Electrochemical and chemical oxidation of dithia-, diselena-, ditellura-, selenathia-, and tellurathiamesocycles and stability of the oxidized species

Article abstract of DOI:10.1021/jo9026484

"Chemical Equation Presented" The diverse electrochemical and chemical oxidations of dichalcogena-mesocycles are analyzed, broadening our understanding of the chemistry of the corresponding radical cations and dications. 1,5-Diselenocane and 1,5-dithiocan

Full text of DOI:10.1021/jo9026484

pubs.acs.org/joc
Electrochemical and Chemical Oxidation of Dithia-,
Diselena-, Ditellura-, Selenathia-, and Tellurathiamesocycles
and Stability of the Oxidized Species
Dennis H. Evans,*,† Nadine E. Gruhn,† Jin Jin,‡ Bo Li,‡ Edward Lorance,§
Noriko Okumura,† Norma A. Macıas-Ruvalcaba,† Uzma I. Zakai,†
´
Shao-Zhong Zhang,‡ Eric Block,*,‡ and Richard S. Glass*,†
†Department of Chemistry, The University of Arizona, Tucson, Arizona 85721, ‡Department of Chemistry,
University at Albany, State University of New York, Albany, New York 12222, and Department of
Chemistry, Vanguard University, Costa Mesa, California 92626
§
evansd@purdue.edu; rglass@u.arizona.edu; eb801@albany.edu
Received January 13, 2010
The diverse electrochemical and chemical oxidations of dichalcogena-mesocycles are analyzed,
broadening our understanding of the chemistry of the corresponding radical cations and dications.
1,5-Diselenocane and 1,5-ditellurocane undergo reversible two-electron oxidation with inverted
potentials analogous to 1,5-dithiocane. On the other hand, 1,5-selenathiocane and 1,5-tellurathio-
cane undergo one-electron oxidative dimerization. The X-ray crystal structures of the Se-Se dimer of
the 1,5-selenathiocane one-electron oxidized product and the monomeric two-electron oxidized
product (dication) of 1,5-tellurathiocane are reported. 1,5-Dithiocanes and 1,5-diselenocanes with
group 14 atoms as ring members undergo irreversible oxidation, unlike the reversible two-electron
oxidation of the corresponding silicon-containing 1,5-ditellurocanes. These results demonstrate the
chemical consequences of the dication stabilities Teþ-Teþ>Seþ-Seþ>Sþ-Sþ, as well as Seþ-Seþ>
Seþ-Sþ and Teþ-Teþ > Teþ-Sþ.
1. Introduction
diverse areas such as electrically conducting materials3 and
the pathogenesis of Alzheimer’s disease.4
The formation and reactions of sulfur radical cations have
attracted considerable interest,1 while far less attention has
been given to the corresponding selenium and tellurium
species.2 Particular emphasis has been placed on uncovering
the factors that affect the oxidation potential of the parent
sulfur compound giving rise to the radical cation and the
stability and fate of the radical cation or related species
obtained by such oxidation. Such studies are relevant to
A seminal result in this area is the finding that chemical
one-electron oxidation of 1,5-dithiocane (1a) produces a
radical cation whose EPR spectrum persists for at least 72 h
at room temperature.5 The remarkable persistence of this
radical cation is ascribed to a transannular S-S bond,5,6
which is especially favorable in eight-membered rings due to
the proximity of 1,5-positions and formation of two fused
five-membered rings. Transannular interactions in eight-
membered rings have long attracted interest, for example,
(1) Glass, R. S. Top. Curr. Chem. 1999, 205, 1–87.
(2) Detty, M. R.; Logan, M. E. Adv. Phys. Org. Chem. 2004, 39, 79–145.
(3) (a) Roncali, J. Chem. Rev. 1992, 92, 711–738. (b) Ishiguro, T.; Yamaji,
K. Organic Semiconductors, 2nd ed.; Springer-Verlag: Berlin, Germany, 1998.
(c) TTF Chemistry: Fundamentals and Applications of Tetrathiafulvalene;
Yamada, J., Sugimoto, T., Eds.; Kodansha-Springer: Hong Kong, 2004.
(d) Herranz, M. A.; Sanchez, L.; Martin, N. Phosphorus, Sulfur Silicon Rel.
Elem. 2005, 180, 1133–1148. (e) Barbarella, G.; Melucci, M.; Sotgiu, G. Adv.
Mater. 2005, 17, 1581–1593. (f) Roncali, J.; Blanchard, P.; Frere, P. J. Mater.
Chem. 2005, 15, 1589–1610.
(4) (a) Varadarajan, S.; Yatin, S.; Kanski, J.; Jahanashahi, F.; Butterfield,
D. A. Brain Res. Bull. 1999, 50, 133–141. (b) Butterfield, D. A.; Kanski, J.
€
Peptides 2002, 23, 1299–1309. (c) Pogocki, D.; Schoneich, C. Chem. Res.
Toxicol. 2002, 15, 408–418. (d) Schoneich, C.; Pogocki, D.; Hug, G. L.;
€
Bobrowski, K. J. Am. Chem. Soc. 2003, 125, 13700–13713.
(5) Musker, W. K.; Wolford, T. L. J. Am. Chem. Soc. 1976, 98, 3055–
3056.
(6) (a) Musker, W. K.; Wolford, T. L.; Roush, P. D. J. Am. Chem. Soc.
1978, 100, 6416–6421. (b) Musker, W. K. Acc. Chem. Res. 1980, 13, 200–206.
DOI: 10.1021/jo9026484
r
Published on Web 02/24/2010
J. Org. Chem. 2010, 75, 1997–2009 1997
2010 American Chemical Society

