylic acid. The same reaction carried out in a water/ace-
tone mixture produced 10-[(E)-3-oxo-but-1-enyl]-ferrocenecarboxylic acid.
Received 13 March 2014
Revised 19 April 2014
Accepted 7 May 2014
Available online 21 May 2014
Ó 2014 Elsevier Ltd. All rights reserved.
Keywords:
Ferrocene
Oxidation
Desymmetrization
Carboxylic acid
Michael addition
Ferrocene is one of the most widely established compounds in
the chemical literature, and it is frequently used as a starting mate-
rial in both organometallic and organic chemistry transforma-
tions.1 This sandwich-type compound consists of an iron(II) atom
bonded in an g5 fashion to two cyclopentadienide rings (Cp).
Redox cycling between the Fe2+/Fe3+ states is both chemically
and electrochemically reversible.2 Certainly this final point is one
main reason why ferrocene has been so popular as a redox reporter
in supramolecular structures.3 The Cp rings are aromatic and hence
it is possible to perform typical electrophilic substitution reactions
on one or both of the rings.4 Selective reactions on a single Cp ring
to introduce two disparate functional groups (e.g., PPh2, dihydro-
oxazole) are also possible.5 For this latter case the ferrocene sub-
unit is used as a planar chiral ligand for asymmetric catalysis
applications.6 For each ring to incorporate a different functional
group can be complex, often requiring selective BuLi reactions at
low temperatures.7 A long reaction time and a convoluted purifica-
tion method is possible starting from 1,10-ferrocenedicarboxalde-
hyde.8 A very mild and fast method to achieve a similar outcome
would appear to be appealing, provided the two functional groups
are useful for further reactions.9 It turned out that, as part of our
research effort into producing ferrocene-based porphyrin and
borondipyrromethene (Bodipy) compounds, we discovered such a
simple procedure.
The oxidation of ferrocenecarboxaldehyde using aqueous
KMnO4 in acetone is well documented to produce the correspond-
ing carboxylic acid.10 Repeating the literature method afforded fer-
rocenecarboxylic acid in reasonable yield. However, we found that
the same procedure applied to 1,10-ferrocenedicarboxaldehyde in
acetone or acetonitrile did not produce the corresponding dicar-
boxylic acid (Scheme 1). The 1H NMR spectrum for the major prod-
uct isolated from the acetonitrile reaction showed different
resonances for the two Cp rings, confirming the product to be
unsymmetrical. Furthermore, two downfield resonances at
d
193.21 and 170.77 for the 13C{1H} NMR spectrum were consistent
with the presence of aldehyde and carboxylic acid groups. The
mass spectrum showed a peak at m/z 259, and a broad peak at
ca. 2939 cmꢀ1 in the FT-IR spectrum confirmed production of a car-
boxylic acid group. The combined evidence verified the product to
be compound 1, in which one aldehyde was oxidized but the other
remained untouched. Hence, under these conditions the ferrocene
is desymmetrized since it contains two different functional groups.
The overall yield for 1 is 49% after a fast and chromatography-free
work up, providing material pure enough for further synthetic
reactions. A similar reaction performed using an acetone/H2O mix-
ture (ca. 1:1) did not afford a product with a 1H NMR spectrum
consistent with compound 1. Instead, additional resonances were
observed in the olefinic region (d = 7.35 and 6.30, J = 16.2 Hz), a
⇑
Corresponding author. Tel.: +44 191 222 5706; fax: +44 191 222 6929.
E-mail address: a.c.benniston@ncl.ac.uk (A.C. Benniston).
methyl resonance (d = 2.25) and
a
downfield resonance
http://dx.doi.org/10.1016/j.tetlet.2014.05.040
0040-4039/Ó 2014 Elsevier Ltd. All rights reserved.

3778
A. C. Benniston et al. / Tetrahedron Letters 55 (2014) 3777–3780
Scheme 1. Reagents and conditions: (i) KMnO4, CH3CN/H2O, 0 °C; (ii) KMnO4,
acetone/H2O, 0 °C.
