Synthesis and characterization of a highly reducing neutral "extended viologen" and the isostructural hydrocarbon 4,4″″-di-n-octyl-p- quaterphenyl

Article abstract of DOI:10.1021/ja053084q

The molecule 4,4″″-di-n-octyl-p-quaterphenyl was synthesized in one step by a nickel-catalyzed cross-coupling reaction. Powder X-ray diffraction shows that it crystallizes in a layered structure with the long axis of the molecule nearly perpendicular to t

Full text of DOI:10.1021/ja053084q

Published on Web 11/08/2005
Synthesis and Characterization of a Highly Reducing Neutral
“Extended Viologen” and the Isostructural Hydrocarbon
4,4′′′′-Di-n-octyl-p-quaterphenyl
William W. Porter III,† Thomas P. Vaid,*,† and Arnold L. Rheingold‡
Contribution from the Center for Materials InnoVation and Department of Chemistry,
Washington UniVersity, St. Louis, Missouri 63130, and Department of Chemistry, UniVersity of
California, San Diego, La Jolla, California 92093
Received May 11, 2005; E-mail: vaid@wustl.edu
Abstract: The molecule 4,4′′′′-di-n-octyl-p-quaterphenyl was synthesized in one step by a nickel-catalyzed
cross-coupling reaction. Powder X-ray diffraction shows that it crystallizes in a layered structure with the
long axis of the molecule nearly perpendicular to the layer plane. Differential scanning calorimetry indicates
a transition to a liquid-crystalline phase at 81 °C. Reaction of 4,4′-bis(4-pyridyl)biphenyl with 1-bromooctane
yields the dication 22+ 2Br-, an “extended viologen” isostructural with 4,4′′′′-di-n-octyl-p-quaterphenyl.
Reduction of 22+ 2Br- with sodium amalgam in DMF yields 2, the first neutral extended viologen to be
isolated. The molecule 2 is, to the best of our knowledge, the most reducing neutral organic molecule that
has been synthesized. Single-crystal X-ray diffraction shows that a diradical form, either singlet or triplet,
1
makes an important contribution to the electronic structure of 2. The broadened H NMR spectrum of 2
indicates the presence of a triplet, but it has not been possible to observe the triplet by ESR spectroscopy.
The electronic structure of 2 appears to be closely related to that of a classic molecule, Chichibabin’s
hydrocarbon.
Introduction
We recently examined the neutral form of phenyl viologen
(Figure 1b, R ) H) as an isostructural n-dopant for p-
quaterphenyl (Figure 1a, R ) H).1 An isostructural n-dopant is
a molecule that has the same skeletal structure as a molecular
semiconductor (p-quaterphenyl in this case) but has one or more
carbon atoms replaced by nitrogen, analogous to the replacement
of some silicon atoms by phosphorus in n-doped silicon. The
dopant thus formed must be neutral in charge and a good
electron donor. When a small amount of the isostructural
n-dopant is cocrystallized with the host molecular semiconductor
it will, ideally, donate electrons to the semiconductor and thus
n-dope it. Because the dopant is isostructural with the semi-
conductor molecule, it will cocrystallize substitutionally and
therefore will be spatially fixed within the crystal lattice, which
means that isostructurally doped molecular semiconductors will
be able to support the space-charge regions that are essential
components of traditional inorganic semiconductor devices.2 We
found, however, that neutral phenyl viologen is not a sufficiently
strong electron donor to act as a dopant in p-quaterphenyl. That
finding has led us to the current investigation, in which we have
synthesized a neutral “extended viologen” (2 in Figure 1c).
When a standard viologen in its neutral form gives up two
electrons to form a dication, two quinoid rings are converted to
aromatic rings (see Figure 1b), providing the driving force for
Figure 1. (a) p-Quaterphenyl (R ) H), 4,4′′′′-di-n-octyl-p-quaterphenyl
(R ) n-octyl); (b) phenyl viologen (R ) H), 1,1′-bis(4-n-octylphenyl)-4,4′-
bipyridinium (R ) n-octyl); (c) extended viologen 2 and its dication 22+
.
the reaction and making a neutral viologen a strong electron
donor. When the neutral extended viologen 2 in Figure 1c gives
up two electrons, four quinoid rings become aromatic, presum-
ably providing a stronger driving force and making it a stronger
electron donor than a standard neutral viologen.
† Washington University in St. Louis.
‡ University of California, San Diego.
(1) Porter, W. W., III; Vaid, T. P. J. Org. Chem. 2005, 70, 5028-5035.
(2) Vaid, T. P.; Lytton-Jean, A. K.; Barnes, B. C. Chem. Mater. 2003, 15,
4292-4299.
9
10.1021/ja053084q CCC: $30.25 © 2005 American Chemical Society
J. AM. CHEM. SOC. 2005, 127, 16559-16566
16559

A R T I C L E S
Porter III et al.
Scheme 1. Synthesis of (a) 4,4′′′′-Di-n-octyl-p-quaterphenyl, (b)
1
2+ 2Cl- and 1, and (c) 22+ 2Br- and 2.
Figure 2. Known extended viologens.
