NJC
Paper
C–C bond leading to 1,2-thiaphosphetane 15 or, more favour-
ably, split into phosphorane 14 and 1,3-dithiole-2-thione 1.
A second higher energy alternative for the reaction between
1,3-dithiole-2-thione 1 and 1,3-dithiole-2-ylidene 10 leads to an
unsymmetrical electron distribution in zwitterion 20 bearing a
1,3-dithiolium ring and a carbanionic centre stabilized by three
S-substituents (Scheme 4). This intermediate can undergo an
electron redistribution process affording the most stable sym-
metric isomer 16 through a relatively low-energy barrier.
In order to explore the possibility of formation of 1,3,2-
dithiophospholane intermediates as reported in previous
mechanistic proposals,23 the reaction of dithioliumthiolate 18
with the trimethyl thiophosphate side-product was studied. All
reactions seem to involve higher-energy intermediates, the
most stable of which is dispiro-1,3,2-dithiaphospholane 21,
formed by relatively high-barrier concerted addition of 18 to
the PQS unit (Scheme 4). However, the C–S bond cleavage
required to afford the final TTF product I or intermediate 22
only produces the backward reaction to 18. The abovemen-
tioned intermediate 22 is a very unstable minimum that can
indeed be formed directly from 18 through a rather high barrier
and decomposes leading to TTF and trimethyl dithioperoxo-
phosphate, (MeO)2P(S)–SOMe 23, through a very high-energy TS
(Scheme 4). The other lower barrier way in which 18 reacts with
trimethyl thiophosphate furnishes the Sꢁ ꢁ ꢁS bridged betaine 24
(dSꢁ ꢁ ꢁS = 2.904 Å; WBI = 0.105; r = 2.86 ꢂ 10ꢀ2 e a0ꢀ3) that splits
in a barrierless process into trimethylphosphite and dispiro-
1,2-dithietane 25, which is the dimerization product of the
starting 1,3-dithiole-2-thione 1 via formal head-to-head [2+2]
cycloaddition of the exocyclic CQS bonds.
The rate-determining initial steps of reaction of 1,2-dithiole-
2-thione 1 with trimethylphosphite have been also studied
with substituted derivatives bearing one model phosphanyl
–PH2 or phosphanoyl –P(O)H2 substituent (20 and 40, respec-
tively; Scheme 3). Both the direct desulfurization to the corres-
ponding carbenes (DG‡ = 38.9 and 38.8 kcal molꢀ1 for 20 and 4,
respectively) and the C-addition furnishing betaines (DG‡ = 33.5
and 34.1 kcal molꢀ1 for 20 and 4, respectively) were found to be
slightly kinetically (also thermodynamically, see the ESI†)
favoured with respect to the parent system 1 (DG‡ = 39.5 and
34.4 kcal molꢀ1, respectively).
The full lowest-energy path 1 - 11 - 12 - 13 - 10 - I
as (Scheme 4) was re-examined, together with the competing
(close in energy) direct desulfurization 1 - 10 for three
different cases, using a more inexpensive computational
method (see Computational details): the parent dithiole-2-
thione 1 and two phosphanyl-mono-substituted dithiole-2-
thiones with PMe2 1PMe2 and P(O)Me2 1POMe2 groups (as models
for the real PPh2 and P(O)Ph2-substituted derivatives, respec-
tively). As previously reported for other reactions with different
organo-phosphorus compounds,25 the similar trends and
energy values to those reported (Scheme 4) at the reference
level for the parent derivative 1 (R = H) validate the new
method, where again desulfurization of 1 via spirothiapho-
sphirane 12 is the lowest energy pathway (Scheme 6). In
the case of 1PMe2 and 1POMe2, the direct P(OMe)3-promoted
Fig. 6 Computed
(0.05 a.u.) for HOMOꢀ1 (a) and HOMO (b), as well as BCPs (small green
spheres) and bond paths (c) for 16.
(B3LYP/def2-TZVPP)
Kohn–Sham
isosurfaces
Almost the same energy barrier from 10 was found for the
reaction (in this case endergonic) with the starting material 1
furnishing a symmetric thiocarbonyl ylide 16 that readily
undergoes sequential cyclization to dispirothiirane 17 and
one-step desulfurization by the action of a second trimethyl
phosphite unit, thus yielding the final TTF (I). As expected,
the central S atom in ylide 16 holds a formal positive charge
(qN = +0.325 a.u.), locating its only p-type lone pair (LP)
perpendicular to the C–S–C plane (Fig. 6a), whereas both
adjacent C (spiro)atoms are the main contributors to the
HOMO (Fig. 6b) and formally bear half negative charge
(qN = ꢀ0.755 a.u.) (compared with natural charges in 17:
+0.086 and ꢀ0.456 a.u. at S and C, respectively). The exocyclic
C–S bond order (d = 1.759 Å; WBI = 1.030; r = 19.29 ꢂ
10ꢀ2 e a0ꢀ3) in between the typical single bond border ꢀi3n
thiirane 17 (d = 1.865 Å; WBI = 0.856; r = 15.40 ꢂ 10ꢀ2 e a0
)
and the double bond in 1,2-dithiole-2-thione 1.15 Both dithio-
lenyl rings in 16 are additionally brought together by attractive
interring Sꢁ ꢁ ꢁS contacts establishing formal S–S bonds, as
evidenced by the location of the corresponding bond critical
points (BCP) (Fig. 6c) (d = 3.333 Å; WBI = 0.056; r = 1.39 ꢂ
10ꢀ2 e a0ꢀ3) that push not only S but also even the negatively
charged C atoms (d = 2.664 Å) below the sum of van der Waals
radii (3.60 and 3.40 Å for S–S and C–C, respectively). This
results in a slightly acute C–S–C angle (98.41; compared with
the computed equilibrium bond angle of 99.71 for Me2S at the
same level) and slight pyramidalization at the ring C2 atoms
P
(
C = 350.61). An additional BCP is found between the two
distal CQC bonds (Fig. 6c).
Further evolution from 17 occurs through a very low-barrier
ring-opening reaction to the slightly less stable zwitterionic
isomer 18, whose stability arises, as in 13, from the formation
of a 1,3-ditholium aromatic ring (Scheme 5). Opposite to the
behaviour of dithiole-2-thione 1, this intermediate undergoes
desulfurization to TTF I via the normal nucleophilic P-to-S
attack featuring a planar (almost linear in this case) S and
tetrahedral P geometries in the TS.13 According to the relative
orientation of both dithiolene rings, dispirothiirane 17 also
exists as a slightly less stable (DDG‡ = 0.3 kcal molꢀ1) out–out
isomer (only the most stable in–out isomer is presented in
Scheme 5). The alternative attack of the P(III) reagent to the
dithiolium C atom of 18 proceeds through a higher barrier,
giving rise to thiabetaine 19. The latter can rotate around the
New J. Chem.
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2020