Evans et al.
JOCArticle
1,5-dithiocane 1a, particularly 1d and 1e whose preparation
we have recently described (compound 1f remains unknown).12
In addition, the redox properties of cyclic dichalcogenoethers
with Si or Sn substituents 2-5 were studied electrochemically
inspired by previous work.
FIGURE 1. Simplified MO depiction of S,S 2c,3e bonding.
1,5-hydride shifts in cyclooctyl cations7 and transannular
tertiary amine-carbonyl interactions.8 Electrochemical stu-
dies on 1a have revealed three remarkable features: (1) its
oxidation potential is 0.34 V (in CH3CN vs Ag/0.1 M AgNO3
in CH3CN), which is almost 1 V less anodic than typical
thioethers;9 (2) oxidation of 1a is reversible, whereas typical
thioethers undergo irreversible oxidation; and (3) the rever-
sible oxidation peak for 1a corresponds to a two-electron
oxidation, which is ascribed to potential inversion;that is,
1a undergoes two one-electron oxidations in which E°0 for
the first step is more positive than that for the second step. All
three features are attributed to neighboring group participa-
tion on oxidation, that is, transannular S-S bond formation
lowering the oxidation potential. A 2c,3e S-S bond,10
depicted in Figure 1, is formed in which there is a σ*-electron.
That is, such Si and Sn substituents substantially lower the
ionization energy of dichalcogenoethers in certain cases.13
However, chemical oxidation of these systems results in
decomposition, especially ring cleavage.12 Electrochemical
studies provide a useful probe for discerning the competition
between chalcogen-chalcogen bond formation stabilizing
these systems and ring cleavage facilitated by Si or Sn.
Before presenting these results, a brief review of the work
done by others on these and related compounds is warranted.
Chemical oxidation of 1,5-diselenocane, 1b, produced the
corresponding dication with a transannular bond between
the selenium atoms,14,15 which was characterized spectro-
scopically and by an X-ray crystallographic structure ana-
lysis of its BF4 salt.16 Comproportionation of this dication
with 1b or one-electron oxidation of 1b with NOPF6 pro-
vided the corresponding radical cation as deduced from the
appearance of an EPR signal.17 Cyclic voltammetric studies
of this dication showed a reversible reduction peak in aceto-
nitrile at an unusually low reduction potential of þ0.11 V vs a
Ag/0.01 M AgNO3 in CH3CN reference electrode.17 Ana-
logous dications with a transannular Se-Se bond were
obtained in 6- and 10-membered rings.14 Both radical cation
and dication18 were formed by chemical oxidation of 6. The
dication was also produced by treatment of the correspond-
ing selenoxide with concentrated H2SO4. Cyclic voltam-
metric studies of 6 in acetonitrile18 reveal a reversible
oxidation peak at þ0.33 V vs a Ag/0.01 M AgNO3 in CH3CN
reference electrode. Chemical oxidation of 1c provided the
corresponding dication with a transannular Te-Te bond as
Removal of this antibonding electron is favored electro-
nically because S-S bonding is strengthened despite the
electrostatic disfavoring of a dication relative to a mono-
cation. Transannular bond formation also stabilizes the
dication, which has also been chemically synthesized and
structurally characterized by X-ray analysis.11 The dication
can be reduced before it undergoes chemical decomposition,
resulting in reversible electrochemical behavior.
This paper reports the results of a systematic study of the
redox properties of the selenium and tellurium analogues of
1
deduced from H, 13C, and 125Te NMR spectroscopic stu-
dies.19 The cyclic voltammogram of 1c showed one reversible
(7) Roberts, A. A.; Anderson, C. B. Tetrahedron Lett. 1969, 3883–3885.
(8) (a) Leonard, N. J. Rec. Chem. Prog. 1956, 17, 243–257. (b) Leonard,
N. J. Acc. Chem. Res. 1979, 12, 423–429. (c) Rademacher, P. Chem. Soc. Rev.
1995, 24, 143–150.
(9) (a) Wilson, G. S.; Swanson, D. D.; Klug, J. T.; Glass, R. S.; Ryan,
M. D.; Musker, W. K. J. Am. Chem. Soc. 1979, 101, 1040–1042. (b) Ryan,
M. D.; Swanson, D. D.; Glass, R. S.; Wilson, G. S. J. Phys. Chem. 1981, 85,
1069–1075.
(10) (a) Asmus, K.-D. Acc. Chem. Res. 1979, 12, 436–442. (b) Stowasser,
R.; Glass, R. S.; Hoffmann, R. J. Chem. Soc., Perkin Trans. 2 1999, 1559–
1561. (c) Nakayama, N.; Takahashi, O.; Kikuchi, O.; Furukawa, N. J. Mol.
Struct. (Theochem) 2001, 542, 215–226. (d) Maity, D. K. J. Am. Chem. Soc.
2002, 124, 8321–8328.
(11) (a) Iwasaki, F; Toyoda, N.; Akaishi, R.; Fujihara, H.; Furukawa, N.
Bull. Chem. Soc. Jpn. 1988, 61, 2563–2567. (b) Fujihara, H.; Kawada, A.;
Furukawa, N. J. Org. Chem. 1987, 52, 4254–4257.
(13) (a) Glass, R. S.; Block, E.; Lorance, E.; Zakai, U. I.; Gruhn, N. E.;
Jin, J.; Zhang, S.-Z. J. Am. Chem. Soc. 2006, 128, 12685–12692. (b) Glass,
R. S.; Block, E.; Gruhn, N. E.; Jin, J.; Lorance, E.; Zakai, U. I.; Zhang, S.-Z.
J. Org. Chem. 2007, 72, 8290–8297.
(14) (a) Fujihara, H.; Akaishi, R.; Erata, T.; Furukawa, N. Chem.
Commun. 1989, 1789–1790. (b) Fujihara, H.; Akaishi, R.; Furukawa, N.
Tetrahedron 1993, 49, 1605–1618.
(15) Batchelor, R. J.; Einstein, F. W. B.; Gay, I. D.; Gu, J.-H.; Mehta, S.;
Pinto, B. M.; Zhou, X.-M. Inorg. Chem. 2000, 39, 2558–2571.
(16) (a) Iwasaki, F.; Morimoto, M.; Yasui, M.; Akaishi, R.; Fujihara, H.;
Furukawa, N. Acta Crystallogr. 1991, C47, 1463–1466. (b) Fujihara, H.;
Furukawa, N. Phosphorus, Sulfur Silicon Rel. Elem. 1992, 67, 131–134.
(17) Fujihara, H.; Akaishi, R.; Nakamura, A.; Furukawa, N. Tetrahe-
dron Lett. 1990, 31, 6375–6378.
(12) (a) Block, E.; Dikarev, E. V.; Glass, R. S.; Jin, J.; Li, B.; Li, X.;
Zhang, S.-Z. J. Am. Chem. Soc. 2006, 128, 14949–14961. (b) Block, E.; Glass,
R. S.; Dikarev, E. V.; Gruhn, N. E.; Jin, J.; Li, B.; Lorance, E.; Zakai, U. I.;
Zhang, S. Z. Heteroat. Chem. 2007, 18, 509–515. (c) Liao, C.; Zhang, S.-Z.;
Block, E.; Clennan, E. L. J. Org. Chem. 2008, 73, 8587–8590.
(18) Fujihara, H.; Yabe, M.; Chiu, J.-J.; Furukawa, N. Tetrahedron Lett.
1991, 32, 4345–4348.
(19) Fujihara, H.; Ninoi, T.; Akaishi, R.; Erata, T.; Furukawa, N.
Tetrahedron Lett. 1991, 32, 4537–4540.
1998 J. Org. Chem. Vol. 75, No. 6, 2010

Evans et al.
JOCArticle
oxidation peak at -0.02 V vs Ag/0.1 M AgNO3. Related
dihalogenoditelluranes were also reportedly formed by reac-
tion of 1c with iodine or chlorine.20
Transannular interaction between S and Se, on the one
hand, and S and Te, on the other, has been reported as well.
Cyclic voltammetric studies on 7a in acetonitrile showed a
peak at a potential of þ0.75 V vs a Ag/0.01 M AgNO3 in
CH3CN reference electrode.21 Furthermore, chemical oxida-
tion of 7a, as well as treatment of the selenoxide of 7a with
sulfuric acid or triflic anhydride, gave the corresponding
dication with a transannular Se-S bond as deduced from 1H
and 13C NMR spectroscopic studies.21 Cyclic voltammetric
studies on 7b in acetonitrile showed a peak at þ0.41 V vs a
Ag/0.01 M AgNO3 in CH3CN reference electrode.22 Chemi-
cal oxidation of 7b produced the corresponding Te-Te
bonded dication as deduced from 1H, 13C, and 125Te NMR
spectroscopic studies. Ab initio molecular orbital calcula-
tions on the dications derived from 1a-f with a transannular
chalcogen-chalcogen bond were reported.23
2. Results and Discussion
2.1. Compounds 1a-c. As expected based on previous
work,17,19 1b and 1c show reversible electrochemical oxida-
tion in acetonitrile. To gain more insight into these redox
processes, cyclic voltammograms for the compounds were mea-
sured under a range of scan rates. In addition, the electro-
chemical oxidation of 1a was reinvestigated.
FIGURE 2. Cyclic voltammograms for 1b: (a) 1.6 mM, 0.20 V/s;
(b) 1.6 mM, 2.00 V/s; (c) 7.5 mM, 1.00 V/s, in 0.10 M NaClO4/
acetonitrile with use of a glassy carbon working electrode. The solid
lines are the experimental results and the circles represent the
theoretical simulation with the parameters listed in Table 1.
seen to overlap, forming a single two-electron oxidation
peak. Similar arguments explain the single reduction peak
that is seen. The shape of the voltammetric peaks for 1a and
1c was similar to those of 1b.
Figure 2 shows typical cyclic voltammograms for 1b at
three scan rates and two different concentrations. The initial
positive-going sweep reveals a single peak whose height
corresponds to an overall two-electron oxidation of 1b. On
the return sweep a single overall two-electron reduction peak
is seen. In congruence with the earlier analysis of the
voltammograms of 1a, the oxidation is interpreted in terms
of stepwise removal of two electrons, forming first the cation
radical and then the dication. The individual formal poten-
tials for these two processes, E°1 and E°2, respectively, are
quite similar causing the two separate peaks that are normally
In earlier work,9b voltammograms of 1a were interpreted
in terms of an EE (electron transfer/electron transfer) pro-
cess in which the individual formal potentials differed by
only 20 mV and for which E°1 - E°2 was positive (þ20 mV),
which corresponds to “inverted potentials”.24 This corre-
sponds to the situation wherein the removal of the second
electron from 1a occurs with greater ease than removal of the
first, although the difference was small in this case.
(20) Fujihara, H.; Takuguchi, Y.; Ninoi, T.; Erata, T.; Furukawa, N.
J. Chem. Soc., Perkin Trans. 1 1992, 2583–2584.
(21) Fujihara, H.; Mima, H.; Chiu, J.-J.; Furukawa, N. Tetrahedron Lett.
1990, 31, 2307–2310.
This potential inversion has previously been suggested to
be a manifestation of the 2c,3e-bonding in [1a]•þ. In this
bonding scheme, shown in Figure 1, there are two electrons
(22) Takaguchi, Y.; Fujihara, H.; Furukawa, N. Organometallics 1996,
15, 1913–1919.
(23) (a) Fujihara, H.; Furukawa, N. J. Mol. Struct. (Theochem) 1989,
186, 261–272. (b) Pissarev, S. A.; Shevchenko, N. E.; Nenaidenko, V. G.;
Balenkova, E. S. Russ. Chem. Bull. Int. Ed. 2003, 52, 1667–1673.
(24) For a recent example of potential inversion with numerous refer-
as-Ruvalcaba, N. A.; Evans, D. H. J. Phys. Chem. B 2006,
110, 5155–5160.
ences, see: Macı
´
J. Org. Chem. Vol. 75, No. 6, 2010 1999