The molecular semiconductor we have chosen to investigate
is 4,4′′′′-di-n-octyl-p-quaterphenyl, shown in Figure 1a. While
R,ω-dialkyloligothiophenes have been thoroughly examined
as components of field-effect transistors,3-5 similarly function-
alized oligophenylenes are virtually unexplored. The octyl
groups of 4,4′′′′-di-n-octyl-p-quaterphenyl increase its solubility
and also make it possible to create an isostructural extended
viologen, 2 in Figure 1c. We have also examined the standard
viologen 1,1′-bis(4-n-octylphenyl)-4,4′-bipyridinium, 12+ (Figure
1b), and its neutral form 1 in order to contrast it with 2. The
molecules 1 and 2 are isomers and have the same skeletal
structure; the only difference between the two is that the
positions of two carbon atoms and two nitrogen atoms have
been interchanged. Yet we expect 2 to be a stronger reducing
agent than 1.
There are several known molecules in which two pyridylium
rings are connected by a π-conjugated fragment (Figure 2), and
all of these can be considered extended viologens. Dications
that have phenylene,6-8 thiophene,6,9,10 furan,6 and polyene11-13
units between the pyridylium rings have been synthesized.
Electrochemical studies of several of those derivatives have been
performed, but no extended viologen of any kind has been
isolated as a neutral molecule. Herein we report the synthesis,
isolation, and characterization of the neutral extended viologen
2, which we believe is the most reducing neutral organic
molecule that has been isolated to date. The molecule 2 also
has an interesting electronic structure, closely related to Chich-
ibabin’s hydrocarbon,14 as will be discussed below.
Results and Discussion
a straightforward nickel-catalyzed cross-coupling reaction15
between 4-n-octylphenylmagnesium bromide and 4,4′-dibro-
mobiphenyl. Its UV-vis absorption at 304 nm corresponds to
a HOMO-LUMO gap of 4.1 eV, similar to the value deter-
mined electrochemically (see below).
The viologen 12+ 2Cl- was prepared by a reaction that has
been previously employed for the synthesis of phenyl viologen16
and other aryl viologens. The reaction of 1,1′-bis(2,4-dinitro-
phenyl)-4,4′-bipyridilium dichloride (obtained from the reaction
of 4,4′-bipyridine with 1-chloro-2,4-dinitrobenzene) with 4-n-
octylaniline yields 12+ 2Cl- and 2 equiv of 2,4-dinitroaniline.
The reaction proceeds by a nucleophilic attack by the aniline
nitrogen on a carbon R to the nitrogen of the pyridinium ring,
resulting in an opening of the pyridinium ring, which is followed
by ring closing with release of 2,4-dinitroaniline.17 The ring-
Synthesis and Spectroscopy. Syntheses are outlined in
Scheme 1. The synthesis of 4,4′′′′-di-n-octyl-p-quaterphenyl is
(3) Garnier, F.; Yassar, A.; Hajlaoui, R.; Horowitz, G.; Deloffre, F.; Servet,
B.; Ries, S.; Alnot, P. J. Am. Chem. Soc. 1993, 115, 8716-8721.
(4) Garnier, F.; Hajlaoui, R.; El Kassmi, A.; Horowitz, G.; Laigre, L.; Porzio,
W.; Armanini, M.; Provasoli, F. Chem. Mater. 1998, 10, 3334-3339.
(5) Facchetti, A.; Yoon, M.-H.; Stern, C. L.; Hutchison, G. R.; Ratner, M. A.;
Marks, T. J. J. Am. Chem. Soc. 2004, 126, 13480-13501 and references
therein.
(6) Takahashi, K.; Nihira, T.; Akiyama, K.; Ikegami, Y.; Fukuyo, E. Chem.
Commun. 1992, 620-622.
(7) Kelley, C. J.; Ansu, K.; Budisusetyo, W.; Ghiorghis, A.; Qin, Y.; Kauffman,
J. M. J. Heterocycl. Chem. 2001, 38, 11-23.
(8) Valasek, M.; Pecka, J.; Jindrich, J.; Calleja, G.; Craig, P. R.; Michl, J. J.
Org. Chem. 2005, 70, 405-412.
(9) Nakajima, R.; Iida, H.; Hara, T. Bull. Chem. Soc. Jpn. 1990, 63, 636-
637.
(10) Albers, W. M.; Canters, G. W.; Reedijk, J. Tetrahedron 1995, 51, 3895-
3904.
(11) Happ, J. W.; Ferguson, J. A.; Whitten, D. G. J. Org. Chem. 1972, 37,
1485-1491.
(15) Tamao, K.; Sumitani, K.; Kiso, Y.; Zembayashi, M.; Fujioka, A.; Kodama,
S.; Nakajima, I.; Minato, A.; Kumada, M. Bull. Chem. Soc. Jpn. 1976, 49,
1958-1969.
(16) Kamogawa, H.; Sato, S. Bull. Chem. Soc. Jpn. 1991, 64, 321-323.
(17) Marvell, E. N.; Caple, G.; Shahidi, I. J. Am. Chem. Soc. 1970, 92, 5641-
5645.
(12) Arrhenius, T. S.; Blanchard-Desce, M.; Dvolaitzky, M.; Lehn, J.-M.;
Malthete, J. Proc. Natl. Acad. Sci. U.S.A. 1986, 83, 5355-5359.
(13) Nanasawa, M.; Miwa, M.; Hirai, M.; Kuwabara, T. J. Org. Chem. 2000,
65, 593-595.
