2252
R. Tkachov et al. / Tetrahedron 73 (2017) 2250e2254
verify whether the heating and the base excess is necessary for the
opening of both TPD rings. To this end, the reaction of TPD with an
excess of potassium methoxide in deuterated methanol was
monitored by 13C NMR spectroscopy at room temperature and it
indicated unambiguously the formation of anion 32ꢀ as the only
sulfur-containing reaction product (Scheme 1) besides dimethyl
carbonate as second reaction product. This was concluded from the
accompanied by the formation of a red-colored crystalline product
which precipitated from the reaction mixture. The product shows
only one signal in the 13C NMR spectrum at
d
¼ 268.8 ppm (SI,
Fig. S3) and was assigned to potassium tetrathiooxalate (K2[4]) by
comparison with the literature data15 and its structure was further
verified by ESI-MS, IR measurements (SI) and X-ray crystal struc-
ture analysis (vide infra).
disappearance of the TPD signals at
¼ 113.2 ppm, which are assigned to the C]C double bond and the
carbonyl group, respectively (SI, Fig. S1). The appearance of new
d
¼
192.0 ppm and
Only small amounts of other reaction compounds (unknown
intermediates or side products) were found highlighting a high
selectivity of the decarbonylation reaction. While reflux for 30 min
leads to only a partial conversion of starting K2[3] to product (K2[4])
(SI, Fig. S4), almost full conversion of K2[3] was observed within 3 h
under refluxing condition in methanol. Although the decarbon-
ylation of K2[3] was not described in the literature so far, the
dimethyl derivative 4,5-bis(methylthio)-1,3-dithiol-2-one can be
decarbonylated into the dimethyl tetrathiooxalate under photo-
chemical16 or pyrolytic17 conditions. This may confirm indirectly
the general feasibility of the transformation discovered in our work.
To elucidate a possible role of oxygen, the same reaction was
conducted in the presence of air. This reaction resulted into a black
colored mixture which contains a number of unknown products.
Darkening of the reaction mixture was observed even faster in a
vessel exposed to air at ambient atmosphere. Hence, the oxidation-
based route C is unlikely a useful process during the polymerization
although we do not exclude that some useful oxidations may pro-
ceed during the recovery/purification of the polymer from the re-
action mixture (such as changes of the redox state of Ni atoms). It is
noteworthy that the elimination of the carbon monoxide from
K2[3] can be regarded as a redox reaction (disproportionation) as it
resulted in CO (product of reduction) and accompanied by the
oxidation of two carbons of double C]C-bond into carboxylic acid
derivative resulting in the dianion 42ꢀ rather than in the tetraanion
24-, the product of a hypothetical complete methanolysis of TPD. As
such, this reaction explains the formation of formally oxidized
product under anaerobic conditions. Furthermore, these experi-
ments show that full ring opening of TPD may occur via route D, i.e.,
without the charge-compensating assistance of Ni(II) salts (route
B); however, in the presence of Ni(II) salts route B may also be
operative.
d
signals at
d
¼ 198.9 ppm and
d
¼ 125.4 ppm (SI, Fig. S2) coincides
with signals of analogous sodium salt of the 32ꢀ anion.14
Our in situ 13C NMR experiments reveal that dianion 32ꢀ forms
immediately after mixing of the educts at room temperature. Dia-
nion 32ꢀ is stable in methanol solution under inert environment for
at least one week when stirred at room temperature. Moreover, we
have observed that the base excess has no influence on the reaction
result. Only signals resulting from dianion 32ꢀ are observed in the
13C NMR spectra when 2, 4.6 or 12 equivalents of KOMe are used in
the reaction. As such, our data show that only one ring opens in TPD
at room temperature independent of the reaction time and stoi-
chiometry of the base. It was proposed that strong electrostatic
repulsions between negatively charged methoxide and dianion 32ꢀ
prevent the opening of the second ring of TPD. On the other hand, it
is known that a high quality poly[Kx(Ni-ett)] polymer is formed
upon prolonged heating of TPD in the presence of the base and
NiCl2 suggesting the full ring opening of TPD under these condi-
tions. Thus, it is to be questioned what exactly causes the full ring
opening of TPD under polymerization conditions.
