Ferrocene-based nanoelectronics: Regioselective syntheses and electrochemical characterization of α-monothiol and α,ω- dithiol, phenylethynyl-conjugated, 2,5-diethynylpyridyl- And pyridinium-linked diferrocene frameworks having an end-to-end distance of ～4 nm

Article abstract of DOI:10.1021/om700807x

Regiospecific synthetic methods have been developed for the assembly of unsymmetric conjugated molecular frameworks containing 2,5-diethynylpyridyl- and 2,5-diethynylpyridinium-linked diferrocene structures and possessing either mono- or dithioacetate end-groups that are suitable for chemisorption onto Au(111) substrates after conversion to the corresponding thiol derivatives. Electronic spectra and solution electrochemistry of these and model compounds establish the electron-withdrawing character of a 2,5-dimethoxyphenylethynyl substituent on ferrocene that serves to shift the Fe(II)/Fe(III) redox couple to higher potentials. Further, while the unsymmetric nature of the 2,5-diethynylpyridyl bridge in 3 does not differentially perturb the redox couples of the two ferrocenes (ΔE_1/2 < 10 mV), upon methylation, the corresponding pyridinium moiety of 4 now produces a large separation in the two redox potentials (ΔE_1/2 = 190 mV). For the two regioisomeric monothioacetate compounds bearing a terminal 2,5-diethynylpyridyl-linked diferrocene unit, 5 and 6 (and their respective pyridinium counterparts, 7 and 8), redox potentials of the two ferrocenes are found to be either widely separated or similar in value depending upon the added influence of the 2,5-dimethoxyphenylethynyl group (e.g., ΔE _1/2= 310 mV in 7 vs ～50 mV in 8).

Full text of DOI:10.1021/om700807x

Organometallics 2008, 27, 927–937
927
Ferrocene-Based Nanoelectronics: Regioselective Syntheses and
Electrochemical Characterization of r-Monothiol and r,ω-Dithiol,
Phenylethynyl-Conjugated, 2,5-Diethynylpyridyl- and
Pyridinium-Linked Diferrocene Frameworks Having an End-to-End
Distance of ∼4 nm
Chaiwat Engtrakul and Lawrence R. Sita*
Department of Chemistry and Biochemistry, UniVersity of Maryland, College Park, Maryland 20742
ReceiVed August 13, 2007
Regiospecific synthetic methods have been developed for the assembly of unsymmetric conjugated
molecular frameworks containing 2,5-diethynylpyridyl- and 2,5-diethynylpyridinium-linked diferrocene
structures and possessing either mono- or dithioacetate end-groups that are suitable for chemisorption
onto Au(111) substrates after conversion to the corresponding thiol derivatives. Electronic spectra and
solution electrochemistry of these and model compounds establish the electron-withdrawing character of
a 2,5-dimethoxyphenylethynyl substituent on ferrocene that serves to shift the Fe(II)/Fe(III) redox couple
to higher potentials. Further, while the unsymmetric nature of the 2,5-diethynylpyridyl bridge in 3 does
not differentially perturb the redox couples of the two ferrocenes (∆E1/2 < 10 mV), upon methylation,
the corresponding pyridinium moiety of 4 now produces a large separation in the two redox potentials
(∆E1/2 ) 190 mV). For the two regioisomeric monothioacetate compounds bearing a terminal
2,5-diethynylpyridyl-linked diferrocene unit, 5 and 6 (and their respective pyridinium counterparts, 7
and 8), redox potentials of the two ferrocenes are found to be either widely separated or similar in value
depending upon the added influence of the 2,5-dimethoxyphenylethynyl group (e.g., ∆E1/2 ) 310 mV in
7 vs ∼50 mV in 8).
quantitative agreement between theory and experiment for
the efficiency of electron transport in all-organic saturated
or conjugated frameworks chemisorbed to gold electrodes
through Au-S interactions, such as those containing n-alkyl,
poly(phenyl), and poly(phenylethynyl) oligomers of general
structure Au/-S-(CH2)n-S-/Au, Au/-S-(C6H4)n-S-/Au,
and Au/-S-(C6H4-Ct C)n-C6H4S-/Au, respectively, is
still off by orders of magnitude.4,5 More specifically,
experimentally measured current (I)-voltage (V) curves for
all-organic LML heterojunctions uniformly exhibit a noncon-
ducting region centered at zero bias that extends for several
hundred mV and a first conductance maximum occurring at a
relatively large bias of 1.5 V with a peak conductance in the
nS range, which is about 500 times smaller than calculated.
Several possible reasons for these discrepancies include remain-
Introduction
Experimental and theoretical interest in the conduction physics
of metal lead/molecule/metal lead (LML) heterojunctions
remains unabated, and for good reason.1 After electron transfer
theory, which describes the coupling of electronic and nuclear
motions as the basis for electronic transfer rates, a thorough
understanding of the dependence of transport through LML
heterojunctions as a function of molecular structure is the next
logical step by which to further build the foundations of modern
chemistry, as conductance directly measures the connection of
wave functions between elements of a molecular framework
and the response of these connections to an applied electron
flux.2 Further, just as the successful experimental validation of
electron transfer theory found a wide range of applications, a
unified molecular conductance theory will be critical to a large
number of emerging, and as-of-yet-unforeseen, technologies,
such as “molecular electronics”, in which single molecules or
molecular assemblies are envisioned as serving as components
within nanoscale electronic devices.1–3
(4) (a) Collier, C. P.; Wong, E. W.; Belohradsky, M.; Raymo, F. M.;
Stoddart, J. F.; Kuekes, P. J.; Williams, R. S.; Heath, J. R. Science 1999,
285, 391–394. (b) Cui, X. D.; Primak, A.; Zarate, X.; Tomfohr, J.; Sankey,
O. F.; Moore, A. L.; Moore, T. A.; Gust, D.; Harris, G.; Lindsay, S. M.
Science 2001, 294, 571–574. (c) Holmlin, R. E.; Haag, R.; Chabinyc, M. L.;
Ismagilov, R. F.; Cohen, A. E.; Terfort, A.; Rampi, M. A.; Whitesides,
G. M. J. Am. Chem. Soc. 2001, 123, 5075–5085. (d) Salomon, A.; Cahen,
D.; Lindsay, S.; Tomfohr, J.; Engelkes, V. B.; Frisbie, C. D. AdV. Mater.
