Fluorinated 1,3λ4δ2,2,4-benzodithiadiazines - A synthetic, structural and theoretical study

Article abstract of DOI:10.1002/ejic.200390016

The series of title compounds has been prepared through both electrophilic (C₆H_nF_5-n-N=S=N-SiMe₃ + SCl₂) and nucleophilic (C₆HF₄_S-N=S=N-SiMe₃ + CsF) intramolecular ortho-cyclisation reactions, and the former route seems to be the more effective. High regioselectivity of the ring-closing procedures is observed in both cases. The compounds were characterised by X-ray crystallography and multinuclear (¹H, ¹³C, ¹⁵N and ¹⁹F) NMR spectroscopy. In accordance with GIAO calculations, ¹⁵N{¹H} experiments and the effects observed on complete substitution of hydrogen by fluorine, the high-field signal in the ¹⁵N NMR spectra can be assigned to N-4 and the low-field signal to N-2. In the crystal, 5,6,7-trifluoro-(5) and 5,6,8-trifluoro-1,3λ⁴δ²2,4- benzodithiadiazine (6) are planar, whereas the 6,8-difluoro derivative 3 is bent along the S1...N4 line by 8.3°. According to NICS calculations the heterocycle moiety in this compound class is antiaromatic while the carbocycle is aromatic. The fluorine substituents increase the aromaticity - and in some cases (especially when a fluorine atom is present in the 8-position) the antiaromaticity - of the corresponding rings. The ortho-fluoro-containing starting material C₆H_nF_5-n-N=S=N-SiMe₃ (n = 2: 10) cyclises to the fluorinated 2,1,3-benzothiadiazole 27 upon treatment with CsF instead of SCl₂. For starting compound 6-HC₆F₄-S-N=S=N-SiMe₃ (14) the planar (Z,E) configuration features a short intramolecular H...N contact, as evidenced by X-ray diffraction. Both the reaction pathways mentioned are also discussed. Wiley-VCH Verlag GmbH, 69451 Weinheim, Germany, 2003.

Full text of DOI:10.1002/ejic.200390016

FULL PAPER
Fluorinated 1,3λ4δ2,2,4-Benzodithiadiazines ؊ A Synthetic, Structural and
Theoretical Study
Alexander Yu. Makarov,[a] Irina Yu. Bagryanskaya,[a] Frank Blockhuys,*[b]
Christian Van Alsenoy,[b] Yuri V. Gatilov,[a] Vladimir V. Knyazev,[a]
Alexander M. Maksimov,[a] Tatiana V. Mikhalina,[a] Vyacheslav E. Platonov,[a]
Makhmut M. Shakirov,[a] and Andrey V. Zibarev*[a]
Dedicated to Professor Dr. Rüdiger Mews on the occasion of his 60th birthday
Keywords: Antiaromaticity / Fluorinated 1,3λ4δ2,2,4-benzodithiadiazines / Electrophilic cyclisation / Molecular structure /
NICS / Nucleophilic cyclisation
The series of title compounds has been prepared through
both electrophilic (C6HnF5−n−N=S=N−SiMe3 + SCl2) and nu-
cleophilic (C6HF4−S−N=S=N−SiMe3 + CsF) intramolecular
ortho-cyclisation reactions, and the former route seems to be
the more effective. High regioselectivity of the ring-closing
procedures is observed in both cases. The compounds were
characterised by X-ray crystallography and multinuclear (1H,
13C, 15N and 19F) NMR spectroscopy. In accordance with
the heterocycle moiety in this compound class is antiaromatic
while the carbocycle is aromatic. The fluorine substituents
increase the aromaticity − and in some cases (especially
when a fluorine atom is present in the 8-position) the anti-
aromaticity − of the corresponding rings. The ortho-fluoro-
containing starting material C6HnF5−n−N=S=N−SiMe3 (n = 2:
10) cyclises to the fluorinated 2,1,3-benzothiadiazole 27 upon
treatment with CsF instead of SCl2. For starting compound 6-
GIAO calculations, 15N{1H} experiments and the effects ob- HC6F4−S−N=S=N−SiMe3 (14) the planar (Z,E) configuration
served on complete substitution of hydrogen by fluorine, the
high-field signal in the 15N NMR spectra can be assigned to
N-4 and the low-field signal to N-2. In the crystal, 5,6,7-tri-
fluoro- (5) and 5,6,8-trifluoro-1,3λ4δ2,2,4-benzodithiadiazine
(6) are planar, whereas the 6,8-difluoro derivative 3 is bent
features a short intramolecular H···N contact, as evidenced
by X-ray diffraction. Both the reaction pathways mentioned
are also discussed.