Evans et al.
JOCArticle
TABLE 1. Parameter Values Found by Fitting Simulations to Voltammograms of Compounds 1a-c and 2g,h According to an EE Mechanism with
Potential-Dependent Values of the Electron-Transfer Coefficienta
compd
E°1, V
ks,1, cm/s
λ1, eV
E°2, V
ks,2, cm/s
λ2, eV
E°overall, V
E°1 - E°2, V
b
1a
1b
1c
2g
2h
0.404
0.102
-0.294
-0.161
-0.214
0.025
0.019
0.0062
0.042
0.030
-
0.251
0.058
-0.330
-0.372
-0.237
0.00051
0.012
0.00085
0.031
0.0062
0.30
0.45
0.17
0.057
0.45
0.327
0.080
-0.312
-0.266
-0.226
0.153
0.044
0.036
0.211
0.023
b
-
b
-
0.28
0.41
aPotentials are referred to the ferrocene/ferrocenium couple measured in acetonitrile. Concentrations were in the range of 0.5 to 4 mM in 0.10 M
NaClO4 in acetonitrile. For each compound voltammograms from 0.10 to 10 V/s were obtained and analyzed. Compound 1b was studied at three
concentrations (1.6, 4.0, and 7.5 mM). Diffusion coefficients were adjusted to match the anodic peak height and ranged from about 5 ꢀ 10-6 cm2/s
for 2h to 1.5 ꢀ 10-5 cm2/s for 1a. Slight improvements in simulation of the voltammograms of 1a and 1b were obtained by including disproportionation
(2D•þ H D þ D2þ) with disproportionation rate constants of 5.0 ꢀ 106 and 1.9 ꢀ 103 M-1s -1, respectively. Inclusion of disproportionation had no effect
for 1c, 2g,h. For 2g and 2h, it was found that the simulated cathodic peak was larger than the experimental. This could be accounted for by inclusion of a
slow decomposition reaction of the dication as was done previously for 1a. However, addition of this reaction did not affect the electron-transfer
parameters. bBest-fit values of λ1 are large and correspond to the limit where R1 changes negligibly over the potential range encompassed in the
voltammogram. Specifically, ΔR1 e 0.05. That is, for these compounds the electron-transfer kinetics of the first electron transfer follow the
Butler-Volmer equations, eqs 1 and 2.
in the S-S σ molecular orbital and one in the antibonding σ*
molecular orbital. Consequently, removal of this antibond-
ing electron on going from [1a]•þ to [1a]2þ increases the S-S
bond order, thus assisting the removal of the second electron.
This is the qualitative basis for the potential inversion that is
seen in 1a-c.
formal potential but that it will deviate from one-half as the
potential is made either positive or negative of the formal
potential. The key parameter here is the reorganization
energy. Smaller values of λ cause a larger dependence of
R on potential.
In the present case, we were able to obtain much improved
fits of simulations with the experimental data using the
potential-dependent R according to eq 3. These fits are illus-
trated by the points in Figure 2 (examples of fits of simula-
tion with the voltammograms of 1c, 2g, and 2h are shown in
the Supporting Information). The parameters used are given
in Table 1. The extent of potential inversion, E°1 - E°2, was
found to be larger for 1a than was found previously with the
Butler-Volmer formulation. This difference in formal poten-
tials is difficult to quantify in the case of potential inversion
but we expect that the present result (þ0.15 V) is more
accurate than the earlier þ0.02 V.
In the earlier work,9b the electron-transfer reactions were
considered to follow the empirical Butler-Volmer formula-
tion of heterogeneous electron-transfer kinetics.25 Here, for
the reaction of CR þ e- H N
ꢀ
ꢁ
-R1F
RT
kf ,1 ¼ ks,1 exp
ðE - E10Þ
ð1Þ
ꢀ
ꢁ
ð1 - R1ÞF
kb,1 ¼ ks,1 exp
ðE - E10Þ
ð2Þ
RT
(where CR is the cation radical and N is the neutral species),
the forward, kf, and reverse, kb, electron-transfer rate con-
stants are given by eq 1 and 2. Here, ks,1 is the standard
heterogeneous electron-transfer rate constant, the rate con-
stant when the potential, E, is equal to the formal potential,
E°1, and R1 is the electron-transfer coefficient. Both ks,1 and
R1 are adjustable parameters used to fit the simulations to the
experimental data, although R1 is bounded by 0 and 1 and
should be close to one-half. Analogous expressions are
relevant for the second step of oxidation.
The previous simulations for 1a, based on the Butler-
Volmer formalism, provided fits that were approximately
correct but which deviated in important details. In the
present work we also found similar fits of simulation to
experiment for 1a-c when using these equations. The extent
of agreement was definitely less than desired.
The overall potential for the two-electron reaction, E°ov,
which is given by (E°1 þ E°2)/2, was found to be very similar
for simulations based on Butler-Volmer and Marcus for-
mulations although the extent of inversion differed signifi-
cantly and, as mentioned above, the Marcus formulation
provided better fits to the experimental voltammograms.
The lowest vertical ionization energies, IE, for 1a-c have
been determined by gas-phase photoelectron spectroscopy as
8.27,12b,26 7.98,12b and 7.5912b eV, respectively. Their oxida-
tion potentials, E°1, are þ0.404, þ0.102, and -0.294 V,
respectively, and these provide an excellent linear correlation
for this limited series (E°1=-8.081 þ 1.026IE; R2=0.9999;
slightly poorer correlation for E°ov vs IE). The near-unity
slope could mean that the solvation energies of the cation
radicals of 1a-c are almost constant, or that structural
changes are compensating for differences in solvation ener-
gies. As will be seen below, such linear correlations of poten-
tial and ionization energy are not general nor are they
expected to be general.27-29 The potential for the overall
A conclusion derived from the Marcus theory of electron
transfer25 is that the electron-transfer coefficient is not a
freely adjustable parameter but is given by eq 3. Here λ is the
reorganization energy
"
#
1
2
E - E0
2λ
R ¼
þ
ð3Þ
(26) Setzer, W. N.; Coleman, B. R.; Wilson, G. S.; Glass, R. S. Tetra-
hedron 1981, 37, 2743–2747.
(27) Glass, R. S.; Andruski, S. W.; Broeker, J. L.; Firouzabadi, H.;
Steffen, L. K.; Wilson, G. S. J. Am. Chem. Soc. 1989, 111, 4036–4045.
(28) Block, E.; Glass, R. S.; DeOrazio, R.; Lichtenberger, D. L.; Pollard,
J. R.; Russell, E.; Schroeder, T. B.; Thiruvazhi, M.; Toscano, P. J. Synlett
1997, 525–528.
for the electron-transfer reaction. Equation 3 predicts that
R is equal to one-half when the potential is equal to the
(29) Block, E.; Birringer, M.; DeOrazio, R.; Fabian, J.; Glass, R. S.; Guo,
C.; He, C.; Lorance, E.; Qian, Q.; Schroeder, T. B.; Shan, Z.; Thiruvazhi, M.;
Wilson, G. S.; Zhang, X. J. Am. Chem. Soc. 2000, 122, 5052–5064.
(25) Bard, A. J.; Faulkner, L. R. Electrochemical Methods. Fundamentals
and Applications, 2nd ed.; Wiley: New York, 2001; Sections 3.3, 3.4, and 3.6.
2000 J. Org. Chem. Vol. 75, No. 6, 2010

Evans et al.
JOCArticle
two-electron oxidation to the dications, E°ov, represents the
difference in energy between the dication and the corres-
ponding neutral compound. Consequently, the order of
stability of dication relative to the neutral compounds is
[1c]2þ > [1b]2þ > [1a]2þ. Enhanced stability of the dication
relative to the neutral can be related both to relative atom
ionization energies and the strength of the E-E bond in the
dication. The order correlates very well with the relative
atom ionization energies Te < Se < S,30 but is reversed
compared with chalcogen-chalcogen bond energies. It is
well-known31 that the bond dissociation energy for chalco-
gen-chalcogen bonds follows the order S-S > Se-Se >
Te-Te, i.e., Te-Te bonds are weaker than S-S or Se-Se
bonds. Nevertheless, the values of E°ov can reasonably be
interpreted to mean that formation of a Te-Te bond is more
favorable in [1c]2þ than a Se-Se bond in [1b]2þ or an S-S
bond in [1a]2þ. Indeed, ab initio calculations on 1,5-dichal-
cogenocanes32,33 show that the order of stability of the
FIGURE 3. He I Photoelectron spectra of 1d (lower) and 1e (upper).
dications with respect to the neutral species is [1c]2þ
>
[1b]2þ > [1a]2þ. In addition, the transannular bond order34
between chalcogens increases on going from [S8]2þ to [Se8]2þ
to [Te8]2þ. Thus the reversal in bond strength from the usual
S-S>Se-Se>Te-Te to Teþ-Teþ >Seþ-Seþ >Sþ-Sþ
in dications is suggested to play a key role in explaining our
results.
2.2. Compounds 1d and 1e. On the basis of the analysis
presented above for 1a-c, determination of the ionization
energies for 1d and 1e might provide insight into their redox
chemistry. Consequently the photoelectron spectra were
measured for 1d and 1e, as shown in Figure 3. The lowest
ionization energies for 1d and 1e are 8.11 and 7.73 eV,
respectively. On the basis of the correlation of lowest vertical
ionization energy and oxidation potentials reported above
for 1a-c, the oxidation potentials predicted for 1d and 1e
are þ0.240 and -0.054 V, respectively.
Interestingly, as seen in Figure 3, there is evidence for more
than one conformer for 1d and 1e. To investigate this further,
conformational analysis of 1d and 1e was carried out. On the
basis of previous work35 nine candidate conformations for
the eight-membered ring were geometry optimized to obtain
energy minima. Six conformations that are true energy
minima and are within 4 kcal/mol of the lowest energy
conformation were found for 1d and seven for 1e. Molecular
orbital calculations were then done for each of these con-
formations and the two filled MOs of highest energy were
plotted to aid in the analysis of their composition. For the
FIGURE 4. Three lowest energy conformations and plots of the
HOMO and HOMO-1 for 1d: (a) conformer B0, (b) conformer B,
and (c) conformer A.
three conformations of 1d of lowest energy, plots of their two
highest occupied MOs are shown in Figure 4. Plots of
HOMO and HOMO-1 for the other conformations and the
calculated relative energies of all of the conformers are given
in the Supporting Information. For the four conformations
of lowest energy of 1e, plots of their two highest energy
occupied MOs are shown in Figure 5. Plots of HOMO and
HOMO-1 for all of the other conformations and the calcu-
lated relative energies of all of the conformers are given in the
Supporting Information.
(30) Chang, F. C.; Young, V. Y.; Prather, J. W.; Cheng, K. L. J. Electron
Spectrosc. Relat. Phenom. 1986, 40, 363–383.
(31) (a) Pauling, L. The Nature of the Chemical Bond, 3rd ed.; Cornell
University Press: Ithaca, NY, 1960, p 85. (b) www.chem.ox.ac.uk/icl/nometal-
sold/lecture1/page11.html. (c) McDonough, J. E.; Weir, J. J.; Carlson, M. J.; Hoff,
C. D.; Kryatova, O. P.; Rybak-Akimova, E. V.; Clough, C. R.; Cummins, C. C.
Inorg. Chem. 2005, 44, 3127–3136.
The results for HOMO and HOMO-1 for 1e differ sig-
nificantly from those of 1a. The HOMO and HOMO-1 for 1a
are the symmetric and antisymmetric combinations of the
p-type lone-pair orbitals on each of the transannular sulfur
atoms. Similarly, HOMO and HOMO-1 for 1b and 1c are
also the symmetric and antisymmetric p-type lone-pair orbi-
tals on the selenium and tellurium atoms, respectively.
However, for 1e the HOMO is predominantly the 5p lone-
pair orbital on Te and HOMO-1, the 3p lone-pair orbital
on S. That is, there is no significant mixing of these lone-pair
(32) Nakayama, N.; Takahashi, O.; Kikuchi, O.; Furukawa, N. Heteroat.
Chem. 2000, 11, 31–41.
(33) Selected calculations at the B3PW91/CEP- 121G level correlate with
the results at the RHF/3-21 G* level reported in ref 32. See Table S3 in the
Supporting Information.
(34) Cameron, T. S.; Deeth, R. J.; Dionne, I.; Du, H.; Jenkins, H. D. B.;
Krossing, I.; Passmore, J.; Roobottom, H. K. Inorg. Chem. 2000, 39, 5614–
5631.
(35) (a) Clennan, E. L.; Hightower, S. E.; Greer, A. J. Am. Chem. Soc.
2005, 127, 11819–11826. (b) Laali, K.; Chen, H. Y.; Gerzins, R. J. J. Org.
Chem. 1987, 52, 4126–4128. (c) Munson, E. J.; Kheir, A. A.; Haw, J. F.
J. Phys. Chem. 1993, 97, 7321–7327. (d) Olah, G. A.; Svoboda, J. J.; Ku, A. T.
J. Org. Chem. 1973, 38, 4447–4450.
J. Org. Chem. Vol. 75, No. 6, 2010 2001