(14) Tschitschibabin, A. E. Chem. Ber. 1907, 40, 1810-1819.
9
16560 J. AM. CHEM. SOC. VOL. 127, NO. 47, 2005

Synthesis of “Extended Viologen”
A R T I C L E S
opening and ring-closing reactions occur with a variety of para-
substituted anilines,18 and the method therefore should be
applicable to the synthesis of a variety of aryl viologens.
Recrystallization of 12+ 2Cl- from ethanol/ethyl acetate without
the exclusion of water results in the dihydrate 12+ 2Cl-‚2H2O.
The spectroscopy of 12+ 2Cl- is similar to that reported for
phenyl viologen dichloride.1 The chloride was converted to the
hexafluorophosphate to improve its solubility in THF for
electrochemical measurements.
Reduction of 12+ 2Cl- by zinc in ethanol yields dark red 1,
which can be recrystallized as thin plates from hot toluene.
Crystals of 1 maintain their dark red color for extended periods
in air, but red solutions of 1 rapidly turn deep green upon
exposure to air. Spectroscopy is similar to that of the reported
neutral phenyl viologen.1
The extended bipyridine 4,4′-bis(4-pyridyl)biphenyl was
synthesized by a slightly modified literature method19 in which
we used 4-iodopyridine in place of 4-bromopyridine.20 Its
reaction with 1-bromooctane in DMF yields 22+ 2Br-. An
extended viologen similar to 22+ with methyl groups in place
of the octyl groups has been reported,7 and like that methyl
derivative, 22+ is highly fluorescent. When 22+ 2Br- in ethanol
is excited at 348 nm it emits light at 427 nm with a quantum
yield of 0.95, determined versus an anthracene standard.21
The neutral extended viologen 2 is synthesized by the sodium
amalgam reduction of 22+ 2Br- in DMF. The reagent 22+ 2Br-
has low to moderate solubility in DMF at room temperature,
while 2 has essentially none, so the product crystallized from
the reaction mixture as the reaction proceeded. The product 2
can be purified by recrystallization from hot pyridine. Spec-
troscopy and further characterization of 2 are discussed below.
Electronic Structure of 2.
Figure 3. (a) Closed-shell, quinoid form of 2. (b) Diradical (singlet or
triplet) form of 2.
indicates the presence of unpaired electrons on the ring system
of 2 that are intrinsic to neutral 2, and that was investigated
further.
ii. Resonance Structures, Spin State, and Calculations.
Two resonance structures for 2 that should be considered are
shown in Figure 3. Figure 3a is the closed-shell, fully quinoid
structure that might be expected to be a good representation of
the electronic structure of 2. In standard neutral viologens, for
example, the spectroscopic properties and bond lengths from
X-ray crystallography are in full agreement with a quinoid
structure.1,22 However, if one imagines a neutral extended
viologen with a very long chain of phenylene rings, it becomes
clear that at some point the energetic stabilization gained by
converting a long chain of quinoid rings to aromatic rings will
be greater than the energy lost by breaking one π-bond, and a
diradical of the type shown in Figure 3b will become favored
over the fully quinoid structure. (In fact, in extended quinone
systems that consist of four rings a diradical structure is in
equilibrium with the quinoid structure, while three-ring and
smaller systems are quinoid and the five-ring system is almost
completely diradical.23,24) It appears that in the four-ring system
of 2 the diradical resonance structure makes an important
contribution to the electronic structure.
The diradical of Figure 3b can represent two different spin
states of the molecule: a singlet diradical or a triplet diradical.
The singlet diradical shown in Figure 3b, along with other
diradical resonance structures, would be a resonance structure
with the quinoid form of the molecule, and their weighted
average would describe the electronic structure of the singlet
state of the molecule. However, a singlet diradical should not
exhibit the broadened and shifted 1H NMR resonances observed
for 2. On the other hand, it is possible that another spin state of
2, a triplet diradical, is thermally accessible and is in equilibrium
i. NMR. The first indication that 2 has an unusual electronic
structure came from its H NMR spectrum, taken in pyridine-
1
d5 at 90 °C (see Supporting Information). A resonance for the
methyl group (of the octyl chain) is observed at 0.92 ppm, five
of the methylenes are overlapping at 1.30 ppm, and it appears
that the two remaining methylenes give rise to a resonance at
1.5 ppm and a broadened resonance at 6.4 ppm. There are no
observable resonances for the protons on the quinoid rings. The
identity of 2 was checked by oxidizing a sample with I2, which
yielded a compound with a 1H NMR spectrum consistent with
1
1
with the singlet and causes the broadening of the H NMR.
2
2+ 2I-. It is unlikely that the broadened and shifted H NMR
Density functional calculations of 2 (with methyl groups
substituted for the octyl groups) were performed with Gaussian
0325 at the UB3LYP/6-31G* level. If the spin state of 2 is set
to a singlet, geometry optimization yields a structure in which
all the rings are coplanar and the molecular orbitals are
consistent with the quinoid structure of Figure 3a. If the
optimization is started with the two outer rings twisted at 90°
relative to the inner rings, the first several cycles of the
optimization resemble a singlet diradical in which there is
unpaired spin density at both ends of the molecule. However,
as the geometry is optimized, the outer rings twist into the
configuration where all the rings are coplanar and the electronic
structure converts to the quinoid form. A calculation that
of 2 is due to adventitious oxidation of a small fraction of 2 to
2+ or 22+ and fast electron exchange of the cation or dication
with neutral 2. There are two reasons that adventitious oxidation
is an unlikely cause: (a) 1H NMR spectra of different samples
of 2, including ones from different preparations, are identical.
