On the causes of potential inversion in 1,2,4,5-tetrakis(amino)benzenes

Article abstract of DOI:10.1021/jo902411b

(Graph Presented) Three 3,6-difluoro-1,2,4,5-tetrakis(amino)benzene compounds, bearing dimethylamino (1), piperidin-1-yl (3), or morpholin-1-yl (5) substituents, have been synthesized and subsequently defluorinated to give the corresponding 1,2,4,5-tetrak

Full text of DOI:10.1021/jo902411b

pubs.acs.org/joc
On the Causes of Potential Inversion in 1,2,4,5-Tetrakis(amino)benzenes
Christopher J. Adams,*,† Rosenildo C. da Costa,† Ruth Edge,‡
Dennis H. Evans,§ and Matthew F. Hood†
†School of Chemistry, University of Bristol, Bristol BS8 1TS, U.K., ‡EPSRC National EPR Service, School
of Chemistry, University of Manchester, Oxford Road, Manchester M13 9PL, U.K., and Department of
Chemistry, Purdue University, 560 Oval Drive, West Lafayette, Indiana 47907-2084
§
chcja@bris.ac.uk
Received November 11, 2009
Three 3,6-difluoro-1,2,4,5-tetrakis(amino)benzene compounds, bearing dimethylamino (1), piper-
idin-1-yl (3), or morpholin-1-yl (5) substituents, have been synthesized and subsequently defluori-
nated to give the corresponding 1,2,4,5-tetrakis(amino)benzene compounds 2, 4, and 6; the crystal
structures of compounds 1, 4, and 6 have been obtained. Cyclic voltammetry shows that all six
compounds will lose two electrons to form dications, and the use of suitable oxidizing agents has
allowed isolation and crystallographic characterization of the dications 22þ and 62þ (as [PF6]2 salts)
and 42þ (as a [I5][I3] salt). The separation ΔE between the loss of the first electron and the second
varies between compounds, from 0.23 V in 1 to 0.01 V in 6. Eletrochemical studies involving the use of
the noncoordinating electrolyte [Bu4N][B{C6H3(CF3)2}4] show that it is possible to increase this
separation, stabilizing the intermediate monocationic phase, and this has allowed the isolation and
crystallographic characterization of the radical salts 2[B{C6H3(CF3)2}4] and 4[B{C6H3(CF3)2}4], the
first radical cations of this family to be isolated. DFT studies of the ion pairing between oxidized
forms of 1 and 2 and anions imply that the location of the ion pairing is different in the two species.
Introduction
and more easily oxidized, and that, therefore, in situations
where an electron donor was required, such compounds
might fulfill this role (as indeed N,N,N0,N0-tetramethyl-1,4-
diaminobenzene sometimes does). This was demonstrated by
Staab and co-workers in 1986, when they reported the
synthesis and characterization of 1,2,4,5-tetrakis(dimethyl-
amino)benzene (2).2 In contrast to N,N,N0,N0-tetramethyl-
1,4-diaminobenzene, this compound did not form a radical
cation but underwent a two-electron oxidation at -0.03 V vs
SCE in acetonitrile. The resulting dication 22þ has 12 π-
electrons, which would render it antiaromatic were they all
conjugated, so it undergoes a structural distortion to create a
molecule that effectively consists of two 6 π-electron 1,3-
bis(dimethylamino)allyl cations, which are linked by single
It has been known for many years that aromatic compounds
bearing amino substituents are very electron-rich systems that
may be used as electron donors. The archetypal system is N,N,
N0,N0-tetramethyl-1,4-diaminobenzene, whose radical cation
is generated from the neutral species at an oxidation potential
of 0.17 V vs SCE and which is colloquially known as “Wur-
ster’s blue” after its discoverer and its color. Although oxida-
tion involves lossof the aromaticityassociated with the neutral
molecule, generation of the quinonoidal mono- and dications
places the nonbonding nitrogen lone-pair electrons into bond-
ing orbitals and is thus relatively easy to do.1
It stands to reason that the addition of further amino
groups to such a system would render it more electron-rich
(2) Elbl, K.; Krieger, C.; Staab, H. A. Angew. Chem., Int. Ed. Engl. 1986,
25, 1023–1024.
(1) Michaelis, L. Chem. Rev. 1935, 16, 243–286.
1168 J. Org. Chem. 2010, 75, 1168–1178
Published on Web 01/14/2010
DOI: 10.1021/jo902411b
r
2010 American Chemical Society

Adams et al.
JOCArticle
SCHEME 3. Buchwald-Hartwig Synthesis of 6
SCHEME 1a
aTwo-electron oxidation of 14 π-electron 1,2,4,5-tetrakis(amino)-
benzene compounds such as 2 generates dications composed of two
monocationic 1,3-diaminoallyl subunits, each containing 6 π-electrons,
linked by two carbon-carbon single bonds. To emphasise the simila-
rities to the N,N,N0,N0-tetramethyl-1,4-diaminobenzene dication, the
dication 22þ may be drawn in two quinonoidal forms each with two
localized positive charges.
on hexafluorobenzene.7 We therefore set out to synthesize
some of these compounds and to investigate their properties to
see whether they might be of use as electron donors.
Results and Discussion
SCHEME 2. Azophenine-Type Systems (Left) Become
Isoelectronic with 22þ upon Diprotonation
Synthesis. The Buchwald-Hartwig coupling of 1,2,4,5-
tetrabromobenzene with a variety of primary amines gives
tetrakis(secondary amines) in good yields. This was first
reported by Wenderski8 with 2,6-dimethyl- and 2,6-diiso-
propylanilines as the amines and greatly expanded upon by
Bielawski,9 who extended the range of substrates to include
sterically bulky alkylamines. However, the literature con-
tains only the synthesis of a single 1,2,4,5-tetrakis(tertiary
amino)benzene compound, in which 1,2,4,5-tetrabromoben-
zene was coupled with morpholine to yield the correspond-
ing tetrakis(morpholin-1-yl) compound 6 (Scheme 3).10
In order to extend the scope of this reaction, we examined
the reactions of 1,2,4,5-tetrabromo- and 1,2,4,5-tetraiodo-
benzene with a variety of secondary amines in the presence of
a range of different catalysts and under a variety of condi-
tions without being able to find another system which
effectively produces the desired product. The only successful
coupling reactions used morpholine as the amine to produce
the aforementioned compound 6; even other cyclic second-
ary amines such as thiomorpholine and piperidine did not
work. Generally, byproducts were observed that had re-
sulted from partial protodehalogenation of the benzene ring,
and which were thus di- or triaminated.
Given the failure of the palladium-catalyzed methodo-
logy, we were required to use the lithiation and protodefluori-
nation pathway outlined in Scheme 4, which was developed
by Sorokin and co-workers to prepare a variety of such
compounds with piperidine, pyrrolidine, dimethylamine,
and dibutylamine as substrates. The 3,6-difluorinated com-
pounds may be converted to the corresponding protio
derivatives by reaction with sodium biphenylide;6 using this
reaction sequence, we have synthesized the 1,4-difluorotetra-
kis(amino)benzene compounds 1, 3, and 5 and subsequently
defluorinated them to yield the corresponding compounds 2,
4, and 6 (Scheme 4).
bonds at the 1 and 3 positions so that there is no through-
conjugation (Scheme 1). The theoretical description of the
bonding in 22þ was deduced by Dahne and Leupold over 40
€
years ago3 and more recently elaborated upon by Braunstein
and co-workers,4 who synthesized some azophenine-type
systems with alkyl substituents.5 These complexes are readily
diprotonated, at which point they are isoelectronic with 22þ
(Scheme 2); 22þ is also isoelectronic with the family of 2,5-
diamino-p-benzoquinones, and the whole family may be
classified as coupled trimethines.3
Given the use of N,N,N0,N0-tetramethyl-1,4-diaminoben-
zene and similar 1,4-diamines as electron donors, it would
seem logical to use the more electron-releasing tetrakis-
(amino) systems in the same way, but this has not happened.