( Wiley-VCH Verlag GmbH, 69451 Weinheim, Germany,
along the S1···N4 line by 8.3°. According to NICS calculations 2003)
1. Introduction
atives) nuclei attached to C-5 and C-8 Ϫ i.e., those closest
to the heterocycle Ϫ with respect to those attached to C-
6 and C-7,[1,3,4] this probably being due to paratropic ring
currents.[5] At the same time, the compounds in series 1 are
thermally stable up to ca. 100 °C. At higher temperatures
and in dilute hydrocarbon solutions, compounds 1 quantit-
atively yield persistent 1,2,3-benzodithiazolyl radicals
(Scheme 1),[10,12] which are also the products of photolysis
reactions of 1.[10,13] The key intermediates are singlet 1,2,3-
benzodithiazol-2-ylnitrenes (Scheme 1), identified under
matrix isolation conditions.[13] The other aspects of the het-
eroatom reactivity of 1 have been studied only to a limited
extent. Since both the π-excess and the antiaromaticity are
destabilising factors, one might believe that the heteroatom
reactivity of 1 should be high; consequently, many new re-
actions should be found in which this reactivity takes on
many different forms. Additionally, several new structures
should no doubt be observed among the reaction products
of 1. Currently, however, the only fact that is known is that
the compounds in series 1 are able to iminate some SII [4]
Hyperelectronic (π- and σ-excessive) 1,3λ4δ2,2,4-
benzodithiadiazines[1Ϫ4] 1 (Scheme 1) reveal a nontrivial
combination of definite features of antiaromaticity[5] with
moderate thermal stability. In particular, the compounds in
series 1 possess a cyclic 12π-electron system with a notably
[6Ϫ8]
low IE1
in planar or nearly planar molecular con-
formations observed both in the gas phase[9] and in the
crystal,[1Ϫ4] low-lying π*-excited states[1Ϫ4,10] localised pre-
dominantly on the heterocycle,[11] and enhanced magnetic
shielding of the 1H, 13C and 19F (for the monofluoro deriv-
[a]
Institute of Organic Chemistry, Russian Academy of Sciences,
630090 Novosibirsk, Russia
Fax: (internat.) ϩ 7-3832/344-752
E-mail: zibarev@nioch.nsc.ru
[b]
Department of Chemistry, University of Antwerp (UIA),
Universiteitsplein 1, 2610 Wilrijk (Antwerpen), Belgium
Fax: (internat.) ϩ 32-3/820-2310
E-mail: Frank.Blockhuys@ua.ac.be
Supporting information for this article is available on the
WWW under http://www.eurjic.org or from the author.
Eur. J. Inorg. Chem. 2003, 77Ϫ88  2003 WILEY-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim 1434Ϫ1948/03/0101Ϫ0077 $ 20.00ϩ.50/0
77

F. Blockhuys, A. V. Zibarev et al.
FULL PAPER
and PIII [14] derivatives oxidatively, another reaction most intermediates.[3,4] The tetrafluoro derivative 7 was obtained
probably mediated by the nitrenes mentioned above. Re- by the nucleophilic ring-closure of the intermediate
versible cycloaddition to norbornadiene[1] as well as reduc- [PhFϪSϪNϭSϭN]Ϫ anion generated from PhFϪSϪNϭ
tion and hydrolysis to 2-aminobenzenethiols (isolated in the SϭNϪSiMe3 and CsF.[2] Both approaches were applied in
form of their corresponding disulfides)[3] have also been de- this work.
scribed.
2.1. Electrophilic Cyclisation
The ArϪNϭSϭNϪSiMe3 starting materials 8Ϫ10
(Scheme 2) for the electrophilic ring-closure procedure were
synthesised by a general method[3,4,16,17] from the corres-
ponding ArϪNϭSϭO compounds and LiN(SiMe3)2. Earl-
ier, the less common Me3SnN(SiMe3)2 was used to trans-
form the polyfluorinated ArϪNϭSϭO into ArϪNϭSϭ
NϪSiMe3.[18,19] Further treatment of 8Ϫ10 with SCl2 af-
forded the desired compounds 3Ϫ5 (Scheme 2). From these
experiments it appears that the introduction of up to three
fluorine atoms into the carbocycle does not deactivate
Scheme 1
ArϪNϭSϭNϪSiMe3 toward the electrophilic cyclisation
needed to obtain 1. It should be noted that only 75% of the
The fluorinated compounds in series 1 are all unknown
except for the 5-, 6- and 7-fluoro derivatives[3,4] and the
5,6,7,8-tetrafluoro derivative.[2] Nevertheless, they are inter-
esting not only from a general point of view, but also in the
context of further investigation of the peculiarities of the
molecular and electronic structures of these compounds.
For instance, it was found by gas electron diffraction (GED)
that a free molecule of the parent 1,3λ4δ2,2,4-benzodithiadi-
azine 2 is bent significantly, while a free molecule of its
tetrafluoro derivative 7 is planar.[9] The bent geometry of
the former was attributed to pseudo-JahnϪTeller distortion
to minimize the antiaromaticity,[4,9] whereas the planar geo-
metry of the latter was thought to be due to a conflict be-
tween the pseudo-JahnϪTeller effect and the perfluoro (or
π-fluoro) effect, caused by the presence of a fluorine atom
in the 8-position, resolved in favour of the latter.[9] To verify
this conclusion it is necessary to perform GED on 6,8-di-
fluoro- (3) and 5,6,7-trifluoro-1,3λ4δ2,2,4-benzodithiadiaz-
ine (5), the free molecules of which would be expected to
be planar and bent, respectively, and these experiments are
currently underway.