Evans et al.
JOCArticle
FIGURE 6. Cyclic voltammograms of 1d and 1e in acetonitrile with
two different switching potentials (past first anodic peak: circles;
past second anodic peak: full curves). The ordinate is the current
divided by the concentration of 1d or 1e. Glassy carbon working
electrode. 1d: 5.6 mM; 0.10 M NBu4PF6. 1e: 2.2 mM, 0.10 M
NaClO4.
FIGURE 5. Four lowest energy conformations and plots of the
HOMO and HOMO-1 for 1e: (a) conformer B0, (b) conformer A,
(c) conformer B, and (d) conformer G.
orbitals. For 1d, significant mixing of the 3p sulfur and 4p
selenium lone-pair orbitals occurs in two of the lowest energy
conformers (B0 and B) but mixing is minimal in conforma-
tion A. These results are due to orbital energy differences and
geometry effects on orbital overlap. This analysis supports
the notion that the photoelectron spectra of 1d and 1e consist
of ionizations from more than one conformation.
FIGURE 7. Comparison of background-corrected voltammogram
of 5.5 mM 1d (full curve) and simulation according to simulation
parameter values listed in Table 2 (circles). Acetonitrile with 0.10 M
NBu4PF6. Glassy carbon working electrode. Here 140 Ω of resis-
tance was compensated in the experiment. The remaining 70 Ω was
included in the simulation.
Furthermore, the corresponding dications from 1d and 1e
were expected to be stable in view of reported ab initio
calculations which showed that the order of stability of the
dications with respect to the neutral species is [1c]2þ>[1b]2þ
[1e]2þ >[1d]2þ >[1a]2þ 32
.
Chemical synthesis and characte-
SCHEME 1a
rization of salts of [1d]2þ and [1e]2þ is described below. These
results led to the expectation that the electrochemical oxida-
tion of 1d and 1e would be similar to that of 1a-c in which
the corresponding dications are reversibly formed at poten-
tials reflective of their lowest ionization energies. However,
the cyclic voltammetry results for 1d and 1e, run under the
conditions outlined above for 1a-c, diverge dramatically
from those observed with the latter compounds. Figure 6
shows voltammograms of 1d and 1e obtained with differing
scan ranges. Rather than reversible oxidation to the corre-
sponding dication, there are two steps of oxidation, both
irreversible. As will be seen, the first corresponds to initial
one-electron oxidation to form the cation radical, which in
turn undergoes rapid dimerization to form a dimer dication.
The second oxidation peak is due to the irreversible oxida-
tion of this dimer dication. There is a single reduction peak,
which is assigned to the reduction of the dimer dication to
form the original neutral compound. The short-scan experi-
ments include only the first oxidation and the reduction
aD = 1d or 1e, D•þ is the corresponding cation radical, D2 is the
2þ
•þ
dimer dication, and D2 is the dimer cation radical. Each electrode
reaction has a standard potential, E°, standard electron-transfer rate
constant, ks, and electron-transfer coefficient, R. Each chemical reaction
has an equilibrium constant, K, forward rate constant, kf, and reverse
rate constant, kb. Subscripts further identify the reaction. For example,
Kdim is the equilibrium constant for dimerization and kf,dim is the
dimerization rate constant.
process and these have proven to be amenable to quantitative
evaluation. The sulfur-tellurium compound is the more
easily oxidized in the first step but this trend is reversed for
the second oxidation peak.
2002 J. Org. Chem. Vol. 75, No. 6, 2010