2003, 15, 1881–1890. (e) Xu, B.; Tao, N. J. Science 2003, 301, 1221–
1223. (f) Kushmerick, J. G.; Blum, A. S.; Long, D. P. Anal. Chim. Acta
2006, 568, 20–27. (g) Venkataraman, L.; Klare, J. E.; Tam, I. W.; Nuckolls,
C.; Hybertsen, M. S.; Steigerwald, M. L. Nano Lett. 2006, 6, 458–462.
(5) (a) Seminario, J. M.; De La Cruz, C. E.; Derosa, P. A. J. Am. Chem.
Soc. 2001, 123, 5616–5617. (b) Ke, S. H.; Baranger, H. U.; Yang, W. J. Am.
Chem. Soc. 2004, 126, 15897–15904. (c) Xue, Y.; Ratner, M. A. Phys.
ReV. B: Condens. Mat. Mater. Phys 2004, 70, 205416/1–205416/17. (d)
Reimers, J. R.; Solomon, G. C.; Gagliardi, A.; Bilic, A.; Hush, N. S.;
Frauenheim, T.; Di Carlo, A.; Pecchia, A. J. Phys. Chem. A 2007, 111,
5692–5702.
Unfortunately, after a decade of intense worldwide interest,
investment, and investigation, both the development of a
unified molecular conductance theory and the promise of
molecular electronics have not yet materialized. In this regard,
while significant experimental and theoretical advances have
been made in the study of LML heterojunctions, the
* Corresponding author. E-mail: lsita@umd.edu.
(1) Heath, J. R.; Ratner, M. A. Phys. Today 2003, 56, 43–49.
(2) Nitzan, A.; Ratner, M. A. Science 2003, 300, 1384–1389.
(3) (a) Reed, M. A. Nat. Mater. 2004, 3, 286–287. (b) Wassel, R. A.;
Gorman, C. B. Angew. Chem., Int. Ed. 2004, 43, 5120–5123.
10.1021/om700807x CCC: $40.75
2008 American Chemical Society
Publication on Web 02/08/2008

928 Organometallics, Vol. 27, No. 5, 2008
Engtrakul and Sita
Scheme 1
ing deficiencies and uncertainties with either the experimental
designs employed to date, the nature of the molecular systems
that have so far been chosen for study, or the present state-of-
the-art of computational methods that have been developed to
predict, in quantitative fashion, the conduction characteristics
of such LML heterojunctions. Indeed, given the general
fundamental electronic structure of organic frameworks, and
even fully (highly) conjugated ones, all of these LML systems
can be expected to possess electronic states for the organic
bridge that are well above and below the effective Fermi level
of the two metal leads, thereby, at a minimum, limiting
transmission efficiencies and the ability to influence the energy
levels of these electronic states relative to the effective Fermi
level by application of a bias to a third gate electrode in more
advanced LML configurations.
Given the possible intrinsic limitations of LML heterojunc-
tions that are based on all-organic frameworks, we began an
experimental program several years ago to explore the design,
construction, and conduction physics of LML heterojunctions
that incorporate transition metal-containing molecular frame-
works due to the simple a priori expectation that they should
possess low-lying metal-centered electronic states that can be
more easily brought into resonance with the effective Fermi
level of the leads.6–8 In addition to potentially providing for
higher transmissions at lower bias, intrinsically small energy
differences between these electronic states, in conjunction with
the spin properties of the transition metal ions, might also give
rise to unique and signature spin crossover behavior that could
then be probed and verified by bringing in additional parameters
into play (e.g., variable-temperature and magnetic field
studies).8b,8b,9 Finally, resonant conduction through low-lying
metal-centered states might well occur closer to equilibrium in
these metal-containing LML heterojunctions, and thus, they may
be more amenable to evaluation by present state-of-the-art
computational methods.10 In this respect, in 2001, we had
introduced the design shown in Scheme 1 for a ferrocene-based
molecular diode in which rectification of electron transport
through a molecular framework might be mediated by an applied
gate voltage.6 Our initial communication also provided experi-
mental support for the concept of utilizing the unsymmetric 2,5-
diethynylpyridyl bridge between two ferrocene moieties for this
purpose through the synthesis and electrochemical characteriza-
tion of model complexes for States 1 and 2 in solution. Left
unaddressed by these promising results, however, were two
unknowns that could potentially prove to be significant obstacles
to the eventual realization of Scheme 1. First is the synthetic
challenge presented in making the required unsymmetric difer-
rocene frameworks that are long enough (e.g., ∼3–4 nm) to
reliably span the distance between two closely fabricated metal
leads. Second, with this required molecular length, it is not
immediately obvious that the diferrocene structures of Scheme
1 are capable of supporting efficient electron transport either
via an “electron-hopping” mechanism involving distinct and
spatially defined metal-centered molecular states or via a
“through-bond” mechanism involving electron tunneling by way
(6) Engtrakul, C.; Sita, L. R. Nano Lett. 2001, 1, 541–549.
(7) Getty, S. A.; Engtrakul, C.; Wang, L.; Liu, R.; Ke, S.-H.; Baranger,
H. U.; Yang, W.; Fuhrer, M. S; Sita, L. R. Phys. ReV. B 2005, 71, 241401/
1–241401/4.
(8) For other experimental investigations of LML heterojunctions
incorporating inorganic- or organometallic-based molecular frameworks,
see: (a) Park, J.; Pasupathy, A. N.; Goldsmith, J. I.; Chang, C.; Yaish, Y.;
Petta, J. R.; Rinkoski, M.; Sethna, J. P.; Abruna, H. D.; McEuen, P. L.;
Ralph, D. C. Nature 2002, 417, 722–725. (b) Liang, W.; Shores, M. P.;
Bockrath, M.; Long, J. R.; Park, H. Nature 2002, 417, 725–729. (c) Kim,
B.; Beebe, J. M.; Olivier, C.; Rigaut, S.; Touchard, D.; Kushmerick, J. G.;
Zhu, X. Y.; Frisbie, C. D. J. Phys. Chem. C 2007, 111, 7521–7526.