stoichiometric amount of SCl2 was used in these prepara-
tions, so as to suppress any further reaction of 3؊5 with
SCl2, it having previously been found that the compounds
in series 1 readily react with SCl2 to yield 1,2,3-benzodithia-
zolium chlorides (Herz salts).[4]
Scheme 2
This work deals with the preparation and X-ray struc-
tural characterisation of a number of new fluorinated deriv-
atives of 1,3λ4δ2,2,4-benzodithiadiazine 2, as well as with
the quantitative estimation of the (anti)aromaticity of the
compounds in series 1 for both the fluorocarbon and the
hydrocarbon series, by application of the Nucleus-Inde-
pendent Chemical Shift (NICS) concept.[15] On the basis of
the theoretical calculations, a more detailed structural study
was also performed for all fifteen possible fluoro-substi-
tuted 1,3λ4δ2,2,4-benzodithiadiazines.
The cyclisation proved to be effectively regioselective in
the case of 9, which possesses two non-equivalent positions
at which the ring-closure may occur (cf. refs.[3,4]). Only 4
(Scheme 2) was isolated from the reaction mixture, and the
second possible isomer was not detected. It had previously
been shown[3] that the regioselectivity in these types of
cyclisation reactions is consistent with the thermodynamics
of the corresponding intermediate σ complexes as well as
with factors of kinetic control for an orbital-controlled elec-
trophile-nucleophile reaction. At the same time, the ratio of
the isomers observed in the reaction mixture is obviously
affected by the difference in their rates of reaction with
SCl2.[4] The structures of 3 and 5 were confirmed by X-ray
2. Results and Discussion
The monofluorinated compounds of series 1 had previ- crystallography (see below).
ously been synthesised by 1:1 condensation of the corres-
As reaction by-products, the corresponding sulfur diim-
ponding ArϪNϭSϭNϪSiMe3 with SCl2, followed by the ides (ArϪNϭ)2S were identified in all cases. One can ex-
electrophilic ortho-cyclisation of the [ArϪNϭSϭNϪSϪCl] plain their formation in terms of dimerisation of [ArϪNϭ
78
Eur. J. Inorg. Chem. 2003, 77Ϫ88

Fluorinated 1,3λ4δ2,2,4-Benzodithiadiazines
FULL PAPER
SϭNϪSϪCl] intermediates through [2π ϩ 2π] cycloaddi-
tion reactions, followed by decomposition of the nonsym-
metric dimers to (ArϪNϭ)2S and the well-known
[S3N2Cl]Cl.[20] Dilution of the reaction solutions reduces
the yields of the (ArϪNϭ)2S by-products.
2.2. Nucleophilic Cyclisation
The ArFϪSϪNϭSϭNϪSiMe3 starting materials 14Ϫ18
(Scheme 3) for the nucleophilic cyclisation reactions were
obtained through the known 1:1 condensation of the cor-
responding ArFSCl compounds with (Me3SiNϭ)2S.[2,3,19]
The structure of 14 was confirmed by X-ray crystallography
(Figure 1). The planar (Z,E) configuration features a short
intramolecular H···N contact or hydrogen bond (cf. hydro-
carbon analogues[21]).
˚
Figure 1. Molecular structure of 14; selected bond lengths [A] and
bond angles [°]: C1ϪS1 1.762(3), S1ϪN1 1.662(3), N1ϪS2
1.552(3), S2ϪN2 1.494(3), N2ϪSi1 1.748(3), N1ϪH 2.504;
C1ϪS1ϪN1 98.55(15), S1ϪN1ϪS2 116.19(15), N1ϪS2ϪN2
113.58(15), S2ϪN2ϪSi1 129.17(19)
that the preferred direction of the ring-closure is consistent
with essentially the same factors as described above for the
electrophilic cyclisation reactions. The structure of 6 was
confirmed by X-ray crystallography (see below).
In the cases of 16 and 17, only polyfluorinated 2,1,3-
benzothiadiazoles 19 and 20 (i.e., the smaller-sized hetero-
cycles) could be detected in the reaction mixtures, along
with the corresponding ArFSSArF and ArFNH2 derivatives
(in different ratios). In the case of 14 (a potential precursor
of 5: see Scheme 2), however, only the disulfide 25
(Scheme 3) was identified by 19F NMR in the crude reac-
tion product, and was subsequently isolated; in the case of
18 only amine 26 (Scheme 3) was obtained in high yield.
Similar thiadiazoles, disulfides and/or amines had previ-
ously been obtained from unsuccessful attempts to prepare
fluorinated naphthalene- and pyridine-fused 1,3λ4δ2,2,4-di-
thiadiazines from the corresponding ArFϪSϪNϭSϭ
NϪSiMe3 derivatives upon treatment with CsF[3,19] (cf. also
the formation of the corresponding disulfide, rather than
the target compound 2, from 2-O2NC6H4ϪSϪNϭSϭ
NϪSiMe3 on treatment with CsF[2]). The results were at-
tributed then to both low nucleophilicity and thermal
stability of the [ArFϪSϪNϭSϭN]Ϫ anions and explained
in terms of: (1) ipso-cyclisation followed by ring-opening
reactions to form [ArFϪNϭSϭNϪS]Ϫ anions, (2) elimina-
tion of sulfur and ortho-cyclisation of [ArFϪNϭSϭN]Ϫ an-
ions to form thiadiazoles, or (3) decomposition to ArFNH2.