(d = 197.51) in the 13C NMR spectrum, which is consistent with the
existence of a CH3CO group (see Supporting information).11 The
mass spectrum showed a peak at m/z 299, and once again the
FT-IR spectrum corroborated the presence of a carboxylic acid
(2934 cmꢀ1).12 The spectroscopic data are fully consistent with
compound 2 with a trans double bond. We speculate that oxidation
of one aldehyde to the carboxylic acid activates the other carbonyl
in 1 to a condensation reaction with the acetone present in solu-
tion. Presumably for ferrocenecarboxaldehyde the reaction of the
aldehyde with acetone is too slow to compete with oxidation.
Rather surprisingly, compounds 1 and 2 could not be found in
the literature using popular search tools (i.e., SciFinder, Reaxys).
The closest analogue to 2 is the structural isomer (C15H14FeO3) in
which one ring has a formyl and the other (E)-b-(methoxycar-
Figure 1. Selected partial 400 MHz 1H NMR spectra for 3 (black) and light-driven
degradation products (red). Assignments of proton resonances for the terpy ligand
in 3 are shown.
bonyl)ethenyl subunit.8 This material was produced by
a
reaction could also be followed by UV–vis spectroscopy in MeOH,
and as the reaction proceeded, a new band at kABS = 580 nm
appeared (see Supporting information). The new absorption is
remarkably similar to the metal-to-ligand charge transfer band
controlled Wittig reaction with 1,10-ferrocenedicarboxaldehyde,
yielding a mixture of compounds which had to be separated by
chromatography. Again for compound 1 the structural analogue
(C12H10FeO3) has both the formyl and carboxylic acid substituted
on the same Cp ring.13
observed for [Fe(terpy)2]2+ 16
.
In addition, the mass spectrum of 3
in methanol left in the light showed peaks at m/z 977 and 489 that
correspond to the ions [Fe(3)2-H]+ and [Fe(3)2]2+
Having established the simple procedure for desymmetrizing
the ferrocene core, we were interested to see if useful chemical
transformations could be carried out on compounds 1 and 2. Not-
ing that the carboxylic acid in 1 may act as an anchoring group to a
semiconductor (e.g., TiO2, NiO), the construction of a ligand from
the aldehyde was attempted (Scheme 2). In particular, the estab-
lished Kröhnke method for synthesizing 2,20:60,200-terpyridine (ter-
py) is well known.14 Reaction of 1 with 2-acetylpyridine in
aqueous NaOH/NH4OH, in a two-step one-pot reaction, gave after
an easy column-free purification the terpy-type ligand 3 in 62%
yield. The material is stable in the solid state, but when in a solu-
tion of MeOH, DMSO or water, air-equilibrated or N2-purged, the
compound 3 is extremely light-sensitive. Simply leaving a solution
exposed to ambient light for several minutes produced a dark pur-
ple coloration. When kept in the dark no alteration in color was
observed. Interestingly, there is a report that the analogue of 3
without the carboxylic acid very slowly changes its color under
light illumination.15 No further details were given, but it was
assumed that oxidation of the ferrocenyl group gave the blue color.
Here we report a more thorough study into the instability of the
Fc-tpy group in solution. 1H NMR spectra for 3 as the light-driven
reaction proceeded are shown in Figure 1. Resonances for the start-
ing material decreased and new peaks appeared with a pattern
very similar to authentic [Fe(Fc-terpy)2]2+ (Fig. 1).15 The conversion
yield, estimated from ferrocene integral values, is ca. 62%. The
.