(9) Liu, R.; Ke, S.-H.; Baranger, H. U.; Yang, W. Nano Lett. 2005, 5,
1959–1962.
(10) Liu, R.; Ke, S.-H.; Baranger, H. U.; Yang, W. J. Am. Chem. Soc.
2006, 128, 6274–6275. Liu, R.; Ke, S.-H.; Yang, W.; Baranger, H. U.
J. Chem. Phys. 2006, 124, 024718/1–024718/5.

Ferrocene-Based Nanoelectronics
Organometallics, Vol. 27, No. 5, 2008 929
Chart 1
Chart 2
Scheme 2
Scheme 3a
i
a
(a) (i) HCt CSi( Pr)3, Pd(PPh3)2Cl2, Cu(OAc)2 · H2O, diisopropyl-
amine (DIPA), 70° C, 24 h. (ii) TBAF (1.1 equiv), CH2Cl2, rt, 15
min.
of a single electronic state that is delocalized over the entire
length of the system. Fortunately, in addressing the latter of
these concerns first, we recently reported single-molecule
measurements that show that conductance through LML het-
erojunctions derived from chemisorbtion of the monoferrocene
dithiol 1 across the nanogap of electromigrated Au leads can
exceed 70% of the theoretical limit at a bias of less than 100
mV.7
Importantly, while the overall bias-dependent conductance
of LML heterojunctions derived from 1 is well explained by
transport through a single molecular state that is strongly coupled
to the macroscopic metal (Au) electrodes,11 conductance through
the corresponding ferrocene-absent LML heterojunction that is
based on the all-organic phenylethynyl dithiol molecular
framework of chemisorbed 2 shown in Chart 1 is at least 2
orders of magnitude lower and Coulomb-blockaded over several
hundred mV.
bring the design shown in Scheme 1 a step further to final
evaluation through the development of regiospecific synthetic
routes to the unsymmetric thioacetate (thiol)-terminated difer-
rocene compounds 3-8 shown in Chart 2 that are suitable for
chemisorption onto metal (Au) supports, which in the case of
self-assembled monolayer (SAM) test structures based on the
monothiols 6-8, can be incorporated in regiospecific fashion
with the nitrogen directed either toward or away from the Au
surface with known certainty. Electronic spectra and solution
electrochemistry of 3–8 and related model compounds further
serve to establish the electron-withdrawing character of a 2,5-
dimethoxyphenylethynyl substituent on ferrocene that manifests
as a shift of the Fe(II)/Fe(III) redox couple to higher potentials.12
Finally, for the sake of completeness and future reference, a
detailed account of the synthesis of 1 is reported here for the
first time as well.
Encouraged by the high transmission efficiency observed for
LML heterojunctions based on 1, the present report serves to
(11) (a) Landauer, R. IBM J. Res. DeV. 1957, 1, 233. (b) Datta, S.
Electronic Transport in Mesoscopic Systems; Cambridge University Press:
Cambridge, 1995.
(12) Results obtained regarding the fabrication and electrochemical
characterization of SAMs derived from chemisorption of compounds 5 and
6, as well as their pyridinium-protonated forms, will be presented elsewhere.

930 Organometallics, Vol. 27, No. 5, 2008
Engtrakul and Sita
Scheme 4a
a
(a) Pd(PPh3)2Cl2, CuI, THF/diethylisopropylamine (Hunig’s base) (1:1), 55° C, 24 h. (b) Pd(PhCN)2Cl2, CuI, P(t-Bu)3, diisopropylamine (1.2
equiv), THF, 50° C, 24 h.
Scheme 5a
Scheme 7a
a
(a) HCt CSi(Me)3, Pd(PPh3)2Cl2, CuI, THF/Hunig’s base (1:1), 50°
C, 24 h.
Scheme 6a
i
a
(a) (i) HCt CSi( Pr)3, Pd(PPh3)2Cl2, CuI, DIPA, rt, 24 h. (ii)
HCt CSiMe3, rt, 24 h. (b) 1 M NaOH, THF, MeOH, rt, 3 h.
Results and Discussion
(a) Synthesis of Diferrocene Frameworks. From the outset,
it was quickly realized that construction of the desired target
2,5-diethynylpyridine-linked diferrocene adsorbates required to
test the proposal of Scheme 1 would present some significant
synthetic challenges. The heart of these problems centered on
the need to identify a suitable ferrocene building block that could
be used to construct extended conjugated molecular frameworks
in an orthogonal (i.e., unsymmetric) fashion. Thus, while several
different structural classes of ferrocene with two symmetrically
disposed alkyne substituents in the 1,1′-positions have been
previously reported,13 to the best that we can ascertain, only
one family of unsymmetrically substituted 1,1′-dialkyne fer-
a
Pd(PPh3)2Cl2, Cu(OAc)2, THF, Hunig’s base, 90 °C, 24 h. (b) TBAF
(2.2 equiv), CH2Cl2, RT, 15 min. (c) (i) 11, Pd(PPh3)2Cl2, CuI, THF,
Hunig’s base, 55 °C, 24 h.
(13) (a) Doisneau, G.; Balavoine, G.; Fillebeen-Khan, T. J. Organomet.
Chem. 1992, 425, 113–117. (b) Ingham, S. L.; Khan, M. S.; Lewis, J.; Long,
N. J.; Raithby, P. R. J. Organomet. Chem. 1994, 470, 153–159. (c) Lavastre,
O.; Even, M.; Dixneuf, P. H.; Pacreau, A.; Vairon, J.-P. Organometallics
1996, 15, 1530–1531. (d) Long, N. J.; Martin, A. J.; Vilar, R.; White,
A. J. P.; Williams, D. J.; Younus, M. Organometallics 1999, 18, 4261–
4269. (e) Lindner, E.; Zong, R.; Eichele, K. Phosphorus, Sulfur Silicon
Relat. Elem. 2001, 168–169, 543–546. (f) Gonzalez-Cabello, A.; Vazquez,
P.; Torres, T. J. Organomet. Chem. 2001, 637–639, 751-–756. (g)
Schottenberger, H.; Lukassser, J.; Reichel, E.; Muller, A. G.; Steiner, G.;
Kopacka, H.; Wurst, K.; Ongania, K. H.; Kirchner, K. J. Organomet. Chem.