Evans et al.
JOCArticle
TABLE 2. Values of Simulation Parameters for 1d and 1ea
1-Selena-5-thiacyclooctane, 1db
E°/V
0.274
electrode reactions
ks / cm s-1
A
D
•þ þ e- H D
0.03
0.0010
0.000050
0.50
0.40
0.69
D2 þ e- H D2
-0.241
0.269
2þ
•þ
D24þ þ 2e- H D2
2þ
chemical reactions
K
kf
kb
2 D•þ H D2
5.0 ꢀ 106 M-1
1.0 ꢀ 106 M-1 -1
s
0.20 s-1
9.9 M-1 s-1
n.a.
2þ
D2 H D þ D•þ
101 Md
1.0 ꢀ 103 s-1
0.030 s-1
100 s-1
•þ
D22þ f prod
irreversible
irreversible
D4þ f prod0
n.a.
1-Tellura-5-thiacyclooctane, 1ec
electrode reactions
E°/V
ks / cm s-1
A
D
•þ þ e- H D
-0.054
-0.724
0.03
0.05
0.50
0.50
[D2]2þ þ e- H D2
•þ
chemical reactions
K
kf
kb
2 D•þ H D2
1.0 ꢀ 108 M-1
2.1 ꢀ 103 Md
irreversible
1.0 ꢀ 107 M-1 -1
0.010 s-1
s
0.10 s-1
24 M-1 s-1
n.a.
2þ
D2 H D þ D•þ
5.0 ꢀ 104 s-1
•þ
D22þ f prod
aData obtained in acetonitrile at a glassy carbon electrode. Potentials are referenced to ferrocene. For 1d the diffusion coefficient of 1d (and all other
monomeric species in the mechanism) was set equal to 2.0 ꢀ 10-5 cm2/s and 1.5 ꢀ 10-5 cm2/s for dimeric species. For 1e the values were 2.5 ꢀ 10-5 cm2/s
and 2.0 ꢀ 10-5 cm2/s, respectively. For compound 1d, 140 Ω of solution resistance was electronically compensated and 70 Ω was included in the
simulations. For compound 1e, the values were 140 and 20 Ω. See Scheme 1 for definitions of the symbols used for the various species and definition of
parameters. The choice of simulation parameter values for dimerization of the cation radicals is discussed in the text. The reaction D2•þ H D þ D•þ was
made to be fast and essentially irreversible in order to ensure that the overall reduction of D22þ had the stoichiometry D22þ þ 2e- f 2D. This was found
to produce good agreement between simulation and experiment for the reduction peak due to D22þ. For both compounds, a very slow irreversible
decomposition of the dication dimer was found to provide an improvement in the fit of simulation to experiment. For 1d the reaction D24þ þ 2e- H D2
,
2þ
coupled with rapid irreversible dissociation of D24þ, D2 f prod0, improves the fit on the positive-going sweep in the region following the anodic peak.
This oxidation of the dimer dication can be seen in Figure 7. For 1e the second step of oxidation is too far removed from the first to cause any appreciable
4þ
additional current in this region. b0.10 M n-Bu4NPF6; 5.6 mM. c0.10 M NaClO4; 2.2 mM. dValue dictated by E°D•þ/D, E°D 2þ/D2•þ, and Kdim
.
2
A voltammogram of sulfur-selenium compound 1d is
shown in Figure 7. The oxidation peak occurs at about
þ0.4 V and its height corresponds to a one-electron process,
which is thought to be oxidation to the cation radical
followed by rapid dimerization to form the dimer dication
(see Scheme 1). On the return sweep, the only feature is an
irreversible reduction near -0.4 V, which is assigned to
reduction of the dimer dication, proceeding in two steps to
the original neutral molecule (Scheme 1).
The points in Figure 7 are for a simulation based on
Scheme 1. At all of the scan rates employed (0.1 to 10 V/s),
the initial oxidation is completely irreversible, signifying that
the dimerization rate constant, kf,dim, is large enough so that
no peak for D•þ þ e- f D is seen at the largest scan rate and
Kdim is sufficiently large that no such peak is seen at the
smallest scan rate (which would be seen if dimerization were
between simulation and experiment for six other scan rates
(0.1, 0.3, 0.5, 2, 3, 5 V/s). Comparison of simulations and
experiments for these other scan rates are included in the
Supporting Information.
These results are presaged by scattered earlier reports.
One-electron oxidation of 1a with NOPF6 yielded a red
diamagnetic solid, which was suggested to be the dimer of
the radical cation of 1a.6 On dissolution, it formed mono-
meric 1a•þ. Electrochemical studies9 of 1a revealed reversible
dimerization of [1a]•þ in solution with Kdim ≈ 5000 and a
reduction potential for the dimer of -0.6 V. Ab initio
calculations on the structure of this dimer suggested an
intermolecular þS-Sþ bond stabilized by transannular S
interactions, that is, a 4c, 6e-bond.36 However, here we
report the much more favorable dimerization of the radical
cations of 1d and 1e, whose rates of dimerization and
equilibrium constant are 3 orders of magnitude more favor-
able than those for the radical cation of 1a. The consequences
of these differences are dramatically reflected in the cyclic
voltammograms under ordinary conditions in which 1a, 1b,
and 1c show a two-electron oxidation to the corresponding
monomeric dications as discussed in Section 2.1 (albeit at
high concentrations (2 mM), dimerization of [1a]•þ dom-
inates over further oxidation to [1a]2þ so that an overall one,
not two-electron, oxidation is observed for 1a). In addition,
these results reflect interesting differences in the behavior of
significantly reversible). The values used in Figure 3, Kdim
=
5.0 ꢀ 106 M-1 and kf,dim= 1.0 ꢀ 106 M-1 s-1, are not unique
in their ability to account for the data, but they fall within a
range of possible values corresponding to the lack of detec-
tion of a peak for reduction of the monomeric cation radical.
The reduction peak near -0.4 V is accounted for by reduc-
tion of the dimer dication to an intermediate dimer cation
radical followed by dissociation to the neutral molecule and
the monomer cation radical, which in turn is reduced
(Scheme 1). For a summary of all of the simulation para-
meters, see Table 2.
The same parameter values used in Figure 7 (listed in
Table 2) were also found to provide reasonable agreement
(36) Mantina, M.; Chamberlin, A. C.; Valero, R.; Cramer, C. J.; Truhlar,
D. G. J. Phys. Chem. A 2009, 113, 5806–5812.
J. Org. Chem. Vol. 75, No. 6, 2010 2003

Evans et al.
JOCArticle
The key issue concerning the dimer dications [D2]2þ
obtained from 1d and 1e is their structure. Therefore chemi-
cal synthesis of these species was attempted. Oxidation of 1d
with tris(pentafluorophenyl)boron afforded the crystalline
dimer [(1d)2]2þ[HOB(C6F5)3]-2, whose structure was deter-
mined by X-ray methods. The molecular structure is shown
in Figure 9. This is the first reported X-ray structure for a
dimeric dication derived from 1,5-dichalcogenacyclo-
octanes. Each eight-membered ring adopts a BC conforma-
tion with the Se-Se bond occupying the equatorial position
in each ring. The structure features an elongated Se-Se bond
length of 2.641(1) A (twice the covalent radius of Se is
2.25-2.45 A; the Se-Se bond length in [1b]2þ is 2.382(2) A)
and S Se distances of 2.830(3) A, which is shorter than the
3 3 3
sum of the van der Waals radii for S and Se (3.70 A).36 In
addition, the S Seþ-Seþ S arrangement is almost linear.
3 3 3
3 3 3
FIGURE 8. Comparison of background-corrected voltammogram
of 2.2 mM 1e (full curve) and simulation according to simulation
parameter values listed in Table 2 (circles). Acetonitrile with 0.10 M
NaClO4. Glassy carbon working electrode. Here 140 Ω of resistance
was compensated in the experiment. The remaining 20 Ω was
included in the simulation.
Consequently, this structure strongly resembles that calcula-
ted10c for the most stable structure for [(1a)2]2þ, in which 4c, 6e-
bonding37 in S Sþ-Sþ S was suggested. To be sure that
3 3 3
3 3 3
[(1d)2]2þ did not disproportionate in solution to 1d and [1d]2þ
its solution NMR spectrum was measured. To provide
comparison NMR spectroscopic data for [1d]2þ, this dication
was synthesized by reaction of 5-thiaselenocane 5-oxide 8
(the thermodynamic product of peracid oxidation of 1d; see
Scheme 2)12c with triflic anhydride at -50 °C, which gave
[1d]2þ[CF3SO3]-2 in essentially quantitative yield.
the corresponding radical cations. Thus [1a]þ. undergoes
disproportionation to 1a and [1a]2þ but the radical cations
of 1d and 1e dimerize.
Controlled potential electrolysis of 1d indicated that other
reactions were occurring. For oxidation, the number of
electrons per molecule exceeded 1.3 after 1700 s and the
current had not approached zero at that time. However,
cyclic voltammograms exhibited the reduction feature seen
in Figure 7. Reductive electrolysis at a potential where the
dimer dication was reduced recovered most of the charge and
produced a solution whose voltammogram was almost
identical with that from the original solution before the
oxidation-reduction sequence. Thus, oxidation to the dimer
dication followed by reduction back to the neutral roughly
describes the stoichiometry of the reactions.
The results with the sulfur-tellurium compound, 1e, were
quite analogous to those seen for 1d. An example of com-
parison of simulation with an experimental background-
corrected voltammogram of 1e is presented in Figure 8.
The simulation parameters are included in Table 2 and
comparisons of simulations with experiments for five other
scan rates are included in the Supporting Information. In this
case also, the values of Kdim and kf,dim were not uniquely
determined but are consistent with a fast, essentially irrever-
sible dimerization reaction.
Controlled potential electrolytic oxidation of 1e pro-
ceeded smoothly and the current approached zero. However,
only about one-half electron per molecule was required. The
electrolyzed solution showed by cyclic voltammetry a peak
for reduction of the dimer dication but about 0.1 V more
negative than seen in voltammograms of neutral 1e. Con-
trolled potential reduction of this electrolyzed solution re-
sulted in recovery of most of the charge. The only oxidation
peak seen in the reduced solution was due to oxidation of 1e,
although its position was about 50 mV more positive than
that seen in voltammograms of neutral 1e. These results are
essentially in agreement with the voltammetry and the diffe-
rences are likely due to reactions that occur on the longer
electrolysis time scale of (>2000 s) but are not important in
the shorter cyclic voltammetry time scale (<∼20 s).
While the 1H NMR spectra of the monomeric dications of
1
1a-c have been reported to be complex, the 400 MHz H
NMR spectrum of [1d]2þ[CF3SO3]- showed five well-
2
defined multiplets centered at δ 4.30 (4H), 3.93 (2H), 3.80
(2H), 3.23 (2H), and 2.95 (2H). The 13C NMR spectrum
showed three peaks at δ 52.1, 47.5, and 32.2 while the 77Se
NMR spectrum showed a single peak at δ 936. Through use
of HMQC 2D NMR methods, the δ 52.1 13C signal could be
correlated with the δ 4.30 1H peak, the δ 47.5 13C signal with
the δ 3.93 and 3.80 peaks, and the 13C signal at δ 32.2 with the
δ 3.23 and 2.95 peaks. Additional 2D NMR methods showed
that the δ 4.30 peak was coupled to the δ 3.23 and 2.95 peaks
but not to the δ 3.93 and 3.80 peaks; similarly, the tightly
coupled multiplets at δ 3.93 and 3.80 were coupled to the
δ 3.23 and 2.95 peaks but not to the 4.30 peak.
The structural assignment for [1d]2þ[CF3SO3]-2 is consis-
tent with several observations: (1) deshielding is seen in
the 77Se NMR peak (δ 936) compared to that seen in the start-
ing material, 5-thiaselenocane 5-oxide (δ 156), and in model
compounds Et3Seþ-OTf (δ 373) and [1b]2þ[PF6]- (δ 806.5);
2
(2) increasing shielding is seen in 13C and 1H NMR spectra on
going from related sulfonium to selenonium to telluronium ions
(e.g., Me3Sþ, δ 28 (13C) and 3.9 ppm (1H); Me3Seþ δ 22 (13C)
and 2.7 ppm (1H); Me3Teþ, δ 4 (13C) and 2.3 ppm (1H));
(CH3)2(CH3CH2)Sþ, δ 38 (13C); (CH3)2(CH3CH2)Seþ, δ 35.8
(13C); (CH3)2(CH3CH2)Teþ, δ 18.2 (13C)).35b-d On the basis of
the above trends, the δ 52.1 13C peak and the δ 4.30 1H peak,
which is a triplet, are assigned to the methylene groups attached
to sulfur while the 47.5 13C peak and δ 3.93 and 3.80 1H peaks
are assigned to the methylene group attached to selenium.
The splitting pattern of the latter two sets of multiplets are
quite similar to the ddd pattern seen for the CH2S(O) protons in
(37) (a) Nakanishi, W.; Hayashi, S.; Toyota, S. J. Org. Chem. 1998, 63,
8790–8800. (b) Hayashi, S.; Nakanishi, W. J. Org. Chem. 1999, 64, 6688–
6696. (c) Nakanishi, W.; Hayashi, S.; Morinaka, S.; Sasamori, T.; Tokitoh,
N. New J. Chem. 2008, 32, 1881–1889.
2004 J. Org. Chem. Vol. 75, No. 6, 2010