The reason for this is that there has been no generally
applicable high yielding synthetic route to these
compounds-2 was made in 27% yield by a route which
cannot be extended to derivatives containing substituents
other than methyl groups. However, modern synthetic
methodologies have provided two potential routes to such
compounds. The first is the well-known Buchwald-Hartwig
amination reaction, which involves the palladium-catalyzed
coupling of an aryl halide with an amine. The second
involves the protodefluorination of 1,2,4,5-tetrakis(amino)-
3,6-difluorobenzene compounds,6 which are simply 1,2,4,5-
tetrakis(amino)benzene compounds in which the hydrogen
atoms on the central benzene ring have been replaced by
fluorine atoms. This is a readily prepared class of compound,
being synthesized by the nucleophilic attack of lithium amides
Electrochemical Studies. Our primary motivation for
synthesizing these compounds was to investigate the effect
that the various substituents and substitution patterns would
have on their electrochemical properties. Cyclic voltammetry
€
(3) Dahne, S.; Leupold, D. Angew. Chem., Int. Ed. Engl. 1966, 5, 984–993.
(4) Braunstein, P.; Siri, O.; Taquet, J.-P.; Rohmer, M.-M.; Benard, M.;
Welter, R. J. Am. Chem. Soc. 2003, 125, 12246–12256.
ꢀ
(5) Siri, O.; Braunstein, P.; Rohmer, M.-M.; Benard, M.; Welter, R.
(7) Sorokin, V. I.; Ozeryanskii, V. A.; Borodkin, G. S.; Chernyshev, A. V.;
Muir, M.; Baker, J. Z. Naturforsch. 2006, 61b, 615–625.
(8) Wenderski, T.; Light, K. M.; Ogrin, D.; Bott, S. G.; Harlan, C. J.
Tetrahedron Lett. 2004, 45, 6851–6853.
ꢀ
J. Am. Chem. Soc. 2003, 125, 13793–13803.
(6) Sorokin, V. I.; Ozeryanskii, V. A.; Borodkin, G. S. Synthesis 2006, 97–
102.
(9) Khramov, D. M.; Boydston, A. J.; Bielawski, C. W. Org. Lett. 2006, 8,
1831–1834.
(10) Witulski, B.; Senft, S.; Thum, A. Synlett 1998, 504–506.
J. Org. Chem. Vol. 75, No. 4, 2010 1169

Adams et al.
JOCArticle
SCHEME 4. Synthesis of 1-6 from C6F6
both fluorine atoms (4.30 G) (there are many examples of
pairs of organic radicals which differ only in the substitution
of hydrogen for fluorine, and often the ratio |aF|/|aH| of the
coupling constants is around 2.0-3.0; however, these are
exclusively anionic radicals; in the case of 1 and 2, the ratio is
either 5 or 10 (aH for 1 was either 0.42 or 0.84 G), and we have
been unable to find any other similar such cationic pairs for
comparison; see ref 12 for further discussion) and all 24
hydrogen atoms of the methyl groups (3.50 G). The similar-
ity of the values to those reported for the radical 2þ, which
were a(N) = 3.57 G and a(Me) = 3.13 G,12 implies that the
unpaired electron is distributed similarly in both 1þ and 2þ.
TABLE 1. First and Second Oxidation Potentials (V vs SCE, CH2Cl2,
0.1 M [Bu4N][PF6]) of All Compoundsa
E°1, E°2
ΔE
E°1, E°2
ΔE
1
3
5
methyl
piperid-1-yl
0.430, 0.660 0.230
0.475, 0.625 0.150
morpholin-1-yl 0.685, 0.745 0.060
2
4
6
0.063, 0.171 0.108
0.110, 0.265 0.155
0.445, 0.455 0.010
aE° values were obtained by digital simulation of the experimental
data (see the Supporting Information for further details).
showed that all of them underwent the reversible loss of two
electrons, with varying degrees of separation ΔE (=E°2 -
E°1) between the two processes (Table 1).
The values show the expected trends. The fluorine atoms
move the oxidations of 1, 3, and 5 to more positive potentials
as compared to their protio analogues, and the oxidation
potentials of 5 and 6, with the less electron-donating mor-
pholinyl substituents, are 0.3-0.4 V higher than those of the
methyl and piperidinyl compounds. The most interesting
feature lies in the values of ΔE, which vary from compound
to compound, and which vary from two one-electron pro-
cesses in 1 to a single two-electron process in 6.
The phenomenon in which ΔE is reduced is known as
potential compression, and when ΔE e 0, potential inversion is
seen and both electrons are apparently lost at once.11 One of
the factors that can cause potential compression and inver-
sion is a significant geometric change following loss of the
first electron. Normally, it is harder to remove a second
electron from a compound than the first because of the
increased electrostatic attraction between a cation and an
electron compared to the neutral species and the electron;
this is equivalent to saying that the SOMO is stabilized in the
cation compared to its energy as the HOMO in the neutral
compound. However, if the geometric change undergone
following the first oxidation causes a change in the composi-
tion and energy of the molecular orbitals such that stabiliza-
tion of the SOMO of the cation is less than expected, then the
system will lose the second electron at a potential near that at
which the first is lost; should the SOMO actually be desta-
bilized (relative to the HOMO of the neutral compound),
then the second electron will be easier to remove than the first
and potential inversion will result. One reason for the smaller
value of ΔE for 2 than 1 could therefore be that 2 undergoes a
larger geometric change than 1 on undergoing one-electron
oxidation. This hypothesis was tested by measuring the EPR
spectrum of the radical cation 1þ (Figure 1), which revealed
coupling constants to all four nitrogen atoms (a = 3.50 G),
FIGURE 1. Simulated (bottom) and experimental (top) EPR spec-
tra of 1þ.
Therefore, we conclude that both radicals have similar elec-
tronic structures (and, by extension, geometric structures),
and that the difference in ΔE of 1 and 2 is not due to them
undergoing radically different structural changes. We were
unable to obtain well-resolved EPR spectra for 3þ or 5þ.
Another factor that may influence ΔE is ion pairing
between the cationic species generated upon oxidation and
the anion of the supporting electrolyte. The work, in parti-
cular, of Geiger and co-workers has demonstrated how
changing from electrolytes containing anions such as
[ClO4]- and [PF6]- to those with much larger noncoordinat-
ing anions such as [B(C6F5)4]- and [B{C6H3(CF3)2}4]- can
dramatically increase ΔE, separating two-electron processes
into two one-electron processes.13 This occurs because ion
pairing has an important effect in stabilizing charged species.
(12) Elbl-Weiser, K.; Neugebauer, F. A.; Staab, H. A. Tetrahedron Lett.
1989, 30, 6161–6164.