2001, 637–639, 558–576.
rocene derivatives is known.14 Further complicating potential
synthetic routes to these compounds is the known ease with
which 1,1′-diethynylferrocene apparently undergoes facile nu-
cleophilic attack at the ꢀ-carbon of an ethynyl group and
subsequent cyclization to produce a [4]ferrocenophane structure
(14) (a) Inouye, M.; Hyodo, Y.; Nakazumi, H. J. Org. Chem. 1999, 64,
2704–2710. (b) Inouye, M.; Itoh, M. S.; Nakazumi, H. J. Org. Chem. 1999,
64, 9393–9398.

Ferrocene-Based Nanoelectronics
Organometallics, Vol. 27, No. 5, 2008 931
Scheme 8a
a
(i) Pd(PPh3)2Cl2, Cu(OAc)2, THF, Hunig’s base, 90° C, 24 h. (ii) TBAF (2.2 equiv), CH2Cl2, rt, 15 min.
Scheme 9a
specifically, over time, a large number of coupling reactions
served to empirically show that, after introduction of the first
alkyne substituent starting with 10, the reactivity of the
remaining iodo group of the monocoupled product toward
subsequent Sonogashira coupling is greatly reduced for most
classes of alkynes, an observation apparently not reported upon
previously in the literature. On the other hand, this reduced
activity of the iodo group in 9 allows one to perform Sono-
gashira couplings of the acetylene moiety with a wide variety
of aryl iodides without any significant self-condensation po-
lymerization occurring. Thus, using S-acetyl-4-iodothiophenol
(11),19 ferrocene 12 could be prepared in high yield according
to eq 1 of Scheme 4. Unfortunately, attempts to couple 2 equiv
of 12 with 2,5-diethynylpyridine (13)20 failed to directly produce
the diferrocene 14 (see eq 2 in Scheme 4). Likewise, while
coupling of 2 equiv of 9 with 2,5-diiodopyridine (15) smoothly
provided the 2,5-diethynylpyridine-linked diferrocene 16 (eq 3
in Scheme 4), subsequent attempts to extend the molecular
framework of this compound through additional alkyne cou-
plings mostly came up empty handed, as exemplified by the
best of these experiments in which only a trace amount of the
diferrocene 17 could be identified after attempted coupling with
1,4-diethynylbenzene21 using Buchwald’s and Fu’s22 (PhCN)2-
PdCl2/P(t-Bu)3 catalyst combination according to eq 4 of
Scheme 4. Interestingly, this general lack of reactivity of the
iodo group of 1-alkynyl,1′-iodo ferrocenes toward alkyne
coupling did not appear to extend to silylacetylenes, as shown
by the ability to convert diferrocene 18 (Vide infra) into the
alkyne-coupled product 19 in high yield as shown in Scheme
5. Unfortunately, all attempts to desilylate 19 for further
framework extension led to a complex product mixture,
potentially as a result of a similar tendency of the deprotected
alkyne moiety to undergo nucleophilic attack as previously
depicted for 1,1′-diethynylferrocene in Scheme 2.
a
(a) 11, Pd(PPh3)2Cl2, CuI, THF, Hunig’s base, 55° C, 24 h. (b) (i)
MeI, CH2Cl2/CH3CN, 50° C, 18 h. (ii) H2O, NH4PF6.
according to Scheme 2.15 Consideration of this latter fact thus
ultimately led us to settle upon 1-ethynyl-1′-iodoferrocene
(9)16,17 as the basic ferrocene-containing building block. Com-
pound 9 has been prepared in low yield through a three-step
process starting from 1,1′-diiodoferrocene (10);18 however, for
obtaining large quantities of this desired material, the two-step
route shown in Scheme 3 proved to be highly efficient. In this
process, a substoichiometric amount of triisopropylsilylacetylene
and dilute conditions were used in palladium-catalyzed Sono-
gashira coupling to first produce a crude product containing
unreacted 10 and a mono-silylacetylene-coupled intermediate,
which was not isolated, but rather, desilylated with tetrabutyl-
ammonium fluoride (TBAF) to produce a mixture of 9 and 10
that could be easily separated by column chromatography on
silica gel. In this fashion, pure 9 was reproducibly obtained in
a 60% yield based on recovered 10.
Although all synthetic routes to the desired diferrocene
adsorbates now looked doomed, a fortunate decision was made
to address another potential problem associated with the final
target molecules, namely, that of limited solubility. On the basis
of literature precedence, it was anticipated that as the conjugated
molecular frameworks became more extended, they should also
become increasingly less soluble. To compensate for this lack
of solubility, a common practice is to place solubilizing
substituents, in the form of either linear alkyl or alkyl ether
chains, on phenyl rings.23 In the present case, methoxy groups
were adopted for this purpose given the ease with which they
In retrospect, the ease with which compound 9 could be
prepared through mono-Sonogashira coupling of 10 with an
acetylene, without any significant coproduction of the possible
1,1′-dialkyne biscoupled product, revealed early on a critical
disadvantage associated with using 9 and this coupling chemistry
for construction of the targeted ferrocene frameworks. More
(19) Pearson, D. L.; Tour, J. M. J. Org. Chem. 1997, 62, 1376–1387.
(20) Bunten, K. A.; Kakkar, A. K. Macromolecules 1996, 29, 2885–
2893.
(21) Bodwell, G. J.; Miller, D. O.; Vermeij, R. J. Org. Lett. 2001, 3,
2093–2096.
(15) Pudelski, J. K.; Callstrom, M. R. Organometallics 1994, 13, 3095–
3109.
(22) Hundertmark, T.; Littke, A. F.; Buchwald, S. L.; Fu, G. C. Org.
Lett. 2000, 2, 1729–1731.