The decomposition of [ArFϪSϪNϭSϭN]Ϫ anions before
cyclisation yields ArFSSArF.[3,19]
The same explanation is obviously valid for this work.
The difference between the successful ArFϪSϪNϭSϭ
NϪSiMe3-based syntheses of the title compounds (ArF
ϭ
Scheme 3
C6F5,[2] 5-HC6F4) and their failed counterparts (ArF ϭ 6-
HC6F4, 4-HC6F4) is in accordance with the known relative
The compounds 14Ϫ18 were treated with CsF in boiling propensities of fluorine atoms in different positions of par-
acetonitrile (cf. ref.[2]) in an attempt to close the ring and tially fluorinated aromatic compounds to nucleophilic sub-
obtain the target heterocycles; this was successful only in stitution.[22] In the case of 14 one can also consider that the
the case of 15 (Scheme 3). The cyclisation of 15 to 6 again intramolecular H···N interaction (see Figure 1) stabilises
proved to be highly regioselective, since the second possible the [ArFϪSϪNϪϪSϵN] thiazylamide form of the interme-
isomer was not observed. It had previously been shown,[19] diate anion (cf. refs.[17,23,24]) more than the [ArFϪSϪNϭ
for the closely related nucleophilic cyclisation reactions af- SϭN]Ϫ sulfur diimide one, the former obviously being un-
fording polyfluorinated naphtho[1,2-c][1,2,5]thiadiazoles, able to cyclise in either ortho or ipso fashion. Taking into
Eur. J. Inorg. Chem. 2003, 77Ϫ88
79

F. Blockhuys, A. V. Zibarev et al.
FULL PAPER
account the high yield of cyclisation of the [4-
F3CC6F4NSN]Ϫ anion to the corresponding thiadiazole un-
der the same conditions,[25] one can also explain the pre-
dominant formation of amine 24 from compound 17
(Scheme 3) by the direct decomposition of the [4-
F3CC6F4NSNS]Ϫ anion to the radical anion [4-
F3CC6F4N]·Ϫ, followed by reaction with the solvent to give
the final product 24. This explanation is also applicable to
the formation of 26 from 18 (Scheme 3).
Thus, the synthetic usefulness of the nucleophilic ap-
proach to fluorinated 1,3λ4δ2,2,4-benzodithiadiazines is
limited. On the other hand, this type of cyclisation reaction
is a convenient method for the synthesis of fluorinated
2,1,3-benzothiadiazoles.[16,19,25] For example, compound 10
(the precursor of 5 under electrophilic cyclisation condi-
tions; Scheme 2) afforded fluorinated thiadiazole 27 in good
yield upon treatment with CsF (Scheme 4).
Figure 2. Molecular structures of 3, 5 and 6; for selected bond
lengths and bond angles see Table 1
Scheme 4
2.3. X-ray Molecular Structures and Comparison with
Calculated Gas-Phase Data
practically identical, as also found for 2,[1] whereas for 3
According to X-ray crystallography[1Ϫ4] and gas electron and 6 they are slightly different, as found for 7.[2]
diffraction[9] data, as well as post-HF and DFT calcula-
Thus, in the crystal, compound 2 and its 5-F3C, 6-F, 7-
tions,[4,9,11] the compounds in series 1 are structurally nonri- Br, 7-H3CO, 5,6,7-F3 (5) and 5,6,8-F3 (6) derivatives are
gid because they possess a low-energy vibrational mode that planar, whereas the 5-H3CO, 5-Br, 6-H3C, 8-Br, 6,8-F2 (3)
allows the molecule to be deformed easily upon going from and 5,6,7,8-F4 (7) derivatives are bent along the S1···N4 line
the gas phase to the solid. For example, a molecule of 2 is by up to 10.8° (for the 5-H3CO derivative) (see refs.[1Ϫ4]
significantly bent along the S1···N4 line in the gas phase[9] and this work). Of the two crystallographically independent
and perfectly planar in the crystal.[1] In contrast, a molecule molecules of the 6-Br derivative, one is perfectly planar
of 7 is flat in the gas phase[9] and folded in the solid state.[2] whereas the other is bent by 3.1°, which directly indicates
(For an in-depth discussion of the structures of 2 and 7 in the importance of packing effects.[4]
the two phases, including an analysis of the experimental
Table 1 also compares the solid-state geometries with
gas-phase geometries, see ref.[9]; the results of GED meas- those obtained in the gas phase by calculation, for those
urements on 3 and 5 will be published elsewhere.) The 6- derivatives for which crystal structures exist. The choice of
fluoro derivative of 2 is nearly planar in the crystal.[3] This the method/basis set combination was based on previous
structural dichotomy is also observed for compounds 3, 5 calculations on 2 and 7, which showed that B3LYP/6-
and 6 in the crystal (Figure 2 and Table 1). According to 311ϩG* performs quite well for these systems, unlike MP2,
X-ray diffraction data, molecules 5 and 6 are planar, whilst which incorrectly reproduces the conformation of the hetero-
a molecule of 3 is bent along the S1···N4 line by 8.3°; this cycle.[9] (Cartesian coordinates and energies for 2 and its
is the largest deviation from planarity so far found in the fifteen fluorinated derivatives can be found in the Sup-
solid state for fluoro-containing compounds of series 1.[2,3] porting Information; symmetries and relative energies can
For compound 4 the crystals obtained were not suitable for be found in Table 3). The calculations rigorously maintain
X-ray diffraction.