(13) Nafady, A.; Chin, T. T.; Geiger, W. E. Organometallics 2006, 25,
1654–1663.
(11) Evans, D. H. Chem. Rev. 2008, 108, 2113–2144.
1170 J. Org. Chem. Vol. 75, No. 4, 2010

Adams et al.
JOCArticle
TABLE 2. Oxidation Potentials for 2 (V vs SCE) with Different Solvents and Electrolytes [Bu4N]þA-
solvent
CH2Cl2
CH3CN
entry in Figure 2
a
b
c
d
e
f
A-
E°1
E°2
ΔE
[B{C6H3(CF3)2}4]-
-0.010
0.408
[PF6]-
0.070
0.171
0.101
[ClO4]-
0.120
0.165
[B{C6H3(CF3)2}4]-
0.042
0.053
0.011
[PF6]-
0.065
0.070
0.005
[ClO4]-
0.042
0.050
0.418
0.045
0.008
The instability of the monocation in systems that display
potential inversion can be quantified by using the free energy
of disproportionation, ΔdispG, associated with the reaction
2Bþ f B þ B2þ, where B is the analyte:
þ
Δ
dispG ¼ Δf GB þ Δf GB2þ -2Δf GB ¼ FΔE
Where no potential inversion is displayed, ΔE (and thus
dispG) is positive and the reaction stays on the left-hand
Δ
side; the monocation is stable. When there is potential
inversion, both ΔE and ΔdispG are negative, and the system
will spontaneously form dications. The electrostatic interac-
tion between an electrolytic anion and an analyte cation has
the effect of making ΔfG more negative (in other words, the
free energy of formation of the ion pair is more negative than
that of the individual ions because of the electrostatic inter-
action between them), and this effect will be greater (because
of the larger electrostatic effects) for the interaction of a
dication with an anion than for a monocation with an anion.
Thus, if the effect of ion pairing on ΔfGB2þ is to make it a great
deal more negative than ΔfGBþ, then this can render ΔdispG
negative and cause potential inversion.
In order to test whether this was the case, we studied the
electrochemistry of 2 in CH2Cl2 and CH3CN solutions with
different supporting electrolytes (Table 2 and Figure 2). This
showed that, in dichloromethane, when [Bu4N][ClO4] or
[Bu4N][PF6] was used, two closely spaced processes were
observed (ΔE = 0.045 or 0.101 V), but when [Bu4N][B{C6-
H3(CF3)2}4] was used, two separate processes were observed
(ΔE = 0.418 V). ESR measurements on the species generated
following the first process confirmed it to be the radical
cation 2þ because it had identical coupling constants to those
reported previously. In acetonitrile, no matter which electro-
lyte was used, essentially only one process was observed
(ΔE<0.011 V). The electrochemical reversibility of the
processes was dependent somewhat upon the scan speed
and the nature of the working electrode; the second wave
showed much better reversibility at a glassy carbon (as
opposed to a platinum) electrode and at slower scan speeds
(50 rather than 500 mV/s). The lower electrochemical rever-
sibility of the second oxidation compared to the first may be
caused by a larger structural change associated with the
second process than the first.
FIGURE 2. Oxidations of 2 at 50 mV s-1 in different combinations of
solvent and electrolyte: (a) CH2Cl2, [Bu4N][B{C6H3(CF3)2}4];(b) CH2-
Cl2, [Bu4N][PF6]; (c) CH2Cl2, [Bu4N][ClO4]; (d) MeCN, [Bu4N]-
[B{C6H3(CF3)2}4]; (e) MeCN, [Bu4N][PF6]; (f) MeCN, [Bu4N][ClO4].
dimethylhexane)14 and also confirms the fact that [ClO4]-
tends to cause a slightly smaller value of ΔE than [PF6]- in
systems with planar organic cationic π-radicals.15 [ClO4]-
is slightly smaller than [PF6]- (ionic radii of 0.290 and
0.301 nm, respectively) which may allow better ion pairing.
In acetonitrile, all three electrolytes give little potential
ꢁ
separation. This is consistent with the findings of Barriere
and Geiger,16 who demonstrated that in solvents of high
polarity such as acetonitrile with medium or large electrolyte
anions (such as [PF6]- or [B{C6H3(CF3)2}4]-) the dominant
effect on ΔE was the solvation of the cationic products of
electrolysis. Under these circumstances, varying the anion
has little effect on ΔE. They also found that with medium-
sized electrolyte anions (including [PF6]- and [ClO4]-) the
choice of solvent has relatively little effect upon ΔE, again
consistent with the data in Table 2, where the change in ΔE
upon moving from CH2Cl2 to MeCN is 0.1 V or less for [PF6]-
and [ClO4]- but more than 0.4 V for [B{C6H3(CF3)2}4]-.
These results clearly indicate that in dichloromethane ion
pairing has a large effect on the electrochemistry of 2, with
the increased ion pairing in the presence of [PF6]- and
[ClO4]- resulting in a potential compression which is re-
moved by the use of the noncoordinating electrolyte. The
ordering of ΔE with the three electrolytes studied ([ClO4]- <
[PF6]- < [B{C6H3(CF3)2}4]-) mirrors that found for the in-
organic system [Rh2(TM4)4]2þ (TM4 = 2,5-diisocyano-2,5-
(15) Bancroft, E. E.; Pemberton, J. E.; Blount, H. N. J. Phys. Chem. 1980,
84, 2557–2560.
(14) Hill, M. G.; Lamanna, W. M.; Mann, K. R. Inorg. Chem. 1991, 30,
ꢁ
(16) Barriere, F.; Geiger, W. E. J. Am .Chem. Soc 2006, 128, 3980–3989.
4687–4690.
J. Org. Chem. Vol. 75, No. 4, 2010 1171

Adams et al.
JOCArticle
˚
TABLE 3. Selected Bond Lengths (A) and Torsion Angles (°) for 1, 4,
and 6 (Σ(N1) = Sum of the Three Bonding Angles at N1, etc.)
1
4
6
C(1)-N(1)
C(3)-N(2)
C(1)-C(2)
C(2)-C(3)
C(1)-C(3A)
Σ(N1)
1.4267(15)
1.4158(15)
1.3899(17)
1.3950(17)
1.4110(16)
342.8
1.4340(13)
1.4266(13)
1.3976(14)
1.4005(14)
1.4141(14)
339.6
1.425(4)
1.432(4)
1.399(4)
1.401(4)
1.415(4)
342.7
Σ(N2)
C(2)-C(1)-C(3)-C(2)
347.0
1.54
340.1
1.15
339.6
0.66
Assuming that the geometric change experienced by 2 is
the same in all solvents, we can therefore say first that in
dichloromethane the potential compression is caused not by
the geometric change but by ion pairing (because it is
removed by the use of the noncoordinating anion), and
second that in acetonitrile it is caused by cation-solvent
interactions (because it is present even in the presence of the
noncoordinating anion, and there is expected to be little ion
pairing in this medium).
FIGURE 3. Molecular structure of 1 (50% probability ellipsoids).
The same numbering scheme has also been used for 4 and 6.