(16) Butler, I. R.; Boyes, A. L.; Kelly, G.; Quayle, S. C.; Herzig, T.;
Szewczyk, J. Inorg. Chem. Commun. 1999, 2, 403–406.
(17) For syntheses of the bromo analogue of 9, see ref 16. Rosenblum,
M.; Brawn, N. M.; Ciappenelli, D.; Tancrede, J. J. Organomet. Chem. 1970,
24, 469–477.
(23) (a) Creager, S.; Yu, C. J.; Bamdad, C.; O’Connor, S.; MacLean,
T.; Lam, E.; Chong, Y.; Olsen, G. T.; Luo, J. Y.; Gozin, M.; Kayyem, J. F.
J. Am. Chem. Soc. 1999, 121, 1059–1064. (b) Sikes, H. D.; Smalley, J. F.;
Dudek, S. P.; Cook, A. R.; Newton, M. D.; Chidsey, C. E. D.; Feldberg,
S. W. Science 2001, 291, 1519–1523. (c) Yu, C. J.; Chong, Y.; Kayyem,
J. F.; Gozin, M. J. Org. Chem. 1999, 64, 2070–2079.
(18) Butler, I. R.; Wilkes, S. B.; McDonald, S. J.; Hobson, L. J.; Taralp,
A.; Wilde, C. P. Polyhedron 1993, 12, 129–131.

932 Organometallics, Vol. 27, No. 5, 2008
Engtrakul and Sita
Scheme 10a
a
(a) Pd(PPh3)2Cl2, CuI, THF/Hunig’s base (1:1), 50° C, 24 h. (b) (2,5-Dimethoxyphenyl)ethyne, Pd(PPh3)2Cl2, Cu(OAc)2, THF, Hunig’s base,
60° C, 24 h. (c) (i) MeI, CH2Cl2, CH3CN, 50° C, 18 h, (ii) H2O, NH4PF6.
Scheme 11a
a
(a) Pd(PPh3)2Cl2, CuI, THF/Hunig’s base (1:1), 55° C, 18 h. (b) Pd(PPh3)2Cl2, Cu(OAc)2, THF, Hunig’s base, 55° C, 24 h. (c) TBAF (1.1
equiv), CH2Cl2, rt, 15 min. (d) 11, Pd(PPh3)2Cl2, CuI, THF, Hunig’s base, 55° C, 24 h. (e) (i) MeI, CH2Cl2/CH3CN, 50 °C, 18 h. (ii) H2O, NH4PF6.
Scheme 12a
a
(a) Pd(PPh3)2Cl2, CuI, THF/Hunig’s base (1:1), 55° C, 18 h.
could be incorporated into synthetic schemes and the minimum
disturbance that they were expected to exert on packing within
SAM structures. Thus, the bisalkyne derivative 20 in Scheme
6 was first prepared from 1-bromo-4-iodo-2,5-dimethoxyben-
zene24 in a similar fashion to the method reported by Höger
and Enkelmann25 for similar compounds. Fortunately, with this
building block in hand, it was quickly discovered that, like
silylacetylenes, it was far more reactive in Sonogashira coupling
with 9, 10, and 1-alkynyl,1′-iodo ferrocenes in general, so that
1,1′-bisalkyne ferrocenes could now be reliably prepared. For
example, as Scheme 7 shows, coupling of 2 equiv of 20 with
10 now proceeded in modest yield to provide the symmetrically
derivatized ferrocene 21, which after clean desilylation provided
the deprotected bisalkyne 22. Compound 22 was then used in
turn to prepare dithioacetate-protected 1 through coupling with
2 equiv of 11.
The successful synthesis of the symmetric diferrocene
framework of 1 encouraged us to return to the synthesis of the
diferrocene thioacetates using the bisalkyne 20. As shown in
Scheme 8, coupling of 20 with the diferrocene core structure
(24) Orito, K.; Hatakeyama, T.; Takeo, M.; Suginome, H. Synthesis 1995,
10, 1273–1277.
(25) Hoger, S.; Enkelmann, V. Angew. Chem., Int. Ed. Engl. 1996, 34,
2713–2716.

Ferrocene-Based Nanoelectronics
Organometallics, Vol. 27, No. 5, 2008 933
Scheme 13a
a
(a) Pd(PPh3)2Cl2, CuI, THF/Hunig’s base (1:1), 55 °C, 18 h. (b) Pd(PPh3)2Cl2, Cu(OAc)2, THF, Hunig’s base, 55 °C, 24 h. (c) TBAF (1.1
equiv), CH2Cl2, rt, 15 min. (d) 11, Pd(PPh3)2Cl2, CuI, THF, Hunig’s base, 55 °C, 24 h. (e) (i) MeI, CHCl2/CH3CN, 50 °C, 18 h. (ii) H2O, NH4PF6.
Scheme 14a
a
(a) Pd(PPh3)2Cl2, CuI, THF/Hunig’s base (1:1), 55 °C, 18 h. (b) TBAF (1.1 equiv), CH2Cl2, rt, 15 min. (c) Diiodobenzene, Pd(PPh3)2Cl2, CuI,
THF/Hunig’s base (1:1), 55 °C, 18 h. (d) 11, Pd(PPh3)2Cl2, CuI, THF/Hunig’s base (1:1), 55 °C, 24 h.
Figure 1. Comparison of electronic spectra for 4 (blue) and 44
(red) in CH2Cl2.
Figure 2. Comparison of electronic spectra for the 1-ethynylpyri-
dine-1′-(dimethoxyphenyl)ethyne monoferrocenes, 28 (green) and
29 (red), and the 2,5-diethynylpyridyl-derivatized monoferrocenes,
45 (black) and 46 (blue), in CH2Cl2.
16now did indeed produce a reliable, although low yield,
quantity of the desired extended bisalkyne 23 after standard
desilylation. It is highly possible, however, that recently
reported modifications and advances in catalysts for the
Sonogashira coupling reaction may serve to considerably
increase the yield of 23 in the future.26 For now, the fact
that 23 could be easily separated from a small amount of a
regioisomeric mixture of monocoupled products through
column chromatography allowed us to carry on with the
synthesis of the final target molecules, 3 and 4 shown in
Scheme 9. Thus, 23 was coupled with 2 equiv of 11 to
provide 3 in reasonable yield, and 3 was then methylated to
provide 4 with a similar degree of success.