Evans et al.
JOCArticle
-
FIGURE 9. Drawing of [1d]22þ[HOB(C6F5)3
]
based on X-ray crystallographic analysis. Se-Se 2.6471(6) A, S Se 2.830(3) A; the
2
3 3 3
S
Seþ-Seþ S arrangement is almost linear; see the Supporting Information for further details.
3 3 3
3 3 3
apparently due to the great sensitivity of the various oxidized
tellurium species to oxygen, treatment of 1e with NOBF4 led
to incorporation of oxygen giving the oxygen-bridged dimer
[1e-O-1e]2þ[BF4-]2 (10). Both structures were established
by X-ray methods (see the Supporting Information for X-ray
structures and experimental procedures which led to the two
compounds). Insertion of an oxygen atom between tellurium
atoms on oxidation of diaryl tellurides with NOBF4, or upon
exposure of tellurium compounds to O2, has been previously
reported.38 These results are summarized in Scheme 3.
However, crystalline [1e]2þ[CF3SO3]-2 could be obtained
by oxidation of 1e with nitrosonium triflate [NO]þ[CF3-
SO3]- at 0 °C (Scheme 3). The structure of the dication could
SCHEME 2
be determined by X-ray crystallographic analysis, as shown
1
in Figure 10. The H NMR spectrum of [1e]2þ[CF3SO3]-
2
shows six seven-peak multiplets, each with an enhanced
central peak (3.94, 3.78, 3.51, 3.29, 3.04, 2.87); 2D NMR
experiments show strong coupling of the 3.94/3.78 and 3.51/
3.29 pairs, which are not coupled to each other, and pre-
sumably correspond to the methylene protons adjacent to
sulfur and tellurium, respectively (see the Supporting Infor-
mation for spectra). The 13C NMR spectrum shows three
peaks at δ 42.6, 39.1, and 29.3. The NMR spectra are
quite similar in appearance to those described above for
5-thiaselenocane 5-oxide except that two of the central peaks
overlap giving a seven-peak rather than eight-peak pattern. The
coalescence of the diastereotopic methylene group protons
adjacent to sulfur (giving a triplet rather than two eight-peak
patterns) is similar to that seen for the methylene group protons
adjacent to sulfur in the published 1H NMR spectrum of [1a]2þ
,
which appear as a singlet.11b
2þ
The 77Se NMR spectrum of the dimeric dication [(1d)]2
[1d]2þ[CF3SO3]- except that both spectra are shifted
shows a peak at δ 867.3 ppm, which is different from that of
[1d]2þ (936 ppm). The 77Se NMR signal for [(1d)2]2þ is bet-
ween that for [1d]2þ and that for [1b]2þ (806.5 ppm) support-
ing the existence of an Seþ-Seþ moiety in the solution of
[(1d)2] 2þ. Furthermore, the electrochemical studies with
[(1d)2]2þ showed a reduction peak at -0.51 V just as obser-
ved in the cyclic voltammeric studies with 1d. Consequently
the structure of the dimer dication obtained electrochemi-
cally from 1d can confidently be assigned to that shown in
Scheme 2.
We were unsuccessful in converting 1e in solution by
chemical oxidation to a dimer similar to that prepared from
1d. Treatment of 1e with tris(pentafluorophenyl)boron af-
forded crystalline 5-thiatellurocane 5-oxytris(pentafluoro-
phenyl)boron S[(CH2)3]2Te-O-B(C6F5)3 (9). In addition,
2
slightly upfield and now the diastereotopic methylene pro-
tons on the carbon attached to sulfur appear as separate
multiplets. Despite repeated attempts, the 125Te NMR spec-
trum of [1e]2þ[CF3SO3]-2 could not be obtained.
2.3. Compounds 2a-h, 3b, 4b, 5b, and 5c. Chemical oxida-
tion of mixed sulfur-, selenium-, or tellurium-, and silicon- or
tin-containing heterocycles has been reported elsewhere.12
Electrochemical oxidation of these compounds, reported
here, complements the chemical studies in several ways.
Oxidation potentials provide quantitative information
about the relative ease of oxidation and, if the oxidation is
(38) Kobayashi, K.; Deguchi, N.; Horn, E.; Furukawa, N. Angew. Chem.,
Int. Ed. 1998, 37, 984–986.
J. Org. Chem. Vol. 75, No. 6, 2010 2005

Evans et al.
JOCArticle
SCHEME 3
FIGURE 10. Drawing of [1e]2þ(CF3SO3-)2 based on X-ray crystal-
lographic analysis.
selective cleavage of the C-Si or C-Sn bond.1,45 Such
selective redox cleavages have been used to synthetic advan-
tage. Electrochemical oxidation of aryl trimethylsilylmethyl
selenides does not result in C-Si bond cleavage but rather an
alternative cleavage pathway. The initially formed Se radical
cation is suggested40 to dimerize to an Se-Se bonded dica-
tion. This dication undergoes Se-C cleavage to afford
diselenide and Me3SiCH2þ, as shown in eq 4. In addition,
we have reported elsewhere13 that 2a, 2e, and 2g on chemical
oxidation undergo nucleophile-induced ring contraction.
electrochemically reversible, then E° reflects the free energy
differences between oxidized products and starting material.
Reactive species produced by oxidation may be reduced
before they decompose, especially when using cyclic voltam-
metry. Such studies may also provide mechanistic informa-
tion as illustrated by cyclic voltammetric analysis as reported
above.
It is known that the ionization energies and oxidation
potentials of R-silylated thioethers39 or selenides40 are modes-
tly lowered while those of R-stannylated thioethers41 are
substantially lowered compared with those of the nonmeta-
lated species. This lowering is geometry-dependent and
greatest when the C-Si or C-Sn σ-orbitals are aligned with
the sulfur or selenium p-type lone-pair orbitals.41a,c,42 In-
deed, we reported elsewhere12 that the ionization energies of
2a-h, 3b, 4b, 5b, and 5c are lowered by this effect. The
consequence of such lowering on the electrochemical oxida-
tion potentials for these compounds is reported here. The
electrochemical oxidation of these compounds is of further
interest because, once the chalcogen is oxidized, it may be
stabilized by transannular bond formation or undergo C-Si
or C-Sn bond cleavage. In cases in which transannular
bond formation is not an option, chalcogen radical cations
formed either electrochemically, by photochemical electron
transfer43 or by one-electron metallic oxidants,44 undergo
Compounds 2a-e, 3b, 4b, 5b, and 5c showed an irrever-
sible oxidation and compound 2f showed two irreversible
oxidations at the peak potentials listed in Table 3. These
oxidations were irreversible even at higher cyclic scan rates or
in CH2Cl2 instead of CH3CN as solvent. These irreversible
oxidations contrast strongly with the previously reported
reversible oxidations of 1a9 and 1b17 whose oxidation peak
potentials are included in Table 3 for comparison. The
irreversible oxidations seen for the first group suggest that
either C-M bond cleavage or oxidative dimerization40 is
faster than transannular chalcogen-chalcogen bond forma-
tion on oxidation or that the oxidized species with a trans-
annular chalcogen-chalcogen bond is kinetically unstable.
The first process is analogous to the known cleavage of
R-silylated and R-stannylated thioethers on one-electron
oxidation.1,45 The second process is analogous to the known
cleavage of aryl trimethylsilyl selenides on electrochemical
oxidation (eq 4).39 It should be noted that the HOMO
ionization energies and electrochemical oxidation potentials
do not correlate, unlike the situation with 1a-c.46 For
example, consider the tellurium compounds 1c, 2g, and 2h,
in which 1c has the highest ionization energy but least
(39) (a) Cooper, B. E.; Owen, W. J. J. Organomet. Chem. 1971, 29, 33–40.
(b) Yoshida, J.; Isoe, S. Chem. Lett. 1987, 631–634. (c) Block, E.; Yencha,
A. J.; Aslam, M.; Eswarakrishnan, V.; Luo, J.; Sano, J. J. Am. Chem. Soc.
1988, 110, 4748–4753. (d) Block, E.; Aslam, M. Tetrahedron 1988, 44, 281–
324. (e) Koizumi, T.; Fuchigami, T.; Nonaka, T. Bull. Chem. Soc. Jpn. 1989,
62, 219–225. (f) Yoshida, J.; Maekawa, T.; Murata, T.; Matsunaga, S.; Isoe,
S. J. Am. Chem. Soc. 1990, 112, 1962–1970.
(40) Jouikov, V. V.; Fattakhova, D. S. J. Organomet. Chem. 2000, 613,
220–230.
(41) (a) Glass, R. S.; Radspinner, A. M.; Singh, W. P. J. Am. Chem. Soc.
1992, 114, 4921–4923. (b) Yoshida, J.; Ishichi, Y.; Nishiwaki, K.; Shiozawa,
S.; Isoe, S. Tetrahedron Lett. 1992, 33, 2599–2602. (c) Glass, R. S.; Guo, Q.;
Liu, Y. Tetrahedron 1997, 53, 12273–12286.
(42) Nishiwaki, K.; Yoshida, J. Chem. Lett. 1996, 171–172.
(43) (a) Hasegawa, E.; Brumfeld, M. A.; Mariano, P. S. J. Org. Chem.
1988, 53, 5435–5442. (b) Yoon, U.-C.; Kim, H.-J.; Mariano, P. S. Hetero-
cycles 1989, 29, 1041. (c) Yoon, U. C.; Kim, U. C.; Choi, J. J.; Kim, D. U.;
Mariano, P. S.; Cho, I.-S.; Jeon, Y. T. J. Org. Chem. 1992, 57, 1422–1428.
(d) Ikeno, T.; Harada, M.; Arai, N.; Narasaka, K. Chem. Lett. 1997, 169–170.
(44) (a) Narasaka, K.; Okauchi, T.; Arai, N. Chem. Lett. 1992, 1229–
1232. (b) Narasaka, K.; Arai, N.; Okauchi, T. Bull. Chem. Soc. Jpn. 1993, 66,
2995–3003.
(45) (a) Yoshida, J.; Nishiwaki, K. J. Chem. Soc., Dalton Trans. 1998,
2589–2596. (b) Yoshida, J.; Kataoka, K.; Horcajada, R.; Nagaki, A. Chem.
Rev. 2008, 108, 2265–2299.
(46) For example, those compounds that have oxidation peak potentials
between 0.327 and 0.39 V (seven compounds: 1a, 2a, 2d, 3a, 3b, 4a, 4b), a
difference of only 60 mV, have first ionization energies varying substantially
by 0.87 eV (7.51-8.38 eV). The λmax of their charge transfer band corre-
spondingly varies by 108 nm (520-628 nm).12b
2006 J. Org. Chem. Vol. 75, No. 6, 2010