Adventitious oxidation would presumably occur to varying
extents and therefore cause varying amounts of broadening and
1
paramagnetic chemical shifts. (b) H NMR spectra of quinoid
1, which is also very air-sensitive in solution and would also
quantitatively react with any O2 present, are always sharp and
complete. Spectra of 1 and 2 were obtained under similar
conditions (in hot pyridine-d5) and were prepared in the same
1
glovebox. It is most likely that the H NMR spectrum of 2
(18) Marvell, E. N.; Shahidi, I. J. Am. Chem. Soc. 1970, 92, 5646-5649.
(19) Biradha, K.; Hongo, Y.; Fujita, M. Angew. Chem., Int. Ed. 2000, 39, 3843-
3845.
(20) Coudret, C. Synth. Commun. 1996, 26, 3543-3547.
(21) Scaiano, J. C., Ed. CRC Handbook of Organic Photochemistry; CRC: Boca
Raton, Florida, 1989.
(22) Bockman, T. M.; Kochi, J. K. J. Org. Chem. 1990, 55, 4127-4135.
(23) Rebmann, A.; Zhou, J.; Schuler, P.; Stegmann, H. B.; Rieker, A. J. Chem.
Res. (S) 1996, 318-319.
(24) Rebmann, A.; Zhou, J.; Schuler, P.; Rieker, A.; Stegmann, H. B. J. Chem.
Soc., Perkin Trans. 2 1997, 1615-1617.
(25) Frisch, M. J. T.; et al. Gaussian 03; Gaussian, Inc., 2004.
9
J. AM. CHEM. SOC. VOL. 127, NO. 47, 2005 16561

A R T I C L E S
Porter III et al.
Figure 4. (a) Closed-shell, quinoid form of Chichibabin’s hydrocarbon.
(b) Diradical (singlet or triplet) form of Chichibabin’s hydrocarbon.
included the command “guess ) mix”, which allows the mixing
of the quinoid and singlet diradical structures, was unstable and
did not find a minimum. A multireference theoretical method
such as CASSCF would be appropriate for further study of 2
in the singlet state. If the spin state of 2 is instead set to a triplet,
the minimized structure has a twist of 33° between the two inner
rings and a twist of 18° between the outer and inner rings. Also,
the bond distances in the triplet are closer to what would be
expected in a diradical structure. Interestingly, the energy of
the minimized triplet diradical is calculated to be 1.0 kcal/mol
lower in energy than the minimized closed-shell singlet. More
evidence for the triplet state was sought in the ESR spectrum.
iii. ESR. A singlet diradical should display no ESR spectrum,
while a triplet diradical has a distinctive ESR signature.26 As a
dilute solution in 1,4-dioxane, 2 exhibited no ESR signal.
However, when the solution was frozen at 0 °C, a single
resonance at g ) 2.0 was observed, indicating the presence of
a monoradical. In one trial its peak-to-peak width (of the
derivative spectrum) was 2 G, and in another trial it was 10 G.
The powder spectrum had a similar resonance at g ) 2.0 with
a peak-to-peak width of 10 G. In no case was a signal observed
at half the usual magnetic field, where the ∆Ms ) 2 transition
of a triplet is expected, nor were the additional resonances
around g ) 2.0 due to the zero-field splitting of a triplet
observed. The doublet (monoradical) signal may originate from
the one-electron reduction of impurities by 2, but it is surprising
then that the signal is not observed in the room-temperature
solution in dioxane.
The ESR results for 2 are puzzling, especially in light of the
solution-phase 1H NMR spectrum, which seems to indicate the
presence of a thermally accessible triplet. However, there is a
molecule with an electronic structure that appears to be closely
related to that of 2, and its ESR spectra have been equally
perplexing. The molecule is Chichibabin’s hydrocarbon (Figure
4), first reported in 1907.14 Both Chichibabin’s hydrocarbon and
2 can be written as closed-shell singlets with their two central
rings in a quinoid form, or they can be written as diradicals
with two aromatic central rings. The electronic state of Chich-
ibabin’s hydrocarbon (singlet, triplet, or an equilibrium mixture
of the two) was a subject of study for decades, mainly by ESR
spectroscopy.27,28 References 27 and 28 summarize the results
of those investigations and give relevant references. Solution-
Figure 5. ORTEP representation of 2.
phase ESR spectra of Chichibabin’s hydrocarbon always display
a signal due to a monoradical, probably due to dimerization of
the diradical or reaction of one end of the diradical with
adventitious impurities. ESR studies of powdered samples by
Brauer showed the presence of a triplet, and the conclusion of
those investigations was that the singlet and triplet were in
thermal equilibrium with the triplet 5.5 kcal/mol higher in energy
than the singlet.29-31 However, later attempts to observe the
triplet ESR signal of Chichibabin’s hydrocarbon were unsuc-
cessful.28
iv. X-ray Crystal Structure of 2. Crystal data is given in
the Supporting Information. Single crystals of 2 were grown
by slow cooling of a hot pyridine solution. In solution, 2 is
blue, while the crystals are metallic green. Interestingly, but
probably coincidentally, Chichibabin’s hydrocarbon forms
similar blue-violet solutions and metallic green crystals.28 Single-
crystal X-ray diffraction revealed the structure shown in Figure
5. All four of the rings are coplanar. There is a slight
pyramidalization at the nitrogens, so that the nitrogen-meth-
ylene bond is at a 9° angle to the plane of the rings. It is the
bond lengths in 2 that lend the most insight into its electronic
structure. Table 1 lists selected bond lengths of 2 along with
relevant bond lengths from other compounds for comparison.