Structural Studies. Having established that in dichloro-
methane the reason for the potential compression for 2, 4,
and 6 is indeed ion pairing between electrolyte anion and
analyte cation, we reasoned that it might be possible to
isolate their radical cations by using an oxidant such as [Fc]-
[B{C6H3(CF3)2}4] (Fc = ferrocenium), which would transfer
the noncoordinating anion, whereas the dications might
be obtained with [Fc][PF6]. This was indeed the case,
and crystals of the radical cations 2þ and 4þ, as [B{C6H3-
(CF3)2}4] salts, and the dications 22þ and 62þ (as [PF6]2 salts)
and 42þ (as a [I5][I3] salt) were isolated and studied by X-ray
diffraction. The structures of the neutral compounds 1, 4,
and 6 were also obtained for the purposes of comparison. We
were unable to isolate oxidized versions of the fluorinated
compounds 1, 3, and 5; mass spectra suggested that decom-
position involving defluorination was occurring following
chemical oxidation. We note that the structure of 22þ as a
mixed triiodide-iodide salt was reported by Staab and co-
workers in their original communication.2
The neutral compounds 1, 4, and 6 all crystallized lying
across a crystallographic inversion center and therefore have
one-half of a molecule in the asymmetric unit, and there are
no significant differences between the metrics of the fluori-
nated compound 1 and the nonfluorinated compounds 4 and
6 (Table 3). In all three compounds, within the aromatic
central ring, the C-C distances C(1)-C(2) and C(2)-C(3)
are 1.39-1.40 A in length and the distance C(1)-C(3A)
between the two halves of the system is marginally longer at
1.41 A, but the differences are small. The sum of the angles at
the nitrogen atoms, indicating the degree of pyramidaliza-
tion and thus the hybridization (for a perfect sp3 atom, this
would be 328.5°), varies between 339 and 347°, with 1 having
the greater values. The torsion angle C(2)-C(1)-C(3A)-
C(2A) which indicates the planarity of the ring varies between
0 and 2°. Essentially, the systems are typical aminobenzenes,
and the structure of 1 is shown in Figure 3 as an example.
The structures of the dications 22þ, 42þ, and 62þ are
radically different to the neutral species. Two-electron
oxidation creates a system with 12 π-electrons, which would
be antiaromatic were they all conjugated. There is therefore a
structural distortion which prevents this happening, which is
FIGURE 4. ORTEP plot (50% probability ellipsoids) of the cation
22þ in the structure of [2][PF6]2, showing the close contact between
endo-methyl groups.
a separation of the molecule into what is effectively two 6 π-
electron 1,3-diaminoallyl cyanine cations (Scheme 1). These
are linked by two C-C single bonds which are long enough
to prevent the π-systems overlapping, and the two halves of
the molecule are also mutually twisted, again so that there is
no effective overlap of the π-electron clouds on each side.
The cation 22þ (Figure 4) lies on a crystallographic center
of symmetry. The C(1)-C(3A) bond which joins the two
halves of the ring is now 1.51 A long, an increase of 0.1 A
when compared to the neutral compounds. The carbon-
nitrogen distances have decreased from about 1.42 A in the
neutral compounds to about 1.32 A, consistent with the bond
order increasing from 1 in the neutral compounds to 1.5 in
the dications, while the allylic C-C distances along the sides
of the central ring are largely unchanged; the C-C bond
order in a 1,3-diaminoallyl cation is approximately the same
as in a benzene ring, so there is little change. The sums of the
three bond angles at the nitrogen atoms are 359°, indicating
sp2 hybridization, and the ring twist (Figure 5) as indicated
1172 J. Org. Chem. Vol. 75, No. 4, 2010

Adams et al.
JOCArticle
˚
TABLE 4. Selected Bond Lengths (A) and Angles (°) for 2[PF6]2,
4[I3][I5] I2, and 6[PF6]2 (Σ(N1) = Sum of the Three Bonding Angles at
3
N1, etc.)
2[PF6]2
4[I3][I5] I2
6[PF6]2
3
C(1)-N(1)
C(3)-N(2)
C(4)-N(3)
C(6)-N(4)
C(1)-C(2)
1.3219(19)
1.3192(19)
1.347(5)
1.320(5)
1.344(5)
1.308(5)
1.382(5)
1.425(6)
1.382(6)
1.420(5)
1.516(6)
1.506(6)
353.5
1.341(4)
1.332(3)
1.348(4)
1.327(4)
1.385(4)
1.419(4)
1.378(4)
1.412(4)
1.498(5)
1.505(4)
359.1
1.398(2)
1.406(2)
C(2)-C(3)
C(4)-C(5)
C(5)-C(6)
C(1)-C(6)
1.508(2)a
C(3)-C(4)
Σ(N1)
Σ(N2)
Σ(N3)
358.6
359.3
358.4
352.8
359.0
358.4
Σ(N4)
360.0
46.8
46.5
41.0
358.2
46.5
46.2
40.1
C(2)-C(3)-C(4)-C(5)
C(2)-C(1)-C(6)-C(5)
C(6)-C(1)-C(3)-C(4)
48.6a
41.6a
aIn this structure, because the cation lies on a crystallographic two-
fold axis, the atoms C(4), C(5), and C(6) should be replaced by the
symmetry generated atoms C(1A), C(2A), and C(3A), which are their
equivalents.
FIGURE 5. Side view of [2][PF]2, showing the twisting of the cation
and the location of the anions.
by the torsion angles is between 40 and 50°. There is a very
short contact (carbon-carbon distance of 3.07 A) between
the endo-methyl groups of neighboring NMe2 substituents;
in the neutral species, these would be twisted out of the
plane of the benzene ring, minimizing such interactions (see
Figure 3), but in the dication, the requirement for conjuga-
tion of the nitrogen lone pair with the central ring prevents
this. The radius of a methyl group is about 2 A,17 but it is not
perfectly spherical, and interleaving of the hydrogen atoms
in 22þ allows the short contact. Similar behavior, both in the
distortion of the aromatic framework and the close methyl-
methyl contacts, is seen upon oxidizing tetrakis(dimethyl-
amino)-p-benzoquinone17 and hexakis(dimethylamino)ben-
zene.18
FIGURE 6. Core of 62þ (50% ellipsoids), showing the bond-length
alternation (distances in A) and the atom numbering scheme used.
The structures of 42þ and 62þ show similar metrics to that
of 22þ, although in these cases, the cations are not on centers
of symmetry so there are more independent values (Table 4).
Interestingly, in both of these cases, there is some evidence of
bond-length alternation (the core structure of 62þ is shown in
Figure 6 as an example), with each diaminoallyl subunit
having one short and one long C-C distance (1.38 and
1.42 A, respectively), and though both C-N distances are
shorter than in the neutral compound, one is longer than the
other (1.30-1.33 vs 1.34-1.35 A). The overall implication is
that one of the quinonoid resonance forms shown in Figure 1 is
making a greater contribution to the structure than the other.
This is probably due to steric effects, the alternant bonding
allowing a greater distance between neighboring piperidyl or
morpholinyl termini than the delocalized bonding would.
The structures of the radical cations 2þ (Figure 7) and 4þ
are interesting (Table 5), in that although the structures of
several doubly oxidized polyaminobenzenes are known there
are apparently no published structures of singly oxidized
FIGURE 7. Crystallographic structure of cation A from poly-
morph I of [1][B{C6H3(CF3)2}4].
€
(17) Bock, H.; Ruppert, K.; Nather, C.; Havlas, Z. Angew. Chem., Int. Ed.