Chart 3
part of our previously published preliminary studies on 2,5-
diethynylpyridine-linked diferrocenes,6 caused us to consider what
(26) (a) See for instance: Thorand, S.; Krause, N J. Org. Chem. 1998,
63, 8551–8553. (b) Mio, M. J.; Kopel, L. C.; Braun, J. B.; Gadzikwa, T. L.;
Hull, K. L.; Brisbois, R. G.; Marworth, C. J.; Grieco, P. A. Org. Lett. 2002,
4, 3199–3202. (c) Jones, T. V.; Blatchly, R. A.; Tew, G. N. Org. Lett.
2003, 5, 3297–3299.
Inclusion of the methoxy groups in the structures of 3 and 4
that were necessitated for synthetic reasons, but which were not

934 Organometallics, Vol. 27, No. 5, 2008
Engtrakul and Sita
Table 1. Cyclic Voltammetry (CV) and Differential Pulse
Voltammetry (DPV) Analysis of 1,1′-Diethynyl-Derivatized
Ferrocenesa
mixture of 35 and its regioisomer 36. Fortunately, 35 could be
isolated in pure form through column chromatography in a 45%
yield, and single-crystal X-ray analysis was used to confirm that
its molecular structure was indeed the required regioisomer.
CV
DPV
b,c
b,c
compound
E1
E2
E1
E2
1/2
1/2
1/2
1/2
With 34 in hand, the synthesis of compound 6 was rapidly
finished according to the route shown in Scheme 13. More
precisely, coupling of 34 with 20, followed by desilylation and
coupling with 11, provided 6 in a 15% overall yield from 34.
Methylation then provided the pyridinium complex 8.
3
4
5
6
7
8
26
27
28
29
43
44
45
46
0.939d
0.944 (2)
1.04 (1)
0.861 (1)
0.877 (1)
0.950 (1)
1.11 (2)
0.941 (1)
0.937 (1)
1.21 (1)
1.09 (1)
0.818 (2)
0.929 (1)
1.01 (1)
1.14 (1)
1.04
1.23
1.25
1.23 (1)
0.987 (1)
0.975 (1)
1.26 (1)
e
e
0.958
1.10d
0.943
0.938
1.20
Finally, in order to compare the rate of electron transfer
through an intervening ferrocene moiety versus a phenyl group
within molecular frameworks of comparable length,12 the
thioacetate-protected ferrocene-terminated phenylethynyl oli-
gomer 42 was prepared in straightforward fashion according to
Scheme 14 through the intermediates 37-41.
1.08
0.840
0.955
1.01
1.12
1.90 (1)
(b) Electronic Spectra. As mentioned previously, methoxy
groups were added to the targeted diferrocence frameworks in
order to enhance solubility characteristics; however, it was
understood from the start that these formal electron-donating
groups might be noninnocent bystanders with regard to contri-
butions made to the overall electronic structure of the system.
In particular, for the pyridinium cations, 4, 7, and 8, it might
be expected that significant donor (2,5-dimethoxyphenyl)/
acceptor (pyridinium) interactions could influence rates of
electron transport across the intervening ferrocene moiety. Thus,
to begin to understand in what manner the methoxy groups
might be perturbing electronic structure, electronic (UV/vis)
spectra of these compounds were recorded and compared with
those taken of the model compounds, 28, 29, and the previously
reported species, 43–46.6 To begin, Figure 1 compares the
electronic spectrum of 44 with that of 4 as recorded in CH2Cl2.
As can be seen, the relatively strong metal-to-ligand charge
transfer (MLCT) absorption observed for 44 at 582 nm (ꢀmax
18 380) is greatly attenuated in 4, which exhibits a corresponding
MLCT at 600 nm (ꢀmax 2 760). No new charge transfer band is
observed for 4 that might be attributable to a direct donor–ac-
ceptor interaction between the 2,5-dimethoxyphenyl and pyri-
dinium groups. These data are taken as evidence that the 2,5-
dimethoxyphenylethynyl “arms” on each ferrocene in 4 actually
serve as electron-withdrawing substituents that decrease the
donor strength of the metal for MLCT. A comparison of the
electronic spectra for 28, 29, 45, and 46 shown in Figure 2
reveals a similar attenuation of the MLCT band between
corresponding structural pairs, and now one can also determine
what role the position of the nitrogen atom within the pyridinium
moiety plays in determining the strength of the MLCT. As can
be seen, when the pyridinium nitrogen is facing toward the
ferrocene as in 28, there is a significant bathochromic shift of
the MLCT band relative to the corresponding regioisomer 29
in which the pyridinium nitrogen is facing away from the
ferrocene [cf. for 28, λmax 554 nm (ꢀmax 2 200) vs for 29, λmax
496 nm (ꢀmax 1 890)]. This trend, which is also observed for
the 45/46 pair of compounds, reinforces our original proposal6
that, on the basis of the resonance scheme presented in Chart
3, a lower energy MLCT will be observed for the ferrocene
that is in an ortho relationship to the pyridinium cation acceptor
due to the “strong electronic communication” that exists by way
of the direct delocalization pathway shown (see structure B in
Chart 3). In corollary fashion, a ferrocene positioned in a meta
relationship to the pyridinium cation acceptor is in weaker, more
indirect, electronic communication.