the equal bond lengths in the NϭSϭN fragment for all six-
˚
The bond lengths and bond angles of 3, 5 and 6 (Table 1) teen compounds; the largest difference is 0.005 A. However,
are typical.[1Ϫ4,9] As in the parent compound 2 and the the localisation of the π-electrons in the heterocycle is ad-
tetrafluoro derivative 7, four of the π-electrons in the heter- equately reproduced. The variation in the NϭS bond
ocycle are localised to some extent in two of the NϪS lengths for the various derivatives is negligible: for N2ϪS3
bonds, giving rise to one long and two short bonds. We also this is limited to 0.005 A, and for S3ϪN4 to 0.006 A. The
note that the bond lengths in the NϭSϭN moiety of 5 are values in Table 1 illustrate this. For the SϪN single bond,
˚
˚
80
Eur. J. Inorg. Chem. 2003, 77Ϫ88

Fluorinated 1,3λ4δ2,2,4-Benzodithiadiazines
FULL PAPER
Table 1. Selected data comparing solid-state (XRD) and calculated gas-phase (B3LYP, re) geometries for 3, 5, 6 and 6-fluoro-
1,3λ4δ2,2,4-benzodithiadiazine[3]
5
3
6
6-Fluoro
Parameter[a] [b]
XRD
1.683(5)
1.539(5)
1.532(4)
1.430(6)
1.389(7)
1.786(5)
B3LYP
XRD
1.668(5)
1.551(5)
1.536(5)
1.425(6)
1.398(7)
1.779(5)
B3LYP
XRD
1.677(6)
1.537(7)
1.552(5)
1.421(6)
1.386(9)
1.784(6)
B3LYP
XRD
B3LYP
S1ϪN2
N2ϪS3
S3ϪN4
N4ϪC4a
C4aϪC8a
C8aϪS1
S1ϪN2ϪS3
N2ϪS3ϪN4
S3ϪN4ϪC4a
N4ϪC4aϪC8a
C4aϪC8aϪS1
C8aϪS1ϪN2
1.713
1.568
1.566
1.402
1.410
1.814
121.7
117.9
121.9
125.6
121.6
103.8
1.697
1.565
1.567
1.409
1.414
1.826
124.9
118.5
122.9
124.7
125.1
103.9
1.696
1.565
1.566
1.401
1.415
1.825
125.0
118.3
122.7
125.6
124.3
104.0
1.673(4)
1.713
1.568
1.568
1.409
1.410
1.813
121.1
118.0
121.9
124.5
122.0
103.6
1.534(4)
1.531(5)
1.425(6)
1.405(6)
1.797(5)
123.9(3)
123.9(3)
124.8(1)
124.3(3)
119.7(2)
121.4(4)
125.7(4)
123.9(4)
105.3(2)
118.3(2)
123.2(4)
123.6(4)
124.6(4)
105.0(2)
119.1(3)
121.2(5)
125.1(5)
125.4(4)
104.1(3)
119.0(2)
123.3(3)
123.7(5)
124.1(4)
105.3(2)
[a]
[a]
˚
Bond lengths in A and bond angles in °. For atomic numbering see Scheme 1 and Figure 2.
˚
S1ϪN2, the variation is larger, amounting to 0.021 A, and constants were used to obtain full assignments. These are
seems to depend on the presence of the fluorine atom in given in Table 2.
In the 15N{1H} NMR spectrum of 2, the signal at δ ϭ
the 8-position and thus on the conformation of the hetero-
cycle. If the atom is present and the compound has Cs sym-
metry, the S1ϪN2 bond is shorter, as is again illustrated in
Table 1. The typical CϪC double/single-bond alternation,
as observed in the calculations on 2 and 7,[9] is found for
all other compounds in the series and thus seems to be inde-
pendent of the number and position of the fluorine atoms.