Engl. 1991, 30, 1180–1183.
radical species. Two different polymorphs of 2[B{C6H3(CF3)2}4]
were crystallographically characterized; the first (polymorph I)
€
€
(18) Speiser, B.; Wurde, M.; Maichle-Mossmer, C. Chem.;Eur. J. 1998,
4, 222–233.
J. Org. Chem. Vol. 75, No. 4, 2010 1173

Adams et al.
JOCArticle
˚
TABLE 5. Selected Bond Lengths (A), Angles, and Torsion Angles (°) for the Radicals [2][B{C6H3(CF3)2}4] and [4][B{C6H3(CF3)2}4] (Σ(N1) = Sum of
the Three Bonding Angles at N1, etc.)
[2][B{C6H3(CF3)2}4]
[4][B{C6H3(CF3)2}4]
polymorph I
polymorph II
independent molecule
A
B
C
D
E
F
A
B
N(1)-C(1)
C(1)-C(2)
1.383(3)
1.385(3)
1.402(3)
1.369(3)
1.445(3)
351.2
1.390(3)
1.386(3)
1.395(3)
1.366(3)
1.453(3)
349.1
1.384(3)
1.384(3)
1.401(3)
1.367(3)
1.447(3)
349.0
1.385(3)
1.386(3)
1.391(3)
1.376(3)
1.450(3)
351.0
1.376(3)
1.390(3)
1.389(3)
1.376(3)
1.448(3)
353.5
1.381(3)
1.390(3)
1.397(3)
1.375(3)
1.443(3)
352.9
1.400(4)
1.379(5)
1.409(5)
1.368(4
1.450(4)
345.2
1.398(4)
1.385(5)
1.407(5)
1.371(4)
1.444(4)
345.7
C(2)-C(3)
C(3)-N(2)
C(3)-C(1A)
ΣN (N1)
ΣN (N2)
C(2)-C(1)-C(3)-C(2)
355.9
5.3
354.6
5.4
356.2
6.3
353.5
6.4
352.3
6.4
352.9
5.7
357.3
4.0
355.8
3.7
FIGURE 8. Computed electrostatic potential surfaces for 2þ (left) and 22þ (right). Red = least positive, blue = most positive.
contains four half-molecules of 2þ (and two complete anions) in
the asymmetric unit, and the second (polymorph II) contains
two half-molecules of 2þ (and one complete anion) in the
asymmetric unit, and there are thus 6 available values for each
metric. The C-C distances range from 1.38 to 1.40 A, except for
the C(1)-C(3A) distance between the two halves of the system,
whichisintherangeof1.43-1.44 A;theC-N distances go from
1.37 to 1.39 A. It was predicted from the ESR spectrum of 2þ
that the nitrogen atoms were “close to planarity” and that there
would be a close contact between the endo-methyl groups of
neighboring NMe2 moieties.19 These inferences are supported by
the crystal structures of 2þ, in which the sum of the angles at
nitrogen varies from 349 to 356°, and the endo-methyl groups lie
close to each other (carbon-carbon distances between 3.06 and
3.14 A). The structure thus has aspects in common with those of
both the neutral compounds and the dications; the planarity of
the nitrogen atoms is characteristic of the latter, but the planarity
of the central ring and the relatively short C(1)-C(3A) bond are
reminiscent of the former. The shortness of this bond is probably
due to the high contribution of the p-orbitals on C(1) and C(3) to
the SOMO.5 The structure of 4þ has similar metrics to that of 2þ,
though there is again some bond-length alternation in the C-C
and C-N distances.
ion pairing on potential inversion appear to be limited to the
reduction of cyclooctatetraene (COT) in the presence of
LiClO4 as electrolyte, in which tight ion pairs between the
COT anions and lithium cations are important;20 we have
been unable to find any modeling studies on the ion pairing
between analyte cations and electrolyte anions such as [PF6]-.
The geometries of both 1 and 2 (as computationally simple
model compounds) were optimized in neutral, cationic, and
dicationic forms, and then solvation was modeled using the
PCM model. Electrostatic potential surfaces were then gene-
rated for the cations to locate the areas of maximum positive
charge on the molecular surface and hence gain some
indication of where any anions might be found. An anion
was then manually placed there, and the geometry of the ion
pair (or ion triplet for dication with two anions) optimized,
and the solvation model applied again.
The potential surfaces (shown for 2þ and 22þ in Figure 8)
reveal that, for the monocation 2þ, the maximum positive
charges are found on each side of the molecule, in the plane of
the ring and next to the central hydrogen atoms. In contrast,
for the monocation 1þ and both dications 12þ and 22þ, the
maximum positive charges are above and below the center of
the ring. Thus, two sites (side-on and face-on) were identified
as possible locations for the anions, and we proceeded to
optimize various ion pairs and triplets using these locations.
It was found that for the monocation 2þ it was possible to
optimize the geometry with an anion in either of the two
Computational Studies. Given that ion pairing does seem to
be the cause of the observed potential inversion in the oxida-
tion of these species, we sought further insight by means of a
computational study. Computational studies on the effect of
(19) Barth, T.; Neugebauer, F. A. J. Org. Chem. 1995, 60, 5401–5406.
1174 J. Org. Chem. Vol. 75, No. 4, 2010
(20) Fry, A. J. Electroanalysis 2006, 18, 391–398.

Adams et al.
JOCArticle
TABLE 6. Calculated Values of ΔdispG (kcal mol-1) for Disproportionation Reactions in Dichloromethane
reaction
1
2
A
PF6
ClO4
PF6
22.85
ClO4
i
ii
2Bþ f B2þ þ B
25.72
-2.03
-4.68
5.95
13.93
11.28
9.76
2(Bþ[A]-) f (B2þ[A]-) þ B þ [A]-
2(Bþ[A]-) f (B2þ2[A]-) þ B
Bþ[A]- þ Bþ f (B2þ[A]-) þ B
2Bþ þ [A]- f (B2þ[A]-) þ B
2Bþ þ 2[A]- f (B2þ2[A]-) þ B
2(Bþ[A]-) f B2þ þ B þ 2[A]-
6.48
6.55
10.09
13.70
13.77
18.50
9.69a/-2.79b
6.08a/-6.40b
12.14a/5.90b
14.59
7.92a/1.62b
7.09a/0.79b
11.98a/8.83b
16.04
15.21
14.73a/8.43b
iii
iv
v
vi
vii
10.98
17.95a/5.47b
aSide-on configuration of 2þ[A]. bFace-on configuration of 2þ[A].
TABLE 7. Calculated Values of ΔdispG (kcal mol-1) for Disproportionation Reactions in Acetonitrile
reaction
1
2
A
PF6
ClO4
PF6
ClO4
i
ii
2Bþ f B2þ þ B
19.93
-6.97
-4.01
5.73
18.43
21.39
-5.47
16.8
2(Bþ[A]-) f (B2þ[A]-) þ B þ [A]-
2(Bþ[A]-) f (B2þ2[A]-) þ B
Bþ[A]- þ Bþ f (B2þ[A]-) þ B
2Bþ þ [A]- f (B2þ[A]-) þ B
2Bþ þ 2[A]- f (B2þ2[A]-) þ B
2(Bþ[A]-) f B2þ þ B þ 2[A]-
1.45
7.32
10.04
18.63
24.50
2.75
5.08a/-8.06b
7.04a/-6.10b
11.72a/5.15b
18.36
3.42a/-3.52b
8.41a/1.47b
11.86a/8.39b
20.30
25.29
-0.08a/-7.02b
iii
iv
v
vi
vii
20.32
3.52a/-9.62b
aSide-on configuration of 2þ[A]. bFace-on configuration of 2þ[A].
locations, though the free energy of formation for the side-on
position wasmorenegative. In contrast, for1þ, the systemwas
only stable when the anion was in the face-on position; if the
optimization was started from the side-on position, the anion
migrated to the face-on location. For both the dications 12þ
and 22þ, stable ion pairs and triplets had one or two anions,
respectively, in the face-on positions above one or both faces
of the ring. Pleasingly, and implying that the calculations were
giving the correct location of the anions, the crystal structure
of [2][PF6]2 does indeed have the [PF6]- anions located above
and below the center of the ring (Figure 5).