Two hypotheses were proposed in literature to explain this
process. According to the first assumption, the added Ni(II) salts
serve as the charge compensating agent so that the elementary
polymerization cycle may include the following sequence: i) the
opening of the first ring of the bicyclic TPD via methanolysis leads
to doubly charged anion 32ꢀ which is inactive to further meth-
anolysis; ii) the reaction of anion 32ꢀ with Ni(II) cations leads to
charge neutralized intermediates; iii) the opening of anion 32ꢀ by
reaction with another equivalent of potassium methoxide. Ac-
cording to that mechanism, the resulting polymerization product
should have the formula poly[K2(Ni-ett)] where each repeating unit
of the backbone bears two negative charges (route B). However, the
experimentally observed K/Ni ratio for the product is usually in the
range of 0.6e0.3 which is much lower than the theoretically pre-
dicted K/Ni ratio of 2.
The unintended oxidation of the dianion (bis(1,3-dithiol-2-one-
4,5-dithiolate)nickel) to the species with a lower degree of charge
by air was proposed as an alternative hypothesis to explain the
observed stoichiometry and, as the process facilitating the final ring
opening via the formation of electrostatically neutral intermedi-
ate.4 However, how the rather large amount of oxygen entered into
the reaction mixture (when the mixture is not exposed to the air
deliberately) and what the influence of the prolonged heating for
the successful polymerization is were never verified experimen-
tally so far.
In order to shed more light on these issues, we monitored
transformations of K2[3] at elevated temperature. To disentangle
the roles of temperature, the oxidation process and the presence of
Ni(II) salts and to disclose the final ring opening of 32ꢀ, an addition
experiment concerning the thermolysis of K2[3] was conducted
under inert conditions in the absence of nickel salts. Since even
trace amounts of oxygen may play a critical role in this process, the
reaction was carried out in exceptionally clean environment e in
glovebox with oxygen level <0.1 ppm. Quite surprisingly, it was
found that refluxing of a K2[3] solution in the glovebox in meth-
anol/potassium methanolate leads to a complete decarbonylation
It should be emphasized that if the formation of poly[Kx(Ni-ett)]
under standard polymerization conditions proceeds exclusively
through the polymerization of tetrathiooxalate dianion 42ꢀ (route
D), it should lead to the polymer poly[Ni-tto] (tto - tetrathiooxalate)
having a completely uncharged backbone and no appreciable
amounts of potassium ions. Since this is not the case, we suggest
that both mechanisms, route B and D, are operative with varying
contributions depending on the particular reaction conditions.
However in all cases, route D dominates, because the usually
observed content of K in poly[Kx(Ni-ett)] is x ¼ 0.3e0.6, which is
closer to the theoretical content of K inherent to route D (x ¼ 0)
than to those of route B (x ¼ 2).
It is also important to emphasize that although all routes should
result in the conductive polymer, we intuitively suggest that route
D, which involves the “true monomer”, (compound K2[4]) rather
than the monomer precursor (TPD) should provide more options
for optimization of reaction conditions toward the synthesis of
polymers with improved properties (e.g. higher purity, higher
molecular weight, better crystallinity). Though our X-ray crystal
structure analysis revealed that the anion [4] adopt a non-planar
structure, ability of this ligand to make planar Ni(II) complexes
with conducting properties was shown.18 In earlier work Reynolds
et al. reported the synthesis of highly conductive oligomers from
the tetraethylammonium salt of tetrathiooxalate and Ni(OAc)2.19
Along these lines, our preliminary experiments demonstrated
that K2[4] smoothly polymerizes in the presence of Ni(II) salts
Tkachov, Roman