A closer look at the various CϪF bond lengths displays
some interesting patterns. The CϪF distance for a particu-
lar position (i.e., C5, C6, C7 or C8) is shorter when that
fluorine atom has another fluorine atom ortho to it than
when it does not, which corresponds to the known
shortening of the CϪF bond in r0 structures observed by
microwave spectroscopy on going from fluorobenzene to
1,2-difluorobenzene.[26] The difference in the bond lengths
between these two situations depends on the position of the
fluorine atom: for F5 and F8 the difference is limited to
0.009 A on average and for F6 and F7 it is 0.012 A on
average. In any of the derivatives of 2, the C8ϪF8 distance
is always the largest, C7ϪF7 and C6ϪF6 are intermediate
and the C5ϪF5 bond length is the shortest. The four tor-
sion angles that define the conformation of the heterocycle
(see ref.[9] for further details) do not vary significantly with
the number and the position of the fluorine atoms. The
largest differences for each angle range from 1.0° for
C5ϪC4aϪC8aϪS1 through 1.9° (N4ϪC4aϪC8aϪS1) and
2.4° (S3ϪN2ϪS1ϪC8a) to 3.8° for C4aϪC8aϪS1ϪN2.
263.1 ppm is a doublet on selective decoupling with 8-H,
but a singlet on selective decoupling with 5-H. From this,
it is possible to assign the signal at δ ϭ 263.1 ppm to N-4
and the signal at δ ϭ 269.2 ppm to N-2. This is also in
agreement with an empirical analysis of the 15N NMR spec-
troscopic data for various 5-, 6-, 7- and 8-R substituted de-
rivatives of 2.[3] Comparison of 15N and 15N{1H} NMR
spectra (see footnotes to Table 2; the latter feature the lack
of the N4-H5 spin-spin coupling, cf. ref.[3]) unambiguously
identifies the high-field resonance of the 6-F derivative of 2
as well as of compounds 3 and 4 as N-4 and the low-field
resonance as N-2. For the previously studied compounds
of series 1, the δ15N values lie in a narrow range of ca.
280Ϫ250 ppm (except for 7[2]), with a weak dependence on
the nature of the carbocyclic substituent. In the case of po-
lyfluorinated derivatives 5, 6 (Table 2) and 7,[2] one nitrogen
resonance is shifted to higher field (up to δ ϭ 233 ppm for
7[2]). It is known for related heterocycles that replacement
of the four hydrogen atoms by fluorines increases the 15N
shielding in 2,1,3-benzothiadiazole but does not affect the
δ15N of 1,3λ4δ2,5,2,4-benzotrithiadiazepine.[27,28] Con-
sequently, and in agreement with the discussion above, it is
feasible that the high-field signal in the 15N NMR spectra
of 5Ϫ7 belongs to N-4 and the low-field signal to N-2. In
agreement with the experiment, the GIAO calculations sys-
tematically predict that the N-2 nucleus of the studied com-
pounds should be definitely deshielded relative to the N-4
nucleus for all but the 5-fluoro derivative, since the calcula-
tions for this last compound fail to predict the coincidence
of both signals. For the 7-fluoro derivative the experimental
assignment is difficult, but from the excellent qualitative
agreement between theory and experiment for the other de-
rivatives, one can be confident in again assigning the high-
˚
˚
2.4. Spectral Properties
1
The experimental assignment of the H and 13C NMR
spectra of compound 2, the parent compound in the 1
series, is based on double resonance, COLOC and off-res-
onance techniques. For the fluorinated derivatives, mono-
resonance and off-resonance techniques as well as analyses field signal to N-4 and low-field signal to N-2. Quantitative
of H-1H, H-19F, 13C-19F and 19F-19F spin-spin coupling prediction of the δ15N values, on the other hand, is prob-
1
1
Eur. J. Inorg. Chem. 2003, 77Ϫ88
81

F. Blockhuys, A. V. Zibarev et al.
FULL PAPER
Table 2. Calculated and experimentally determined NMR chemical shifts (ppm) for 1,3λ4δ2,2,4-benzodithiadiazine 2 and its fluoro-
substituted derivatives
2