Having calculated the free energy of formation for all
species, we were then able to calculate ΔdispG for various
disproportionation reactions of 1þ and 2þ (Table 6). For the
reaction 2Bþ f B2þ þ B, ΔdispG was found to be between 20
and 30 kcal mol-1; in other words, disproportionation is
highly disfavored.21
It is apparent from the data in Table 6 that only by
allowing ion-pairing of all cations do the calculations imply
that disproportionation might be favorable (reactions ii and
iii, in which ΔdispG is closest to 0). If there is no ion pairing,
then ΔdispG is on the order of 5-20 kcal mol-1 and the
radical cation is stable with respect to disproportionation.
The relative contribution of each of the reactions to the
observed solution behavior will depend upon the concentra-
tions of the species in solution, but overall, it is possible to say
that ion pairing is necessary for potential inversion to occur
for these molecules.
The data do not, however, definitively explain the ob-
served experimental behavior. For example, similar values
were obtained when acetonitrile was used as the solvent
model (Table 7), whereas the experimental data show that
in this solvent disproportionation occurs even in the absence
of ion pairing. This is likely to be due to inadequate modeling
of the highly directional acetonitrile-ion interactions by the
solvent continuum model used; the experimental work of
Geiger has shown these to be very strong,16 and even a small
fractional error in the calculation might cause the disagree-
ment seen.
One qualitative observation that can be made is that the
ion pairing in 1þ and 2þ is likely to be different; effectively,
the presence of the fluorine atoms in 1 makes the 3 and 6
positions of the central ring “nonstick”, and less effective ion
pairing for 1 may be the reason why it has a greater value of
ΔE than 2.
Conclusions. The studies reported herein have shed light
on the causes and the nature of the observed oxidations of
1,2,4,5-tetrakis(amino)benzenes. The original report by
Staab included an electrochemical study in acetonitrile,
under which conditions a single two-electron oxidation is
seen,2 and this has subsequently been attributed to the
geometric changes observed on changing from neutral com-
pound to dication. While the change in orbital energies
associated with the geometry change is undoubtedly a factor
in determining the electrochemical behavior, we have shown
herein that the separation of the two oxidations is also both
solvent- and electrolyte-dependent, and that these two fac-
tors must therefore also play a part in determining ΔE. Using
the combination of a noncoordinating anion and a nonpolar
solvent shows that in the system studied herein geometric
change on its own is not sufficient to cause potential inversion,
and that extra stabilization of the dication (by ion pairing or
strong solvent-molecule interactions) is required for this
phenomenon to be observed. This may well prove to be the
case for a much wider variety of organic redox-active systems
than just that reported here if the trends observed in the
electrochemical studies of inorganic compounds with noncoor-
dinating anions are followed by their organic counterparts.
Experimental Section
All reactions were carried out under a nitrogen atmosphere,
(21) Macıas-Ruvalcaba, N. A.; Evans, D. H. J. Phys. Chem. B 2005, 109,
´
14642–14647.
using thoroughly dried solvents and glassware. 37 and 46
J. Org. Chem. Vol. 75, No. 4, 2010 1175

Adams et al.
JOCArticle
were synthesized according to the literature procedure, and 6 could
be synthesized by the palladium-catalyzed methodology reported
in the literature10 or from 5 by the procedure outlined below.
[Bu4N][B{C6H3(CF3)2}4],22 Fc[B{C6H3(CF3)2}4],23 and Fc[PF6]24
were synthesized according to the literature procedures.
(0.129 mmol) of Fc[B{C6H3(CF3)2}4]25 to give a bright green
solution. The product was precipitated as a dark green powder
by addition of excess hexane to give 120 mg (49%, 0.061 mmol).
Crystals for X-ray diffraction were grown by layered diffusion
of hexane into an ether solution: MS (ESI) [M]þ 250.22; HRMS
calcd for 2[B{C6H3(CF3)2}4 1113.2800, found 1113.2782. Anal.
Calcd for C46H34BF24N4: C, 49.61; H, 3.44; N, 5.03. Found: C,
49.70; H, 3.55; N, 5.00.
Synthesis of 4[B{C6H3(CF3)2}4]: Seventeen milligrams
(0.041mmol) of4and44 mg (0.042 mmol) ofFc[B{C6H3(CF3)2}4]
were dissolved in a mixture of 6 mL of diethyl ether and 4 mL of
CH2Cl2. A large excess of hexane was then carefully layered onto
the top of the solution, and the solutions were allowed slowly to
diffuse together to give 4[B{C6H3(CF3)2}4] (45 mg, 86%) as large
dark blue crystals: MS (ESI) [M]þ 1273.42; HRMS calcd for
4[B{C6H3(CF3)2}4] 1273.4058, found 1273.4066. Anal. Calcd for
C58H54BF24N4: C, 54.69; H, 4.27; N, 4.40. Found: C, 54.53; H,
4.36; N, 4.67.
Synthesis of 1,2,4,5-Tetrakis(dimethylamino)-3,6-difluoroben-
zene 1: A round-bottomed flask was charged with lithium
dimethylamide (26.7 mmol, 8.0 equiv relative to hexafluoroben-
zene, 40 mL of a 5% solution in hexanes). THF (20 mL) was
added to give a 0.45 M solution of the amine, and the system was
cooled to -20 °C. Hexafluorobenzene (0.62 g, 3.3 mmol) was
added dropwise to give an orange solution, which was further
stirred for 60 h. MeOH was added dropwise until decoloration
of the reaction mixture was achieved (typically 3-5 drops), and
stirring was continued for 5 min. The mixture was poured into a
20% aqueous KOH solution (20 cm3) and extracted with diethyl
ether (2 ꢀ 100 cm3) and hexane (3 ꢀ 100 cm3), washed with brine
and water, and dried over MgSO4. Evaporation of the solvent
gave a mixture of white and orange solids, which were washed
with small amounts of MeOH to leave 0.261 g of product as a
white solid. Storage of the washings at -10 °C yielded another
Synthesis of 4[I3][I5] I2: Crystals of this compound were
3
grown by adding solid iodine to a sample of 4 dissolved in CDCl3
in an NMR tube. The sample was allowed to slowly evaporate,
giving a few crystals of the product suitable for X-ray diffraction.
Synthesis of 1,2,4,5-Tetrakis(morpholin-1-yl)-3,6-difluoroben-
zene 5: A round-bottomed flask was charged with morpholine
(2.813 g, 32.3 mmol, 8.0 equiv relative to hexafluorobenzene).
THF (40 mL) was added, and the system was cooled to -20 °C.