The analogous bonds lengths are given for neutral methyl
viologen22 (which has a fully quinoid structure), phenyl viologen
dication1 (in which both rings are aromatic), and phenyl viologen
radical cation1 (which is an average of the quinoid and aromatic
resonance structures). The central C(3)-C(3) bond of 2, for
example, at 1.438(3) Å, is distinctly longer than the central
double bond of neutral methyl viologen at 1.363(9) Å and
shorter than the single bond of phenyl viologen dication at
1.485(3) Å. It is closest in length to phenyl viologen radical
cation’s central bond at 1.427(4) Å. A similar pattern is seen
throughout the central two rings of 2. If one assumes that a
closed-shell singlet electronic structure would yield the normal
pattern of alternating single and double bond lengths, then the
overall set of bond lengths implies a contribution of at least
50% from the singlet diradical resonance form shown in Figure
3. It seem unlikely that in the solid state, where all the rings
are coplanar, that the triplet state is occupied to any significant
degree. The similarity of the bond lengths of 2 and those of
(26) Wertz, J. E.; Bolton, J. R. Electron Spin Resonance; Chapman and Hall:
New York, 1986.
(29) Brauer, H. D.; Stieger, H.; Hartmann, H. Z. Phys. Chem. 1969, 63, 50-
(27) Platz, M. S. In Diradicals; Borden, W. T., Ed.; Wiley: New York, 1982;
pp 201-209.
65.
(30) Stieger, H.; Brauer, H. D. Chem. Ber. 1970, 103, 3799-3810.
(28) Montgomery, L. K.; Huffman, J. C.; Jurczak, E. A.; Grendze, M. P. J. Am.
Chem. Soc. 1986, 108, 6004-6011.
(31) Sartorius, R.; Reinsch, E. A.; Brauer, H. D. Z. Phys. Chem. 1974, 92, 281-
284.
9
16562 J. AM. CHEM. SOC. VOL. 127, NO. 47, 2005

Synthesis of “Extended Viologen”
A R T I C L E S
Table 1. Comparison of Bond Lengths in 2, Chichibabin’s Hydrocarbon,28 Neutral Methyl Viologen,22 Phenyl Viologen Radical Cation,1
Phenyl Viologen Dication,1 and the Calculated (UB3LYP/6-31G*) Optimized Structure of 2
distance (Å)
2
Chichibabin’s
Hydrocarbon
Methyl
Viologen0
Phenyl
Viologen+
Phenyl
Viologen2
2
+
bond
(X-ray)
(calcd)
C(3)-C(3)
C(2)-C(3)
C(3)-C(4)
C(1)-C(2)
C(4)-C(5)
C(5)-C(6)
C(1)-C(6)
C(6)-C(9)
C(8)-C(9)
C(9)-C(10)
C(7)-C(8)
C(10)-C(11)
C(11)-N(1)
N(1)-C(7)
1.438(3)
1.421(2)
1.425(2)
1.360(2)
1.361(2)
1.426(2)
1.427(2)
1.416(2)
1.441(2)
1.432(2)
1.347(2)
1.346(2)
1.3800(19)
1.3699(19)
1.448(4)
1.420(3)
1.420(3)
1.371(3)
1.372(3)
1.429(3)
1.424(3)
1.415(3)
1.363(9)
1.461(6)
1.444(7)
1.325(8)
1.336(8)
1.427(4)
1.422(4)
1.424(4)
1.348(4)
1.349(4)
1.485(3)
1.396(2)
1.391(2)
1.374(2)
1.376(2)
1.428
1.435
1.435
1.369
1.369
1.437
1.437
1.414
1.447
1.447
1.357
1.357
1.386
1.386
Chichibabin’s hydrocarbon is striking. For that molecule it was
estimated that the contribution from the diradical resonance form
is about 60%.28 The final column of Table 1 lists the bond
distances of the minimized structure from the density functional
calculations referred to above, with the spin state of 2 set to a
singlet. Although the calculated HOMO of 2 looks exactly as
would be expected for a fully quinoid structure (see Supporting
Information) and there is no indication of a singlet diradical,
the calculated bond distances are in reasonable agreement with
the experimental values.
v. Summary. The electronic structure of 2 is somewhat
perplexing, with its 1H NMR spectrum indicating the presence
of at least some triplet diradical (and DFT calculations reinforc-
ing the possibility, with the triplet state at a energy comparable
to the singlet), yet no triplet could be detected in the ESR
spectrum. X-ray crystallography indicates some singlet diradical
contribution to the electronic structure of 2 in the solid state.
The classic molecule Chichibabin’s hydrocarbon is closely
related to 2, and further investigation of both by ESR spectros-
copy and modern electronic structure calculations may be
warranted.