1.13
a Recorded at 298
K
using 1 mM of the compound in 0.1 M
[n-Bu4N][B(C6F5)4] in CH2Cl2 using a glassy carbon working electrode, a
Pt wire counter electrode, and a Ag/Ag+ reference electrode. Values are
reported relative to the redox couple of
a known amount of
decamethylferrocene (Cp*2Fe) added as an internal standard. b Calculated
from Epk ) E1/2 - pulse amplitude/2. c Values in parentheses are the
integrated number of electrons based on the known amount of Cp*2Fe
added as an internal standard. d Two unresolved (overlapping) one-electron
e
processes. E1/2 values not determined since peaks were not fully
resolved by CV.
effect a donor–acceptor interaction within the conjugated frame-
work (i.e., between 2,5-dimethoxyphenyl and pyridinium groups)
might have on the electronic structure, and thus the electron
transport properties, of an intervening ferrocene moiety. To probe
this question, the regioselective syntheses of the model complexes
shown in Scheme 10 were developed. To begin, compound 9 was
first coupled with commercially available 2-bromopyridine to
provide the pyridyl ferrocene 24 in high yield. Under similar
conditions, however, 3-bromopyridine failed to couple to provide
the desired regioisomer. This compound, 25, was obtained by
coupling 9 with 3-iodopyridine27 instead. Next, both 24 and 25
were coupled with (2,5-dimethoxyphenyl)ethyne28 to provide the
regioisomeric compounds 26 and 27, respectively. Finally, com-
pounds 26 and 27 were methylated to produce the desired model
complexes 28 and 29 as shown in Scheme 10.
The regiospecific synthesis of the desired monothiol difer-
rocene adsorbates 5 and 6 for inclusion into SAMs proved to
be straightforward in the case of isomer 5, as shown in Scheme
11. Thus, coupling of the previously reported ferrocene pyridyl
bromide 306 with 9 first provided 18, which was then coupled
with 20 to generate product 31, which was subsequently
desilylated to yield the deprotected alkyne 32. Coupling of 32
with 11 finally provided the monothioacetate 5 in 65% yield,
and this compound was then methylated to produce the
pyridinium complex 7.
In sharp contrast to the straightforward route to 5, synthesis
of the regioisomer 6 proved to be slightly more challenging in
that an analogous coupling of 9 with the previously reported
ferrocene pyridyl bromide 336 failed to provide compound 34
according to Scheme 12. A successful route to 34 was eventually
found, however, by using the iodo analogue 35, which was
inelegantly, but effectively, prepared through coupling of
ethynylferrocene29 with 2,5-diiodopyridine30 to produce a
(27) Trecourt, F.; Breton, G.; Bonnet, V.; Mongin, F.; Marsais, F.;
Queguiner, G. Tetrahedron 2000, 56, 1349–1360.
(28) Buckle, D. R.; Rockell, C. J. M. J. Chem. Soc., Perkin Trans. 1
1985, 11, 2443–2446.
(30) Hama, Y.; Nobuhara, Y.; Aso, Y.; Otsubo, T.; Ogura, F. Bull. Chem.
Soc. Jpn. 1988, 61, 1683–1686.
(29) Polin, J.; Schottenberger, H. Org. Synth. 1996, 73, 262–269.

Ferrocene-Based Nanoelectronics
Organometallics, Vol. 27, No. 5, 2008 935
Figure 3. CV, scan rate 0.1 Vs-1 (top) and DPV (bottom) for (A) 3 and (B) 4 in 0.1 M [n-Bu4N][B(C6F5)4] in CH2Cl2 using a glassy
carbon working electrode. Potentials are referenced to the internal standard, Cp*2Fe.
(c) Solution Electrochemistry. A comprehensive investiga-
tion of the electrochemical properties of the newly synthesized
compounds, 3–8 [X ) COCH3 (Ac)], was conducted in order
to study the redox properties of the ferrocene units within these
extended conjugated frameworks. For comparison purposes, the
electrochemical properties of the model complexes 43–46 were
once again used. Table 1 presents data obtained from dc cyclic
voltammetry (CV) and differential pulse voltammetry (DPV)
analysis of the listed compounds, and Figure 3 presents in
graphical format the CV and DPV data obtained for compounds
3 and 4. To begin, it should be pointed out that since many of
the electrochemical studies involved in situ formation of di-
and tricationic species upon full oxidation, the non-nucleophilic
supporting electrolyte, [(n-Bu)4N][B(C6F5)4] in CH2Cl2, that was
introduced by Geiger and co-workers,31 was employed. Under
these conditions, all redox couples were observed to be either
reversible or quasi-reversible processes at a glassy carbon
working electrode as assessed by the width of peaks at half-
height.32 As shown in Figure 3A, compound 3 displays a single
unresolved redox couple for the two ferrocenes, which means
that the redox potentials of these two moieties must be within
a few tens of millivolts of one another.33 Viewed another way,
as previously determined for 43,6 the unsymmetric 2,5-diethy-
nylpyridyl unit of 3 is sufficiently long enough so that the two
ferrocene groups of 3 do not experience any significant
electostatic repulsion between one another upon oxidation,34 nor
does the unsymmetric nature of this bridging group preferentially
perturb the electronic structure of one ferrocene significantly
over the other. Upon placing a positive charge on the pyridyl
nitrogen, however, the asymmetry of the bridging pyridinium
cation in 4 has a profound differential influence on the two
ferrocene groups, resulting in two one-electron redox couples
now being observed with a ∆E1/2 value of 190 mV (see Figure
3B). Once again, the resonance structures of Chart 3 account
for the stronger perturbation imposed on a ferrocene moiety
that is positioned in an ortho relationship to the pyridinium
cation vis-à-vis a ferrocene that is in a meta position. In
comparing the redox potentials of 3 and 4 with those of 43 and
44, it is interesting, and important, to note that the conjugated
2,5-dimethoxyphenylethynyl “arms” in the former pair once
again appear to act as electron-withdrawing groups that serve
to shift their redox potentials to higher values relative to those
of the latter pair (see Table 1). This apparent electron-
withdrawing effect of a dimethoxy-substituted phenylethynyl
group tells an even more interesting tale when analyzing the
electrochemistry of the monothioacetate compounds, 5 and 6,
and their respective pyridinium derivatives, 7 and 8.
Figure 4 presents the CV and DPV data for compounds 5-8.