A solution of n-BuLi (20 mL, 1.6 M in hexanes, 32.0 mmol, 1.0
equiv relative to the amine) was added dropwise, and the
mixture was stirred at -20 °C for 30 min. Hexafluorobenzene
was added dropwise to give an orange solution, which was
further stirred for 16 h while being allowed to reach room
temperature, and the solution was then refluxed for 90 min.
After cooling to room temperature, MeOH was added dropwise
until decoloration of the reaction mixture was achieved
(typically 3-5 drops) and no further color change was observed.
Addition of a further 20 mL of MeOH caused the precipitation
of the crude product as a white solid (leaving an orange
supernatant), which was isolated by filtration and purified by
being dissolved in CH2Cl2, filtered, and the solvent removed to
give 0.705 g of product. The orange supernatant was evaporated
to dryness and redissolved in CH2Cl2 to give a suspension which
was treated with ultrasound for a few minutes. This was then
filtered, and the gel-like solid was washed with 25 mL of hot
CH2Cl2. The combined organic fractions were decolorized with
charcoal and evaporated to dryness to give a further 0.073 g of
product (63%): 1H NMR (301 MHz, CDCl3) δ ppm 3.79 (t, J =
5 Hz, 16H), 3.14 (t, J = 5 Hz, 16H); 13C NMR (101 MHz,
CDCl3) δ ppm 51.50, 67.89, 133.57 (dd, J = 10.90, 7.79 Hz),
155.17 (dd, J = 246.01, 2.34 Hz); 19F NMR (283 MHz, CDCl3)
δ ppm -133.29; MS (ESI) [M þ H]þ 455.25, [M þ Na]þ 477.23.
Anal. Calcd for C22H32F2N4O4: C, 58.14; H, 7.10; F, 8.36; N,
12.33. Found: C, 58.26; H, 7.17; F, 8.19; N, 12.20.
1
0.400 g of product (total yield 0.661 g, 79%): H NMR (301
MHz, CDCl3) δ ppm 2.80 (t, J = 0.8 Hz, 24 H); 13C NMR
(76 MHz, CDCl3) δ ppm 155.9 (d, J = 244.03 Hz), 134.8 (t, J =
9.23 Hz), 43.9 (br s); 19F NMR(283 MHz, CDCl3) δ ppm -135.1;
MS (ESI) [M þ H]þ 287.20. Anal. Calcd for C14H24F2N4: C,
58.72; H, 8.45; N, 19.56. Found: C, 59.16; H, 8.69; N, 19.36.
Synthesis of 1,2,4,5-Tetrakis(dimethylamino)benzene 2: A
Schlenk tube with a Young’s tap was charged with freshly
distilled DME (40 cm3) and sodium (0.35 g, 15.28 mmol) under
nitrogen. Biphenyl (1.62 g, 10.5 mmol) was added, and the
mixture was vigorously stirred for 2 h to give a dark blue
solution. 2 (0.409 g, 1.43 mmol) was added as a solid, and
stirring was continued for 1 day. A few drops of an aqueous 20%
HCl solution was added slowly and dropwise until decoloration
of the mixture was achieved. The mixture was poured into 20%
aqueous HCl solution (20.0 cm3) and extracted with hexanes to
remove biphenyl. The aqueous phase was basified with ammo-
nia solution and extracted with diethyl ether (3 ꢀ 50 cm3),
washed with brine and then water, and dried over MgSO4.
Evaporation of the combined organic phases gave the product
as an off-white solid, which was recrystallized from dichloro-
methane/MeOH to give the product as a white powder (0.298 g,
83%): 1H NMR (300 MHz, C6D6) δ ppm 2.77 (s, 24 H), 6.68 (s, 2
H); 13C NMR (101 MHz, C6D6) δ ppm 140.6, 109.7, 42.6;
HRMS calcd for 2 250.2152, found 250.2161. Anal. Calcd for
C14H26N4: C, 67.16; H, 10.47; N, 22.38. Found: C, 66.95; H,
10.56; N, 22.13.
Synthesis of 2[PF6]2: To a solution of 16 mg of 2 (0.064 mmol)
in dry dichloromethane was added Fc[PF6] (43 mg, 0.128 mmol).
Upon stirring, the solution became green and then deposited
2[PF6]2 as a purple precipitate, which was isolated by filtration
and washed with diethyl ether to give 18.1 mg (46 mmol, 72%) of
product. Crystals for X-ray diffraction were grown by vapor
diffusion of diethyl ether into an acetonitrile solution: MS (ESI)
[M]þ 250.22; HRMS calcd for 2þ 250.2152, found 250.2150.
Anal. Calcd for C14H26F12N4P2: C, 31.12; H, 4.85; N, 10.37.
Found: C, 31.48; H, 5.13; N, 10.21.
1,2,4,5-tetrakis(morpholin-1-yl)benzene 6 was prepared from
5 in 65% yield by a procedure analogous to that described above
for the synthesis of 2 from 1: 1H NMR (400 MHz, CDCl3) δ ppm
3.14 (t, J = 4.40 Hz, 16 H), 3.81 (t, J = 4.40 Hz, 16 H), 6.53 (s, 2
H); 13C NMR (101 MHz, CDCl3) δ ppm 50.29, 67.59, 109.28,
139.62; MS (ESI) [M þ H]þ 419.27. Anal. Calcd for C22H34-
N4O4: C, 63.13; H, 8.19; N, 13.39. Found: C, 62.83; H, 8.21; N,
12.95.
Synthesis of 2[B{C6H3(CF3)2}4]: To a solution of 2 (31 mg,
0.124 mmol) in 10 mL of diethyl ether was added 135 mg
Synthesis of 6[PF6]2: To a solution of 0.051 g of 6 in 10 mL of
CH2Cl2 was added 0.065 g of Fc[PF6]. The suspension was
stirred for 2 h, before the product was isolated by filtration as
52 mg (0.07 mmol, 60%) of microcrystalline green solid which
was dried in vacuo: 1H NMR (301 MHz, CD3CN) δ ppm 3.80,
3.82 (each s, 8H), 5.95 (s, 2H); crystals for X-ray diffraction were
grown by vapor diffusion of diethyl ether into an acetonitrile
solution; MS (ESI) [M]þ 418.26, [M]2þ 209.13, [M þ PF6]þ
(22) Taube, R.; Wache, S. J. Organomet. Chem. 1992, 428, 431–442.
ꢀ
(23) Chavez, I.; Alvarez-Carena, A.; Molins, E.; Roig, A.; Maniukiewicz,
W.; Arancibia, A.; Arancibia, V.; Brand, H.; Manuel Manrıquez, J.
J. Organomet. Chem. 2000, 601, 126–132.
´
(24) Connelly, N. G.; Geiger, W. E. Chem. Rev. 1996, 96, 877–910.
(25) Chavez, I.; Alvarez-Carena, A.; Molins, E.; Roig, A.; Maniukiewicz,
W.; Arancibia, A.; Arancibia, V.; Brand, H.; Manriquez, J. M. J. Organomet.
Chem. 2000, 601, 126–132.
1176 J. Org. Chem. Vol. 75, No. 4, 2010

Adams et al.
JOCArticle
563.22, [M - C4H8NO þ H]þ 348.19. Anal. Calcd for C22H34-
F12N4O4P2: C, 37.30; H, 4.84; N, 7.91. Found: C, 37.12; H, 4.83;
N, 7.71.