Electrochemistry. As explained in the Introduction, our
original motivation for the synthesis of 2 was our belief that it
would be a stronger reducing agent than a regular viologen such
as 1. That hypothesis was tested by solution-phase electro-
chemical measurements. Cyclic voltammograms of 4,4′′′′-di-
n-octyl-p-quaterphenyl, 12+ 2PF6-, and 22+ 2PF6 as saturated
-
solutions in THF are shown in Figure 6. All are referenced to
ferrocene0/+ at 0 V. Two reversible, one-electron reductions of
4,4′′′′-di-n-octyl-p-quaterphenyl occur at -2.83 and -3.05 V,
and a nonreversible oxidation was observed at approximately
+1.2 V. Two reversible, one-electron reductions of 12+ occur
at -0.70 and -1.00 V. The two compounds exhibit electro-
chemical behavior very similar to that of their respective parent
compounds, p-quaterphenyl and phenyl viologen.1 One revers-
ible, two-electron reduction (or two very closely spaced one-
electron reductions) of 22+ occurs at -1.48 V followed by two
reversible, one-electron reductions (to 2- and 22-) at -2.49 and
-2.86 V. The fact that the reduction of 22+ to 2 appears to
occur in a single, two-electron step makes sense in the light of
the greater spatial separation (and decreased interaction) of the
pyridinium moieties compared to 1. Single-potential, two-
Figure 6. Cyclic voltammetry of 4,4′′′′-di-n-octyl-p-quaterphenyl (top), 1
(middle), and 2 (bottom). All in THF, with potentials relative to ferrocene0/+
.
electron reductions and oxidations of organic molecules tend
to occur for larger molecules and for those that undergo a
structural rearrangement in one or both electron-transfer steps.32
The electrochemical data can be combined to create an
approximate picture of the solid-state energy levels of 4,4′′′′-
(32) Evans, D. H.; Hu, K. J. Chem. Soc., Faraday Trans. 1996, 92, 3983-
3990.
9
J. AM. CHEM. SOC. VOL. 127, NO. 47, 2005 16563

A R T I C L E S
Porter III et al.
di-n-octyl-p-quaterphenyl doped with 1 or 2. The difference
between the reduction and oxidation potentials of 4,4′′′′-di-n-
octyl-p-quaterphenyl indicate that as a molecular solid it had a
band gap of approximately 4.0 eV. Cocrystallization of 4,4′′′′-
di-n-octyl-p-quaterphenyl with a small amount of 1 would create
a donor level about 1.8 eV from the conduction band of 4,4′′′′-
di-n-octyl-p-quaterphenyl (from redox potentials 2.83-1.00 V).
While 2 is a better electron donor than 1 by 0.48 V (-1.48 V
versus -1.00 V), it will create a donor level in 4,4′′′′-di-n-octyl-
p-quaterphenyl that is still about 1.3 eV from the conduction
band, so 2 is not expected to function as an electron donor in
4,4′′′′-di-n-octyl-p-quaterphenyl.
Our attempt to create a good electron donor in 2 was
successful in that it is a better donor than the highly reducing
neutral viologen 1 by 0.48 V. In fact, to the best of our
knowledge, 2 is the most reducing neutral organic molecule yet
synthesized. The reducing strength of a molecule can be judged
by its (0/+) redox potential. There are transition metal orga-
nometallic complexes, such as the permethylated, 19-electron
FeCp*(η6-C6Me6) and CoCp*2 (Cp* ) η5-C5Me5), that are
stronger reducing agents, with (0/+) redox potentials of -2.30
V (vs Fc0/+, in DMF) for FeCp*(η6-C6Me6) and -1.94 V (vs
Fc0/+, in CH2Cl2) for CoCp*2.33 The (0/2+) potential of -1.48
V for 2 puts its reducing strength between that of CoCp*2 and
CoCp2. There are also organic anions that are more reducing
Figure 7. Current-voltage plots for pressed polycrystalline samples of 1,
mass 8, 16, 24, 32, and 40 mg (top). Measured resistance of the five samples,
with a best-fit line extrapolated to zero mass to yield the contact resistance
(bottom).
than 2: naphthalene has a (0/-) couple of -3.10 V (vs Fc0/+
,
in THF),33 so sodium naphthalide is an extremely strong
reductant. But among neutral organic molecules, 2 appears to
be the strongest electron donor.
disk (0.317 cm2) to calculate the sample thickness. An average
conductivity of 4.3 × 10-8(0.6 × 10-8) Ω-1 cm-1 was
calculated from 20 measurements (four times each on five
samples).
Current-voltage plots for 2 are qualitatively similar to those
of 1 and are given in the Supporting Information. The data were
treated in the same manner as that for 1. In this case the contact
resistance is 26 kΩ, which was subtracted from all measured
resistances for calculation of the conductivity. The density
calculated from the crystal structure (1.176 g/cm3) was used to
calculate sample thicknesses. An average conductivity for 2 of
9.3 × 10-7(0.9 × 10-7) Ω-1 cm-1 was calculated from 20
measurements.