5-Fluoro
HF
δ
HF
DFT
DFT
DFT
DFT
DFT
δ
DFT
DFT
DFT
DFT
7-H
6-H
5-H
8-H
C-4a
C-7
C-6
C-8
C-5
C-8a
N-2
N-4
6.78
6.63
5.90
5.79
7-H
6-H
5-H
8-H
C-4a
C-7
C-6
C-8
C-5
C-8a
N-2
N-4
7.17
7.10
6.67
6.44
7-H
6-H
5-H
8-H
C-4a
C-7
C-6
C-8
C-5
C-8a
N-2
N-4
6.58
6.41
5.59
5.42
7-H
6-H
8-H
5-F
C-5
C-7
C-4a
C-8
C-6
C-8a
N-2
N-4
6.70
6.38
5.53
7-H
6-H
8-H
5-F
C-5
C-7
7.18
6.88
6.12
7-H
6-H
8-H
5-F
C-5
C-4a
C-7
C-6
C-8a
C-8
N-2
N-4
6.48
6.08
5.12
36.4
31.5
40.1
138.5
143.3
141.4
152.1
133.3
126.9
119.4
119.0
116.4
268.0
268.0
149.5
132.7
130.9
118.5
118.1
117.2
390.7
338.3
148.5
130.4
129.3
115.9
115.6
114.4
350.1
297.3
133.2
130.5
124.0
123.0
115.3
269.2
263.1
130.7
129.5
124.8
123.0
113.8
396.7
353.5
129.5
125.8
119.5
119.1
114.2
354.2
313.3
C-4a
C-8
C-6
C-8a
N-2
N-4
6-Fluoro
HF
7-Fluoro
δ
δ
HF
HF
HF
7-H
8-H
5-H
6-F
C-6
C-4a
C-8
C-7
C-5
C-8a
N-2[a]
N-4[a]
6.48
5.76
5.71
7-H
5-H
8-H
6-F
C-6
C-4a
C-8
C-7
C-5
C-8a
N-2
N-4
6.87
6.47
6.45
7-H
5-H
8-H
6-F
C-6
C-4a
C-8
C-7
C-8a
C-5
N-2
N-4
6.21
5.39
5.34
6-H
5-H
8-H
7-F
C-7
C-4a
C-5
C-8a
C-6
C-8
N[b]
N[b]
6.25
5.90
5.59
6-H
5-H
8-H
7-F
6.74
6.68
6.22
6-H
5-H
8-H
7-F
C-7
C-4a
C-5
C-8a
C-6
C-8
N-2
N-4
5.92
5.53
5.20
50.5
37.6
45.8
55.5
40.3
54.0
163.9
140.2
124.9
118.4
111.4
110.1
275.7
259.6
161.7
146.0
127.2
115.7
110.8
106.8
415.8
341.8
161.4
143.0
120.4
115.4
108.2
108.1
362.6
308.9
165.2
134.5
124.6
117.5
115.4
112.3
260.7
256.5
C-7
C-4a
C-5
C-8a
C-6
C-8
N-2
N-4
162.3
137.9
125.9
118.3
114.4
112.5
373.3
352.6
163.8
136.9
120.5
116.4
111.5
108.6
335.9
309.7
3
HF
4
δ
δ
7-H
5-H
6-F
6.23
5.53
53.6
47.4
163.8
155.5
139.0
107.8
107.1
97.6
7-H
5-H
6-F
6.26
5.76
41.7
36.6
162.5
154.5
143.2
106.3
103.3
94.8
7-H
5-H
6-F
5.39
4.58
49.0
41.5
161.3
152.3
137.9
104.3
103.7
96.1
5-H
8-H
7-F
6-F
5.86
5.72
29.7
5-H
8-H
7-F
6.57
6.33
18.9
5-H
8-H
7-F
5.43
5.23
29.9
8-F
8-F
8-F
25.4
152.2
150.0
135.0
113.6
113.2
111.0
265.0
258.8
6-F
17.1
148.9
147.7
139.9
115.4
113.7
109.8
394.6
342.8
6-F
22.3
151.8
148.4
137.6
110.3
110.0
109.0
347.3
307.4
C[b]
C-8
C-6
C-4a
C-8a
C-7
C-5
N-4
N-2
C-6
C-8
C-4a
C-5
C-7
C-8a
N-2
N-4
C-6,C-7[b]
C-6,C-7[b]
C-4a
C-7
C-6
C-4a
C-8
C-5
C-8a
N-2
N-4
C-7
C-6
C-4a
C-8
C-5
C-8a
N-2
N-4
C[b]
C-4a
C-5
C-7
C-8a
N-2[c]
N-4[c]
C-5,C-8[b]
C-5,C-8[b]
C-8a
265.4
250.1
298.3
391.5
324.2
270.6
N-2[d]
N-4[d]
5
HF
6
δ
δ
8-H
7-F
5-F
6-F
C-7
C-5
C-6
C-4a
C-8a
C-8
N-2
N-4
5.51
30.1
16.3
8-H
7-F
5-F
6-F
C-7
C-5
C-6
C-4a
C-8a
C-8
N-2
N-4
6.06
23.1
17.9
8-H
7-F
5-F
6-F
C-7
C-6
C-5
C-4a
C-8a
C-8
N-2
N-4
4.99
31.4
23.6
7-H
8-F
5-F
6-F
6.29
41.6
32.0
7-H
8-F
6-F
5-F
C-6
C-8
C-5
C-4a
C-7
C-8a
N-2
N-4
6.37
30.0
24.8
7-H
8-F
6-F
5-F
C-6
C-8
C-5
C-4a
C-7
C-8a
N-2
N-4
5.39
35.7
30.3
3.3
-0.4
3.1
10.7
6.7
12.8
152.8
142.2
140.6
124.5
110.8
108.1
265.1
240.9
150.3
140.6
138.4
128.5
112.3
108.5
388.3
328.0
152.2
140.1
139.9
127.7
109.4
104.5
343.4
291.4
C-6,C-8[b]
C-6,C-8[b]
C-5
152.4
150.4
138.7
127.2
107.7
97.5
151.4
148.7
134.5
131.3
105.3
96.6
151.4
147.3
135.4
128.3
104.2
96.1
C-4a
C-7
C-8a
N-2
264.2
232.8
386.7
287.4
320.8
257.0
N-4
7
HF
δ
8-F,5-F[b]
5-F,8-F[b]
6-F,7-F[b]
7-F,6-F[b]
C-6,C-7[b]
C-7,C-6[b]
C-8,C-5[b]
C-5,C-8[b]
C-4a
18.5
12.2
9.9
8-F
5-F
6-F
7-F
C-6
C-7
C-8
C-5
C-4a
C-8a
N-2
N-4
11.2
11.0
3.9
8-F
5-F
7-F
6-F
C-7
C-6
C-8
C-5
C-4a
C-8a
N-2
N-4
15.8
14.3
8.1
6.9
2.9
5.3
143.0
141.3
140.4
139.5
122.0
98.9
139.8
139.3
139.0
136.3
125.2
100.4
368.5
289.7
141.9
141.1
138.9
136.9
122.9
98.1
C-8a
N-2
N-4
255.5
233.1
309.0
256.7
[a]
The doublet at δ ϭ 275.7 ppm is not changed on going to the 15N{1H} spectrum; thus, the corresponding spin-spin coupling constant
(ca. 2.2 Hz) is not J(N4-H5) (cf. ref.[3]) and the signal does not belong to N-4. Uncertain experimental assignments. The doublet of
3
[b]
[c]
doublets of doublets at δ ϭ 250.1 ppm becomes a doublet of doublets on going to the 15N{1H} spectrum while the triplet at δ ϭ