Evans et al.
JOCArticle
TABLE 3. Ionization Energies and Oxidation Potentials for Dichalcogena-Mesocycles
aPeak potentials (0.100 V/s) for irreversible oxidation. Potentials are referred to the reversible ferrocenium ion/ferrocene potential measured in
acetonitrile. bReference 26. cReversible reactions with E°overall values taken from Table 1. dReference 13.
positive oxidation potential (albeit by only 60 mV) among
the three. In principle, these need not correlate because the
photoelectron spectroscopic measurement is a vertical, non-
adiabatic process in the gas phase. On the other hand, the
reversible electrochemical potentials reflect adiabatic ioniza-
tion in solution containing supporting electrolyte.47 Indeed
lack of such correlation has been reported before.27-29 Also,
there is an unknown and probably varying difference between
the irreversible peak potential and the standard potential for
the first ionization of these compounds that will contribute
to the lack of correlation.47
In marked contrast to the sulfur compounds 2a-d and
selenium compounds 2e and 2f, tellurium compounds 2g
and 2h showed reversible electrochemical oxidation as does
1,5-ditellurocane, 1c. Clearly in these cases, the oxidized
telluro compounds are stable and transannular Te-Te bond
formation is faster than C-Si cleavage on oxidation. Although
there may be several factors contributing to this result, we
emphasize the key role played by transannular bond formation.
The trend in bond strengths for dichalcogen dications discus-
sed above is Teþ-Teþ>Seþ-Seþ>Sþ-Sþ. Consequently,
(47) Glass, R. S.; Wilson, G. S.; Coleman, B. R.; Setzer, W. N.; Prabhu,
U. D. G. Adv. Chem. Ser. 1982, 201, 417–441.
J. Org. Chem. Vol. 75, No. 6, 2010 2007

Evans et al.
JOCArticle
SCHEME 4
this trend in bond strength is proposed to account for the
reversible electrochemical oxidation of 2g and 2h but irre-
versible electrochemical oxidation of 2a-f.
ferrocene. The highly polished glassy carbon electrode was
repolished before use and after each scan with 0.05-μm alumina
paste, rinsed with water, and sonicated for 5 min in water. It was
then rinsed with acetone and dried with a tissue. The areas of the
two electrodes that were used were determined to be 0.0814 cm2
and 0.0707 cm2 based on simulations of voltammograms of
known concentrations of ferrocene in acetonitrile at 298 K,
using the consensus value of 2.5 ꢀ 10-5 cm2/s for the diffusion
coefficient of ferrocene.48 The solution resistance, Ru, was
almost completely compensated via IR compensation through
positive feedback. The residual Ru was included in the simula-
tions of the voltammograms.
Electrochemical Simulations. Digital simulations were con-
ducted with DigiSim from Bioanalytical Systems, version 3.03,
or DigiElch, version 2.0, a software package for the Digital
simulation of common Electrochemical experiments (http://
www.elchsoft.com).49
3. Conclusions
The complex redox chemistry of 1,5-dichalcogenocanes
and related compounds is summarized as shown in Scheme 4.
The preferred pathway followed and stability of the products
is predicated on the formation of the strongest dicationic
bond, namely Teþ-Teþ>Seþ-Seþ>Sþ-Sþ, aswellasTeþ-
Teþ > Teþ-Sþ and Seþ-Seþ > Seþ-Sþ. Furthermore,
electrochemical oxidation under inert conditions and shorter
time frame than chemical oxidations provides insight into
intermediates that react with nucleophiles (or oxygen) under
chemical oxidation conditions.
Computational Methods. The various conformers of the var-
4. Experimental Section
ious heterocycles were geometry optimized with B3LYP50,51
/
CEP-121G52 and harmonic frequencies were determined. These
frequencies assured that the conformers were genuine and
provided free energies which in turn allowed the calculation of
conformer populations.
General Methods. Anhydrous acetonitrile (AN, 99.8%, <
0.001% H2O) was used as received and transferred via syringe
under nitrogen. Sodium perchlorate was dried for 24 h in a
vacuum oven at 70 °C. Tetrabutylammonium hexafluorophos-
phate (Bu4NPF6,) was recrystallized three times from absolute
ethanol and dried in a vacuum oven at 100 °C for 24 h before use.
Compounds 1d and 1e were prepared as described elsewhere.12
Tellurium compounds were manipulated in subdued light and
were chromatographed with use of neutral alumina rather than
silica gel, toward which they are unstable. Solvents were purified
by standard procedures and were freshly distilled prior to use
and deoxygenated with argon.
Orbital energies were estimated both by (1) determining the
ionization potential from the difference in radical cation elec-
tronic energy and neutral electronic energy at the geometry of
the neutral (the ΔSCF method) and the HOMO/HOMO-1
(48) Hong, S. H.; Kraiya, C.; Lehmann, M. W.; Evans, D. H. Anal. Chem.
2000, 72, 454–458.
(49) Rudolph, M. J. Electroanal. Chem. 2003, 543, 23–29.
Cyclic Voltammetry. All electrochemical experiments were
performed with either an EG&G Princeton Applied Research
Potentiostat/Galvanostat, model 283, or a PARSTAT 2273,
using a standard, jacketed three-electrode cell (10 mL). The
experiments were conducted at room temperature (ca. 23 °C) in
a nitrogen-inflated globe chamber (I2R, model X-27-27) with
the cell contents being purged with argon. The working elec-
trode was a nominally 0.3-cm diameter glassy carbon electrode,
while a platinum wire served as an auxiliary electrode and the
reference electrode was an Ag/Agþ electrode (a silver wire
immersed in 0.10 M Bu4NPF6/0.01 M silver nitrate/acetonitrile).
The reference electrode was separated from the test solution by a
porous Vycor frit. The potential of the silver reference electrode
was periodically measured vs the potential of the ferrocenium/
ferrocene couple in acetonitrile and all potentials are reported vs
(50) As implemented in G03: Frisch, M. J., et al. Gaussian 03, Revision
D.01; Gaussian, Inc., Wallingford, CT, 2004.
(51) (a) Becke, A. D. J. Chem. Phys. 1993, 98, 5648–5652. (b) Lee, C.;
Yang, W.; Parr, R. G. Phys. Rev. B 1988, 37, 785–789. (c) Miehlich, B.; Savin,
A.; Stoll, H.; Preuss, H. Chem. Phys. Lett. 1989, 157, 200–206.
(52) Stevens, W. J.; Basch, H.; Krauss, M. J. Chem. Phys. 1984, 81, 6026–
6033. Stevens, W. J.; Krauss, M.; Basch, H.; Jasien, P. G. Can. J. Chem. 1992,
70, 612–630. Cundari, T. R.; Stevens, W. J. J. Chem. Phys. 1993, 98, 5555–
5565.
(53) Ortiz, J. V. J. Chem. Phys. 1988, 89, 6348–6352. Cederbaum, L. S. J.
Phys. (Paris) 1975, B8, 290–303. von Niessen, W.; Schirmer, J.; Cederbaum,
L. S. Comput. Phys. Rep. 1984, 1, 57–125. Zakrzewski, V. G.; von Niessen, W.
J. Comput. Chem. 1993, 14, 13–18. Zakrzewski, V. G.; Ortiz, J. V. Int. J.
Quantum Chem. 1995, 53, 583–590. Ortiz, J. V. Int. J. Quantum Chem. Symp.
1988, 34, 431–436. Ortiz, J. V. Int. J. Quantum Chem. Symp. 1989, 36, 321–
332. Ortiz, J. V.; Zakrzewski, V. G.; Dolgounircheva, O. In Conceptual
Perspectives in Quantum Chemistry; Calais, J.-L., Kryachko E., Eds.; Kluwer
Academic: New York, 1997; pp 465-518.
2008 J. Org. Chem. Vol. 75, No. 6, 2010