Solid-State Conductivity of 1 and 2. Although our main
interest in 1 and 2 was their possible utility as dopants for 4,4′′′′-
di-n-octyl-p-quaterphenyl, our finding that neutral phenyl vi-
ologen is an electrical conductor as a pure material1 led us to
examine the solid-state conductivity of 1 and 2. Measurements
were carried out as described previously.1 Because these were
two-point measurements, a series of sample masses for each
substance was measured to examine the influence of contact
resistance. The cross-sectional area of the pressed pellets in
the apparatus is constant, so the thickness of the sample, and
therefore its measured resistance, will increase in direct pro-
portion to the sample mass if the contact resistance is not
significant. At each mass, the same sample was repacked and
measured for a total of four measurements. Current was
measured as a function of applied potential from 0 V to +2 V
to -2 V to 0 V.
Similar to neutral phenyl viologen, 1 and 2 are good electrical
conductors for even-electron, undoped molecules; typical un-
doped, even-electron conjugated molecules have single-crystal
conductivities in the range from 10-15 to 10-18 Ω-1 cm-1 34
.
The conductivity of 1 is about an order of magnitude smaller
than that of neutral phenyl viologen,1 and it is not surprising
that the presence of insulating octyl groups in 1 decreases its
conductivity relative to the parent neutral phenyl viologen. The
conductivity of 2 is about 20 times higher than that of 1. To
compare the conductivity of 1 and 2, their solid-state structures
must be considered. Powder X-ray diffraction indicates that 1
packs in a layered structure in the solid state, similar to 4,4′′′′-
di-n-octyl-p-quaterphenyl (see below). The solid-state structure
of 2 can also be considered layered, with the layers stacking
along the a crystallographic direction and the molecules of 2
within those layers being very tilted with respect to a. Electrical
conductivity in 2 would occur mainly within the b-c plane.
Although the intermolecular contacts between molecules of 2
Figure 7 shows the results for 1. Each current-voltage plot
in the top graph is the average of the four runs at each mass;
all of them are linear, indicating Ohmic behavior. The inverse
of the slope of each plot (obtained by linear regression) is the
resistance. The resistance as a function of mass is shown in the
bottom graph. The horizontal error bars represent the estimated
weighing error, and the vertical error bars are the standard
deviation over the four measurements at each mass (representing,
essentially, the variation due to sample packing). Extrapolation
of a least-squares-fit line to zero mass yields the contact
resistance, which is a very small negative number (-35 kΩ),
which for the remaining calculations is assumed to be zero. The
conductivity of 1 was calculated by assuming a density of 1.18
g/cm3 and using the known cross-sectional area of the sample
(34) Gutmann, F.; Lyons, L. E. Organic Semiconductors; John Wiley & Sons:
(33) Connelly, N. G.; Geiger, W. E. Chem. ReV. 1996, 96, 877-910.
New York, 1967.
9
16564 J. AM. CHEM. SOC. VOL. 127, NO. 47, 2005

Synthesis of “Extended Viologen”
A R T I C L E S
Figure 8. Powder X-ray diffraction pattern of 4,4′′′′-di-n-octyl-p-quater-
phenyl; λ ) 1.542 Å.
in its layers and those between molecules of 1 in its layers are
different, and a comparison between the conductivity of the two
must be made with that fact in mind, the 20-fold higher
conductivity of 2 does seem to imply that hole transfer between
molecules of 2 in the solid state is particularly facile.
Solid-State Structure and Properties of 4,4′′′′-Di-n-octyl-
p-quaterphenyl. The molecule 4,4′′′′-di-n-octyl-p-quaterphenyl
has not been reported before. In fact, except for isolated reports
of the methyl35 and n-hexyl derivatives,36 no 4,4′′′′-di-n-alkyl-
p-quaterphenyls have been investigated. In contrast, the similar
R,ω-di-n-alkylsexithiophenes and R,ω-di-n-alkylquaterthiophenes
have been thoroughly examined as the active components of
thin-film transistors.5 The unfunctionalized parent compound
p-quaterphenyl does function in thin-film transistors,37 so it
seems that studies of 4,4′′′′-di-n-alkyl-p-quaterphenyls in thin-
film transistors may be fruitful, and we have shown in this paper
that they are very easily synthetically accessible. Furthermore,
as shown below, 4,4′′′′-di-n-octyl-p-quaterphenyl is arranged
in the solid state in the layered structure of most materials
currently in use in organic thin-film transistors.
Figure 9. Set of four molecules of 4,4′′′′-di-n-octyl-p-quaterphenyl in an
energy-minimized configuration, indicating a layer thickness of ap-
proximately 36.6 Å.
Powder X-ray diffraction of 4,4′′′′-di-n-octyl-p-quaterphenyl
(Figure 8) shows strong diffraction peaks at 2θ intervals of about
2.5°, with some absences. The diffraction pattern indicates a
layered structure with a layer thickness of 35.5 Å. Crystals of
4,4′′′′-di-n-octyl-p-quaterphenyl are thin plates and the layers
are presumably parallel to the two long dimensions of the plates.
When the diffraction experiment is performed, the plates are
preferentially lying flat on the sample holder, so only diffraction
peaks due to the interlayer spacings are observed. A layer
thickness of 35.5 Å is consistent with 4,4′′′′-di-n-octyl-p-
quaterphenyl molecules packed in two-dimensional sheets. A
set of four molecules in a configuration of minimal energy
determined by molecular mechanics (in Chem3D) is shown
in Figure 9, which can be viewed as a part of one sheet of
thickness 36.6 Å, in reasonable agreement with the X-ray
diffraction data. The related molecules R,ω-dihexylquater-
thiophene4 and R,ω-dihexylsexithiophene3 have similar solid-
state structures.