Hence, the source of the iron(II) must be ferrocene for the light-
activated reaction to produce [Fe(3)2]2+. The reaction time is very
dependent on the concentration of the starting material and is con-
sistent with a bimolecular mechanism. It is speculated that a terpy
ligand from one molecule of ground-state 3 bites onto the ferro-
cene iron(II) center of another ‘activated’ molecule, leading to deg-
radation of the compound. The identity of the activated ferrocene
is not clear. However, the triplet quantum yield for ferrocene is
very reasonable (/T = 0.66), although the lifetime (albeit in DMSO
solution) is only 0.6 ns. The triplet energy is around 40 kcal molꢀ1
(14,000 cmꢀ1) and the structure for the triplet state is known to be
distorted.17 The reaction must proceed via the triplet state and the
presence of the carboxylic acid appears to promote compound deg-
radation. This idea was tested by reacting under the same condi-
tions
a
1:1 mixture of 1,10-ferrocenedicarboxylic acid and
terpyridine in MeOH. When exposed to ambient light the solution
became dark purple, with a very distinct absorption pattern sug-
gesting a similar degradation of the ferrocene unit and formation
of the [Fe(terpy)2]2+ complex. Non-substituted ferrocene was
indefinitely stable in methanol solution in the presence of
terpyridine.
Performing all steps, starting from 1,10-ferrocenedicarboxalde-
hyde, under ambient light illumination gave after purification by
column chromatography (silica gel, MeOH) a deep purple product
in 11% yield. Suitable crystals for X-ray crystallographic analysis
of the iron complex were obtained, and the obtained structure con-
firms its identity, although it is of relatively low precision, the mar-
ginal-quality crystals giving only weak diffraction. The structure of
the cation is illustrated in Figure 2, highlighting the six-coordinate
iron(II) center and the two Fe-terpy ligands of compound 3. The
Fe1AN1 [1.876(4) Å] and Fe1AN4 [1.875(4) Å] bond lengths are
typically shorter than the remaining FeAN bond lengths [Fe1AN2
1.974(4), Fe1AN3 1.968(4), Fe1AN5 1.972(4), Fe1AN6 1.971(4)
Å]. As observed in previous [Fe(terpy)2]2+ complexes18 the bite
angle of the ligand is not perfectly trans, as reflected in the
N2AFe1AN3 and N5AFe1AN6 bond angles of 161.66(17)° and
162.33(17)°, respectively. A point to note is that each ferrocene
Me
CHO
N
N
O
(i)
N
Fe
Fe
Fe
HO2C
HO2C
N
HO2C
1
3
O
Me
O
Fe
(ii)
HO2C
Me
2
4
Scheme 2. Reagents and conditions: (i) 2-acetylpyridine, aq NaOH/NH4OH, dark;
(ii) (a) diethyl malonate, NaOEt, EtOH; (b) H2O, H+.

A. C. Benniston et al. / Tetrahedron Letters 55 (2014) 3777–3780
3779
CH2(COOEt)2
NaOEt, Ethanol
Fc
Fc
COOEt
O
O
CO2Et
Fc
O
COOEt
Michael Addition
EtO2C
NaCH(COOEt)2
Aldol
OH
H
CO2Et
CO2Et
O
HO
O
OEt
O
CO2Et
Fc
Fc
COOEt
Fc
COOEt
OEt
CO2Et
OEt
CO2Et
COOEt
O
Dieckmann
O
Intramolecular
deprotonation
1. HO-
2. H+
Hydrolysis
H2O
CO2H
Fc
Fc
CO2H
Fc
Figure 2. X-ray determined structure of the dication [Fe(3)2]2+
solvent molecules, and hydrogen atoms are omitted for clarity.
.
Counter-ions,
O
H
O
CO2H
O
O
O
CO2H
Decarboxylation
Scheme 3. Proposed mechanism to explain the formation of compound 4.21
Fc = ferrocenecarboxylic acid.
any suggestion that the starting material was incorrect. A tentative
mechanism is suggested from the initially produced Michael addi-
tion product, followed by aldol addition of the second diethyl mal-
onate molecule and cyclohexanone ring closure through
a
Dieckmann condensation reaction (Scheme 3).20 Hydrolysis of
the esters and acidification is followed by dehydration of the aldol
intermediate and a series of decarboxylations to end up eventually
with the identified product 4.
In conclusion, we have demonstrated via simple oxidation that
1,10-ferrocenedicarboxaldehyde can be converted into desymmet-
rized compounds. There is clear scope to expand on the chemistry
of the two derivatives, especially via the terpy ligand 3 with its
anchoring carboxylic acid site. We expect to produce photoactive
systems for semiconductor attachment and dye-sensitized solar
cell applications from these products.