Evans et al.
JOCArticle
splitting from the Kohn-Sham orbital eigenvalues and (2) the
Outer Valence Green’s Function (OVGF)53 method. The orbi-
tals were visualized with use of Molekel.54
The dications of the various heterocycles were also geometry
optimized with B3LYP/CEP-121G and harmonic frequencies
were determined; from this information, free energy differences
were computed. Additionally, a Natural Bond Orbital (NBO)
analysis55 was conducted to determine the chalcogen-chalcogen
bond occupancy.
Bond strength estimates for dichalcogen dications have been
previously reported32 by computing the energy differences
between the dication and neutral species, both in the chair-
chair conformation. This was done at the RHF/3-21G* level.
The same comparison was made by using the results of this study
(at the B3PW91/CEP-121G level), but the 1a neutral species was
not computed in this study, and neither the neutral nor dication
species of 1f was computed in this study. Note that although
the chalcogen atoms in the dications are formally hyperva-
lent, neither the 3-21G* nor the CEP-121G basis set includes
d-orbitals.
was added to the above solution of NOTfO dropwise with
stirring at 0 °C giving a pale yellow solution. The solution was
kept at room temperature and the solvent was slowly evaporated
whereupon crystals of [1e]2þ[CF3SO3]- formed; mp 152-154
2
°C dec; 1H NMR (CD3CN, rt) δ 3.89-3.98 (m, 2 H), 3.74-3.83
(m, 2 H), 3.46-3.55 (m, 2 H), 3.25-3.34 (m, 2 H), 3.00-3.15 (m,
2 H), 2.80-2.93 (m, 2 H); 13C NMR (CD3CN) δ 42.6, 39.1, 29.3.
The 125Te NMR spectrum could not be determined. The X-ray
structure was determined and is shown in Figure 9 (see the
Supporting Information for further details).
1,10-Bis-5-thio-1-selenocanium Bis(B-hydroxytris[pentafluoro]-
phenyl)boron [1d]22þ[HOB(C6F5)3-]2. To a solution of 5-thiase-
1
lenocane12 (1d; 12 mg, 0.06 mmol; NMR H NMR (CDCl3) δ
2.86 (t, J=5.8 Hz, 4 H), 2.77 (t, J=5.9 Hz, 4 H), 2.14 (quintet,
J = 5.8 Hz, 4 H); 13C NMR (CDCl3) δ 31.0, 30.8, 22.1; 77Se
NMR δ 164.6 (relative to Me2Se)) in CDCl3 (0.6 mL) in an
NMR tube was added solid B(C6F5)3 (31.5 mg, 0.06 mmol) at
room temperature. The solution was kept at rt for 4-5 days
whereupon a small amount of orange crystals of the title com-
pound was obtained, mp 158-159 °C dec; 1H NMR (400 MHz,
CD3CN) δ 2.95-3.06 (m, 4 H), 2.73-2.86 (m, 4 H), 2.10-2.20
(m, 4 H); 13C NMR δ 150.1, 147.7, 139.1, 136.6, 32.1, 30.6, 27.5;
77Se NMR (CD3CN) δ 867.3 ppm (relative to Me2Se). The
X-ray structure was determined and is shown in Figure 10 (see
the Supporting Information for further details).
5-Thioniaseloniabicyclo[3.3.0]octane Bis(trifluoromethanesulfo-
nate) ([1d]2þ[CF3SO3]-2). To a solution of 5-thiaselenocane
1
5-oxide12c (8; 15 mg, 0.07 mmol; NMR H NMR (CDCl3) δ
3.28-3.38 (m, 2H), 3.18-3.27 (m, 2H), 2.55-2.70 (m, 4H),
2.28-2.46 (m, 4H); 13C NMR (CDCl3) δ 53.8, 24.3, 22.0; 77Se
(relative to Me2Se) δ 156) in CD3CN (0.6 mL) was added Tf2O
(20 mg, 0.07 mmol) at -50 °C and a colorless solution formed.
The mixture was monitored by NMR at -50 °C and then at rt.
Acknowledgment. The authors gratefully acknowledge
support of this work by the donors of the Petroleum
Research Fund, administered by the American Chemical
Society (RSG, EB), and the National Science Foundation
(CHE-0201555, CHE-0455575, RSG; CHE-9906566, CHE-
0342660, CHE-0450505, CHE-0744578, EB; CHE-0347471,
CHE-0715375, DHE).
The monomeric dication [1d]2þ[CF3SO3]- was obtained in
2
1
quantitative yield (by NMR); H NMR (400 MHz, CD3CN,
rt) δ 4.30 (t, J = 6.4 Hz, 4 H), 3.93 (apparent ddd, 2 H), 3.80
(apparent ddd, 2 H), 3.23 (apparent dtt, 2 H), 2.95 (apparent dtt,
2 H); 13C NMR (CD3CN) δ 52.1, 47.5, 32.2; 77Se (relative to
Me2Se) δ 936. Crystals suitable for X-ray analysis could not be
obtained. For the 1H NMR spectrum see the Supporting Infor-
mation.
5-Thioniatelluroniabicyclo[3.3.0]octane Bis(trifluoromethane-
sulfonate ([1e]2þ[CF3SO3]-2). To an ice-cooled solution of
TfOH (97 mg, 0.64 mmol) in CH3CN (5 mL) was added sodium
nitrite (14.6 mg, 0.21 mmol). The mixture was stirred at 0 °C for
20 min, and then a solution of 5-thiatellurocane12 (1e; 52 mg,
0.21 mmol; NMR 1H NMR (400 MHz; CDCl3) δ 2.93 (t, J=6.0
Hz, 4 H), 2.74 (t, J=6.0 Hz, 4 H), 2.24 (quin, J=6.0 Hz, 4 H);
13C NMR (100 MHz) δ 32.7 (SC), 30.9 (CCC), 0.71 (TeC); 125Te
NMR δ 254 (relative to Me2Te)) in CH3CN/CH2Cl2 (6 mL 1/1)
Supporting Information Available: Comparison of simula-
tions and background-corrected voltammograms for 1c, 2g, and
2h, plots of HOMO and HOMO-1 of the other conformations of
1d and 1e, and tables of the absolute and calculated relative
energies and their atomic coordinates of all of the conformers of
1d and 1e, comparison of selected calculated chalcogen-chalcogen
bond stability with previous work, comparison of the simulations
and background-corrected voltammograms for 1d and 1e at seve-
ral scan rates, crystal data, and structure refinement for
[1e]2þ[(CF3SO3-)]2 and [1d]2[HOB(C6F5)3-]2, experimental pro-
cedures as well as crystal data and structure refinement for
5-thiatellurocane 5-oxytris(pentafluorophenyl)boron (8), and
1
(54) Molekel 5.2.0.5; Swiss National Supercomputing Centre, Manno,
Switzerland.
5,50-oxy-bis(5-thiatellurocane) bis(tetrafluoroborate), (9), H
and 13C NMR spectra for [1d]2þ[CF3SO3-]2, [1e]2þ [CF3SO3-]2,
9, and 10, as well as a 2D 1H-1H spectrum for the latter, CIF files
for the X-ray analysis for [(1d)2]2þ [HOB(C6F5)3-]2, [1e]2þ
[CF3SO3-]2, 9, and 10 (the crystal structure data has been deposi-
ted at the Cambridge Crystallographic Data Centre with deposi-
tion numbers 735750, 735753, 735751, and 735752, respectively),
and full ref 50. This material is available free of charge via the
Internet at http://pubs.acs.org.
(55) Carpenter, J. E.; Weinhold, F. J. Mol. Struct. (Theochem) 1988, 169,
41–62. Carpenter, J. E., Ph.D. Thesis, University of Wisconsin, Madison,
WI, 1987. Foster, J. P.; Weinhold, F. J. Am. Chem. Soc. 1980, 102, 7211–
7218. Reed, A. E.; Weinhold, F. J. Chem. Phys. 1983, 78, 4066–4073. Reed,
A. E.; Weinhold, F. J. Chem. Phys. 1985, 83, 1736–1740. Reed, A. E.;
Weinstock, R. B.; Weinhold, F. J. Chem. Phys. 1985, 83, 735–746. Reed, A.
E.; Curtiss, L. A.; Weinhold, F. Chem. Rev. 1988, 88, 899–926. Weinhold, F.;
Carpenter, J. E. In The Structure of Small Molecules; Naaman R., Vager, Z.,
Eds.; Plenum: New York, 1988; p 227.
J. Org. Chem. Vol. 75, No. 6, 2010 2009