EPR Spectroscopy: EPR spectra were recorded on a Bruker
EMX Micro X-band spectrometer at room temperature using
1 mm thick quartz flat cells at the EPSRC Multi-Frequency cw
EPR Service Centre at The University of Manchester. In situ
EPR spectroelectrochemical experiments at room temperature
were carried out in dichloromethane with 0.1 M [Bu4N][PF6]
(for 1) or [Bu4N][B{C6H3(CF3)2}4] (for 2) as the supporting
electrolyte, using platinum working and auxiliary electrodes and
a Ag/AgCl reference electrode. The auxiliary and reference
electrodes were both situated in the bulk solution above the
cavity, and the Pt mesh working electrode was located in the flat
part of the cell inside the cavity of the spectrometer. The
potentiostat (Autolab, Type II) was controlled via a PC running
General Purpose Electrochemical System software, version 4.9
(Eco Chemie BV, Utrecht). Simulations were performed using
Bruker Win-EPR Simfonia software.
Electrochemistry: Electrochemical studies were carried out at
room temperature (ca. 293 K) using an EG&G model 273A
potentiostat linked to a computer using EG&G Model 270
Research Electrochemistry software in conjunction with a
three-electrode cell. The auxiliary electrode was a platinum wire
and the working electrode a glassy carbon disk (1.6 mm
diameter). The reference electrode was an aqueous saturated
calomel electrode separated from the test solution by a fine
porosity frit and an agar bridge saturated with KCl. Under the
conditions used, the reversible potential for the ferrocenium/
ferrocene couple at 298 K is þ0.38 V in acetonitrile and þ0.46 V
in CH2Cl2.24 Solutions were 1.0 ꢀ 10-3 mol dm-3 in the
compound and 0.1 mol dm-3 in supporting electrolyte. Voltam-
mograms were simulated using DIGISIM version 3.03b
(Bioanalytical Systems, Inc.); further details are available in
the Supporting Information.
Crystal Structure Determinations: Diffraction intensities were
collected on a Bruker SMART APEX CCD diffractometer,
with graphite-monochromated Mo KR (0.71073 A) radiation.
The data were corrected for absorption. The structures were
solved by SHELXS-97, expanded by Fourier difference synth-
eses, and refined with the SHELXL-97 package incorporated
into the SHELXTL crystallographic package.26 The positions of
the hydrogen atoms were calculated by assuming ideal geome-
tries but not refined. All non-hydrogen atoms were refined with
anisotropic thermal parameters by full-matrix least-squares
procedures on F2. Many of the details of the structural determi-
nations are given in Table 8, and complete crystallographic
details and tables of bond lengths and angles for all compounds
are given in the Supporting Information. Many of the CF3
groups of the [B{C6H3(CF3)2}4] anions showed rotational dis-
order, which was modeled using the approach of Muller.27 The
€
structure of 4[B{C6H3(CF3)2}4] contains a single large residual
electron density peak (of 2.052 e A-3) between two of the
fluorine atoms of a CF3 group. It was not possible to model
this in a satisfactory fashion as there was no indication of the
presence of two more peaks for the remaining fluorine atoms.
Computational Details. All calculations were performed with
Gaussian 03 (revision B.04)28 and used the popular B3LYP
density functional.29-31 All atoms were described by the all-
electron 6-31G* basis set. Geometry optimizations were per-
formed in the gas phase without either solvation or the imposi-
tion of symmetry, and an unsolvated ΔfG for each species was
then calculated according to the method of Ochterski.34 Solva-
tion was then applied using a dichloromethane (ε = 8.93) or
acetonitrile (ε = 36.64) continuum solvation field, using the
(26) Sheldrick, G. M. Acta Crystallogr., Sect. A 2008, 64, 112–122.
€
(27) Muller, P. Cryst. Rev. 2009, 15, 57–83.
J. Org. Chem. Vol. 75, No. 4, 2010 1177

Adams et al.
JOCArticle
default polarizable continuum model PCM32,33 as implemented
in G03 with radii optimized for the HF/6-31G* level of theory
as recommended in the Gaussian 03 manual, which generates
a Gibbs free energy of solvation. This was then added to
the appropriate unsolvated ΔfG value to yield the final sol-
vated ΔfG; a table of these values for all species (Table S1) is
available as part of the electronic Supporting Information.
Attempts at geometry optimization with the application of
solvation failed to converge for the ion pairs (hence the
need to apply solvation after optimization), but comparison of
these unconverged results with those obtained by retrospec-
tive application of the PCM continuum revealed no signifi-
cant differences in either energy or geometry between the two
methods.
Supporting Information Available: Details of crystallo-
graphic structure determinations, computational details, elec-
trochemical simulations, and NMR spectra. This material is
available free of charge via the Internet at http://pubs.acs.org.
(28) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb,
M. A.; Cheeseman, J. R.; Montgomery J. A., Jr.; Vreven, T.; Kudin, K. N.;
Burant, J. C.; Millam, J. M.; Iyengar, S. S.; Tomasi, J.; Barone, V.; Mennucci,
B.; Cossi, M.; Scalmani, G.; Rega, N.; Petersson, G. A.; Nakatsuji, H.; Hada,
M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima,
T.; Honda, Y.; Kitao, O.; Nakai, H.; Klene, M.; Li, X.; Knox, J. E.;
Hratchian, H. P.; Cross, J. B.; Adamo, C.; Jaramillo, J.; Gomperts, R.;
Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.;
Ochterski, J. W.; Ayala, P. Y.; Morokuma, K.; Voth, G. A.; Salvador, P.;
Dannenberg, J. J.; Zakrzewski, V. G.; Dapprich, S.; Daniels, A. D.; Strain,
M. C.; Farkas, O.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.;
Foresman, J. B.; Ortiz, J. V.; Cui, Q.; Baboul, A. G.; Clifford, S.; Cioslowski,
J.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.;
Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.;
Nanayakkara, A.; Challacombe, M.; Gill, P. M. W.; Johnson, B.; Chen,
W.; Wong, M. W.; Gonzalez, C.; Pople, J. A. Gaussian 03, revision B.04;
Gaussian, Inc.: Wallingford, CT, 2003.
Acknowledgment. We wish to thank the ESPRC for fund-
ing (C.J.A., R.C.D.C., M.F.H.). D.H.E. gratefully acknow-
ledges support from the U.S. National Science Foundation,
Grant No. CHE-0715375.
(29) Lee, C. T.; Yang, W. T.; Parr, R. G. Phys. Rev. B 1988, 37, 785–789.
(30) Becke, A. D. J. Chem. Phys. 1993, 98, 5648–5652.
Note Added after ASAP Publication. Due to a calculation
error, the values in Tables 6 and 7 and in the Supporting
Information were incorrect in the version published ASAP on
January 14, 2010; the corrected version was published on the
Web February 12, 2010.
(31) Vosko, S. H.; Wilk, L.; Nusair, M. Can. J. Phys. 1980, 58, 1200–1211.
(32) Cossi, M.; Barone, V. J. Chem. Phys. 2001, 115, 4708–4717.
(33) Cossi, M.; Scalmani, G.; Rega, N.; Barone, V. J. Chem. Phys. 2002,
117, 43–54.
(34) Ochterski, J. W. Thermochemistry in Gaussian; Gaussian, Inc.:
Wallingford, CT, 2000.
1178 J. Org. Chem. Vol. 75, No. 4, 2010