Not unexpectedly, given the unsymmetric nature of the com-
pound, the CV and DPV for 5 revealed two resolved one-
electron redox couples for the two ferrocene groups (∆E1/2
)
126 mV) (see Figure 4A). In going to the pyridinium structure,
this separation of redox couples increases even further to a
remarkable ∆E1/2 value of 310 mV for compound 7 as shown
in Figure 4B. In contrast, as Figure 4C and 4D present, in the
regioisomeric structures, a much smaller ∆E1/2 value is observed
in each case: 98 mV for 6 and only ∼50 mV for 8. While
initially somewhat surprising, the electrochemical behavior of
this latter pair of compounds is readily explained after considering
the redox properties of the regioisomeric model monoferrocene
compounds 28 and 29, along with those of 45 and 46. Thus, once
again due to the contributing resonance form B of Chart III, it was
established that a greater shift in the anodic direction of the half-
wave potential of a ferrocene resulted when the electron-withdraw-
ing pyridinium cation was in direct electronic communication with
the ferrocene (i.e., with the nitrogen atom “facing toward” the iron
center) as opposed to a weaker, more indirect communication (i.e.,
with the nitrogen atom “facing away from” the iron center) (see
Table 1). Additionally, as presented in Table 1, the CV and DPV
analyses for the two isomeric pyridinium cations, 28 and 29,
showed that an electron-withdrawing 2,5-dimethoxyphenylethynyl
(31) LeSuer, R. J.; Geiger, W. E. Angew. Chem., Int. Ed. 2000, 39, 248–
250.
(32) Bard, A. J.; Faulkner, L. R. Electrochemical Methods; John Wiley
& Sons, Inc.: New York, 2001.
(33) Barriére, F.; Geiger, W. E. J. Am. Chem. Soc. 2006, 128, 3980–
3989, references cited therein.

936 Organometallics, Vol. 27, No. 5, 2008
Engtrakul and Sita
Figure 4. CV, scan rate 0.1 V s-1 (top) and DPV (bottom) for (A) 5, (B) 7, (C) 6, and (D) 8 in 0.1 M [n-Bu4N][B(C6F5)4] in CH2Cl2 using
a glassy carbon working electrode. Potentials are referenced to the internal standard, Cp*2Fe.
diferrocene monothioacetates presented in Figure 5A, in which a
close correlation between the E1/2 values of 7 and 8 and those for
the pairs of monoferrocene model compounds, 28/45 and 29/46,
respectively, is observed. In particular, the DPV analyses for
compound 8 and the corresponding monoferrocene compounds 29
and 46 shown in Figure 5B clearly predict the presence of two
one-electron redox couples with a ∆E1/2 value of approximately
50 mV in the diferrocene structure. As a final comment, it should
also be noted that the greater separation in E1/2 values observed for
substituent on the cyclopentadienyl rings of the ferrocene resulted
the regioisomeric compound 7 is the result of the redox potential of
the internal ferrocene group being strongly influenced by both strong
electronic communication with the pyridinium cation and the electron-
withdrawing nature of the 2,5-dimethoxyphenylethynyl substituent.
in the half-wave potential of ferrocene shifting to higher potential
(anodic direction) relative to the half-wave potentials for the
corresponding monoferrocene isomers 45 and 46. Taken altogether,
these data allow for the structure/electrochemical analyses of the

Ferrocene-Based Nanoelectronics
Organometallics, Vol. 27, No. 5, 2008 937
Figure 5. (A) Structural/electrochemical comparison of compounds 7 and 8 and their corresponding monoferrocene models. (B) Comparison
of DPV data for 8 (black) and the two monoferrocene model compounds 29 (red) and 46 (blue), which predicts the two unresolved one-
electron redox couples in the diferrocene compound 8.
been able to document the “noninnocent bystander” status of
the 2,5-dimethoxyphenylethynyl moiety that can be globally
treated as an electron-withdrawing substituent that serves to shift
the Fe(II)/Fe(III) redox couple to higher potentials. This detailed
knowledge of solution electrochemical behavior of compounds
3–8 should benefit efforts directed toward the fabrication and
characterization of SAMs and LML heterojunctions in which
these diferrocene frameworks are incorporated.12
Conclusion
The present report serves several purposes in the quest for
ferrocene-based molecular components that might be suitable
for incorporation into future nanoscale electronic devices. First,
we have documented regiospecific synthetic routes that should
now make a variety of unsymmetric, phenylethynyl-conjugated
mono- and diferrocene molecular frameworks with end-to-end
distances of up to ∼4 nm that are suitably derivatized with
monothiol and R,ω-dithiol functionalities for their inclusion into
a variety of LML heterojunction test structures. Second, we have
Acknowledgment. This work was funded by the Basic
Energy Sciences division of the Department of Energy (DEF-
G0201ER15258) and in part by the NSF-NIRT (DMR-
0102950) and NSF-MRSEC (DMR-0080008) programs, for
which we are thankful.
(34) (a) Flanagan, J. B.; Margel, S.; Bard, A. J.; Anson, F. C. J. Am.
Chem. Soc. 1978, 100, 4248–4253. (b) Daum, P.; Murray, R. W. J. Phys.
Chem. 1981, 85, 389–396. (c) Rosenblum, M.; Nugent, H. M.; Jang, K. S.;
Labes, M. M.; Cahalane, W.; Klemarczyk, R.; Reiff, W. M. Macromolecules
1995, 28, 6330–6342. (d) Watanabe, M.; Nagasaka, H.; Ogata, N. J. Phys.
Chem. 1995, 99, 12294–12300. (e) Lee, W. Y.; Hostetler, M. J.; Murray,
R. W.; Majda, M. Isr. J. Chem. 1997, 37, 213–223. (f) Xu, G. L.; Crutchley,
R. J.; DeRosa, M. C.; Pan, Q.-J.; Zhang, H. X.; Wang, X.; Ren, T. J. Am.
Chem. Soc. 2005, 127, 13354–13363. (g) Xu, G. L.; Xi, B.; Updegraff,
J. B.; Protasiewicz, J. D.; Ren, T. Organometallics 2006, 25, 5213–5215.
(h) Sun, H.; Steeb, J.; Kaifer, A. E. J. Am. Chem. Soc. 2006, 128, 2820–
2821.
Supporting Information Available: Full experimental details,
including synthesis and analytical characterization of all new
compounds and solution electrochemical characterization. This
material is available free of charge via the Internet at http://pubs.acs.
org.
OM700807X