Figure 10. DSC curves for 4,4′′′′-di-n-octyl-p-quaterphenyl in both the
heating and cooling directions. Data from the third of three cycles is shown.
liquid crystalline mesophases at 81 °C and 99 °C. Melting occurs
at 284 °C, so there is a very broad temperature range over which
4,4′′′′-di-n-octyl-p-quaterphenyl exists as a liquid crystal. The
related molecule R,ω-dihexylquaterthiophene undergoes a simi-
lar transition at 84 °C and melts at 179 °C.4
Cocrystallization of 1 and 2 with 4,4′′′′-Di-n-octyl-p-
quaterphenyl. Both 1 and 2 were originally intended to be
dopants for 4,4′′′′-di-n-octyl-p-quaterphenyl, so their cocrystal-
lization with 4,4′′′′-di-n-octyl-p-quaterphenyl was investigated.
Slow cooling a hot solution of a 9:1 mixture 4,4′′′′-di-n-octyl-
p-quaterphenyl and 1 in toluene results in pale red plate-shaped
crystals, while a similar crystallization of 4,4′′′′-di-n-octyl-p-
quaterphenyl and 2 results in blue-green plate-shaped crystals.
Inspection of both materials under an optical microscope
revealed that all the crystals were the same color and each was
uniform in color, implying a random distribution of the dopant
molecules within the host. Treatment of a sample of each set
of cocrystals with ferrocenium hexafluorophosphate in pyridine-
d5 and integration of the 1H NMR (obtained at 95 °C) resonances
of 12+ or 22+ versus 4,4′′′′-di-n-octyl-p-quaterphenyl indicated
doping levels of 11% and 10% for 1 and 2, respectively. Powder
X-ray diffraction of sets of cocrystals was very similar to that
of pure 4,4′′′′-di-n-octyl-p-quaterphenyl, showing that the
layered structure of the host was maintained in the cocrystals.
Figure 10 shows the differential scanning calorimetry (DSC)
curves of 4,4′′′′-di-n-octyl-p-quaterphenyl. Runs from 25 to 260
°C and back to 25 °C were performed three times sequentially,
and the results from each run were similar. Data from the third
run is shown. The molecule is in a crystalline state at room
temperature and then undergoes transitions to two different
(35) Gilman, H.; Weipert, E. A. J. Am. Chem. Soc. 1957, 79, 2281-2283.
(36) Kovyrzina, K. A.; Tsvetkova, T. A. Zh. Org. Khim. 1977, 13, 2395-2398.
(37) Gundlach, D. J.; Lin, Y. Y.; Jackson, T. N.; Schlom, D. G. Appl. Phys.
Lett. 1997, 71, 3853-3855.
9
J. AM. CHEM. SOC. VOL. 127, NO. 47, 2005 16565

A R T I C L E S
Porter III et al.
There was an additional broad peak in the powder diffraction
pattern of both sets of cocrystals at 2θ ) 26.4°, corresponding
to a spacing of 3.37 Å. The origin of this extra peak is unknown,
although it is close to interplanar spacing in an aromatic
π-stacked material. Although substitutional cocrystallization of
both dopants with the host was successful, neither doped
material displayed any electrical conductivity, as would be
expected from the solution-phase electrochemical potentials.
of 2 is somewhat enigmatic, although it is clear that a diradical
form, either singlet or triplet, makes an important contribution
to the electronic structure. The electronic structure of 2 appears
to be closely related to that of the classic molecule Chichibabin’s
hydrocarbon.
Acknowledgment. We thank David Sloop and Jun Lang for
assistance in acquiring ESR spectra, Jeffrey Turner and Karen
Wooley for help in acquiring DSC data, and Brian Barnes and
Lev Gelb for help with density functional calculations. Funding
was provided by the National Science Foundation grant CHE
0133068, Research Corporation grant RI0697, and the Depart-
ment of Education (Graduate Assistance in Areas of National
Need Fund P200A000217).
Conclusions
The new molecule 4,4′′′′-di-n-octyl-p-quaterphenyl was syn-
thesized by a straightforward one-step procedure, and it is
representative of an unexplored class of molecules, the di-n-
alkyl-p-quaterphenyls, that may find use as the active compo-
nents of organic thin-film transistors. It is a liquid crystal over
a wide temperature range, from about 81 °C to its melting point
at 284 °C. The isostructural molecule 2 is a neutral “extended
viologen” and is the first neutral extended viologen to be
isolated. Our aim in synthesizing 2 was to create a molecule
that is a stronger reducing agent than its isomer 1, and indeed
2 is, to the best of our knowledge, the most reducing neutral
organic molecule that has been isolated. The electronic structure
Supporting Information Available: Synthetic, spectroscopic,
1
and other experimental details, H NMR of 2, crystal data for
2 (and a separate CIF file), current-voltage plots for 2, image
of the HOMO of 2, and a complete version of ref 25 are
provided. This material is available free of charge via the
Internet at http://pubs.acs.org.
JA053084Q
9
16566 J. AM. CHEM. SOC. VOL. 127, NO. 47, 2005