Figure 3. X-ray determined molecular structure for racemic compound 4 showing
here the S form.
Acknowledgments
We thank the Engineering and Physical Sciences (EPSRC)-spon-
sored mass spectrometry service at Swansea for collection of mass
spectra, the FP7-PEOPLE-2009-IRSES-246902 grant for the
exchange of researchers, EPSRC for a crystallography equipment
grant, and Diamond Light Source for access to synchrotron
facilities.
group is in its carboxylic acid form and so hexafluorophosphates
(not shown) are present as counter-ions.
The double bond activated by the ketone group in the conden-
sation product 2 is ideally suited for a Michael addition reaction.
Thus, reaction of 2 with diethylmalonate in ethanol, catalyzed by
sodium ethoxide, gave after quenching with water, acidification,
and purification by column chromatography (silica gel, CH2Cl2/
MeOH 9:1) a yellow product in 35% yield. The 1H , 13C and DEPT-
135° NMR spectra showed resonances typical for a cyclohexenone
group (see Supporting information).19a–c The 2-D 1H–1H COSY-45°
and heteronuclear (1H/13C HMQC and HMBC) correlation spectra,
are consistent with racemic 10-(5-methylcyclohex-4-en-3-one)fer-
rocenecarboxylic acid (4). The mass spectrum showed peaks at m/z
321, 339, and 677 that match for the [MꢀOH]+, [M+H]+, and
[2M+H]+ ions. After several attempts, crystals of the compound
suitable for synchrotron X-ray analysis were obtained. The molec-
ular structure is shown in Figure 3, and confirms unequivocally the
identity of the product. The C15AC16 bond length of 1.354(3) Å is
consistent with the expected double bond, and other CAC bond
lengths in the cyclohexenone ring are typical for CAC single bonds.
C12 is a chiral center and both the R and S enantiomers of 4 are
present in the crystal. A centrosymmetric dimer is generated in
the crystal structure, as a result of typical hydrogen bonding at
the carboxylic acid site (see Supporting information).
Supplementary data
Supplementary data (details of the synthesis of the compounds
and their NMR spectra; absorption spectrum for the light-driven
reaction; X-ray crystallographic data. Full crystallographic data
for the structures in this paper have been deposited with the Cam-
bridge Crystallographic Data Centre as supplementary publication
nos. CCDC 991073 and 991074. Copies of the data can be obtained,
free of charge, on application to CCDC, 12 Union Road, Cambridge
CB2 1EZ, UK. Fax: +44 (0)1223 336033 or e-mail: deposit@ccdc.ca-
m.ac.uk) associated with this article can be found, in the online
version, at http://dx.doi.org/10.1016/j.tetlet.2014.05.040.
References and notes
1. (a) Atkinson, R. C. J.; Gibson, V. C.; Long, N. J. Chem. Soc. Rev. 2004, 33, 313–328;
(b) Sutcliffe, O. B.; Bryce, M. R. Tetrahedron: Asymmetry 2003, 14, 2297–2325;
(c) Debroy, P.; Roy, S. Coord. Chem. Rev. 2007, 251, 203–221.
2. (a) Daeneke, T.; Kwon, T.-H.; Holmes, A.; Duffy, N.; Bach, U.; Spiccia, L. Nat.
Chem. 2011, 3, 211–215; (b) Nemykin, V.; Rohde, G.; Barrett, C.; Hadt, R.; Sabin,
J.; Reina, G.; Galloni, P.; Floris, B. Inorg. Chem. 2010, 49, 7497–7509.
3. Bayly, S. R.; Beer, P. D. Struct. Bond. 2008, 129, 45–94.
The intriguing conundrum is the mechanism by which com-
pound 2 is converted into 4, since there is no immediately obvious
route. The purity of the diethyl malonate was checked to eliminate

3780
A. C. Benniston et al. / Tetrahedron Letters 55 (2014) 3777–3780
4. Buchanan, J. G.; Sable, H. Z. In Selective Organic Transformations; Thyagarajan, B.
S., Ed.; Wiley-Interscience: New York, 1972; Vol. 2, pp 1–95.
5. Dong, T.-Y.; Lai, L.-L. J. Organomet. Chem. 1996, 509, 131–134; Butler, I. R.;
Cullen, W. R. Can. J. Chem. 1983, 61, 147–153.
6. Dai, L.-X.; Tu, T.; You, S.-L.; Deng, W.-P.; Hou, X.-L. Acc. Chem. Res. 2003, 36,
659–667.
7. Zhang, W.; Yoneda, Y.; Kida, T.; Nakatsuji, Y.; Yohji, N.; Ikeda, I. Tetrahedron:
Asymmetry 1998, 9, 3371–3380.
8. Argyropoulos, N.; Coutouli-Argyropoulos, E. J. Organomet. Chem. 2002, 654,
117–122.
13. Gyömöre, Á.; Csámpai, A. J. Organomet. Chem. 2011, 696, 1626–1631.
14. Krohnke, F. Synthesis 1976, 1–24.
15. Constable, E. C.; Edwards, A. J.; Martínez-Máñez, R.; Raithby, P. R.; Cargill
Thompson, A. M. W. J. Chem. Soc., Dalton Trans. 1994, 645–650.
16. Braterman, P.; Song, J.-I.; Peacock, R. Inorg. Chem. 1992, 31, 555–559.
17. Maciejewski, A.; Jaworska-Augustyniak, A.; Szeluga, Z.; Wojtczak, J.; Karolczak,
J. Chem. Phys. Lett. 1988, 153, 227–232.
18. Constable, E. C.; Housecroft, C. E.; Neuburger, M.; Phillips, D.; Raithby, P. R.;
Schofield, E.; Sparr, E.; Tocher, D. A.; Zhender, M.; Zimmermann, Y. J. Chem. Soc.,
Dalton Trans. 2000, 2219–2228.
9. In the preparation of neutral receptor systems based on ferrocene the selective
hydrolysis of dimethyl 1,10-ferrocenedicarboxylate was shown to produce the
monoester, see Li, C.; Medina, J. C.; Maguire, G. E. M.; Abel, E.; Atwood, J. L.;
Gokel, G. W. J. Am. Chem. Soc. 1997, 119, 1609–1618.
19. (a) Venkatapuram, P.; Boggu, J. M. R.; Akula, B.; Katta, V. R.; Dandu, B. R.
Molecules 2000, 5, 1281–1286; (b) Hajime, N.; Koichi, T.; Ikua, U. Chem. Pharm.
Bull. 1992, 40, 858–863; (c) Shujian, Q.; Kehui, S.; Hongyin, G.; Qiancai, L.;
Hong, W. Molecules 2007, 12, 988–996.
10. Batsanov, A. S.; Hérault, D.; Howard, J. A. K.; Patrick, L. G. F.; Probert, M. R.;
Whiting, A. Organometallics 2007, 26, 2414–2419.
20. (a) Yamane, H.; Sugawara, J.; Suzuki, Y.; Shimamura, E.; Takahash, N. Agric. Biol.
Chem. 1980, 44, 2857–2864; (b) Xiang, J.-N.; Nambi, P.; Ohlstein, E. H.; Elliott, J.
D. Bioorg. Med. Chem. 1998, 6, 695–700.
ˇ
11. Toma, S.; Sališová, M.; Solcániová, E. J. Organomet. Chem. 1975, 90, 335–341.
12. (a) Kono, H.; Shiga, M.; Motoyama, I.; Hata, K. Bull. Chem. Soc. Jpn. 1979, 43,
21. The alternative ring closure product is from a lactonization of the ketodiester
intermediate. There is no obvious mechanism to convert this derivative into
the final product.
ˇ
1435–1439; (b) Solcániová, E.; Toma, S.; Fiedlerová, A. Org. Magn. Reson. 1980,
14, 181–185.