265.4 ppm is not affected. The quadruplet at δ ϭ 258.8 ppm becomes a triplet on going to the 15N{1H} spectrum while the signal at
[d]
δ ϭ 265.0 ppm is not affected.
82
Eur. J. Inorg. Chem. 2003, 77Ϫ88

Fluorinated 1,3λ4δ2,2,4-Benzodithiadiazines
FULL PAPER
lematic: DFT performs better than HF but the rms devi- forms better, the correct assignment of C-6 and C-8 remains
ation for all the compounds under consideration in Table 2 a problem.
Despite the small problems that remain, the two methods
used here perform generally well on a qualitative and some-
times even on a quantitative level. More sophisticated com-
putational methods were not evaluated since computational
time and efficiency become an important factor in those
cases, and an evaluation of that type is beyond the scope of
this study.
is still δ ϭ 103.9 ppm for HF and 60.0 ppm for DFT.
A similar observation can be made for the δ19F values.
Here, the rms differences are 8.6 and 3.6 ppm respectively.
The spread in the specific values of the different compounds
is larger for this nucleus, but HF seems to do better when
the number of fluorine atoms increases. Both methods do,
however, qualitatively reproduce the spectrum of all com-
pounds, except for 6, in which 5-F and 6-F have been
switched. The assignments of the 19F NMR spectrum of 7,
previously uncertain,[2] can now be confidently made, on
the basis of the superior DFT shifts.
The most interesting feature of the NMR spectra of pre-
viously studied compounds of series 1[1,3,4] is the enhanced
1
shielding of the H, 13C and 19F nuclei (for the monofluoro
derivatives) attached to C-5 and C-8 Ϫ i.e., those closest to
the heterocycle Ϫ as compared to those attached to C-6
and C-7, probably due to paratropic ring currents. However,
this is not the case for 3Ϫ6 (Table 2) and 7.[2] The GIAO
calculations at HF and B3LYP levels of theory (Table 2)
reproduce these shielding patterns. However, the results of
these calculations cannot be interpreted in chemical terms.
One can only think that for 3Ϫ7 the mutual influence of
the fluorine atoms dominates all other effects on δ19F.
The long-wavelength absorption maxima in the UV/Vis
spectra of compounds 3Ϫ6 (Table 5) and 7[2] are found in
the 620Ϫ635 nm range (i.e., essentially the same as for the
non-fluorinated analogues),[1,3,4] with only a weak depend-
ence upon the number and positions of the fluorine sub-
stituents. These results are in agreement with the conclu-
sion,[11] based on resonance Raman data, that the low-en-
ergy electronic transition in the UV/Vis spectra, responsible
for the blue colour of solutions of the compounds in series
1, is localised mainly on the sulfurϪnitrogen chain.
1
For the H NMR spectrum, the difference in perform-
ance of the two methods is a considerably smaller: the rms
values are now 0.56 ppm for HF and 0.49 ppm for DFT,
with HF performing better than DFT for compounds with
a larger number of fluorine atoms. The only false prediction
found for this nucleus, by both methods, is for the 6-fluoro
derivative, in which 5-H and 8-H have been switched, but
since both the calculated and the experimental differences
in chemical shifts are very small, solvent effects can be used
to account for this discrepancy.
For the δ13C values, the rms differences for both methods
are comparable: 2.6 ppm for HF and 2.7 ppm for DFT, but,
in contrast to what was found for hydrogen, the perform-
ance of HF deteriorates relative to that of DFT when the
number of fluorine atoms increases. The rather surprising
fact that DFT falters slightly in the case of the carbon
chemical shifts, when it had been clearly superior for the
other nuclei, may be explained in terms of basic density
functional theory. This indicates that the usual theorems do
not hold in the presence of an external magnetic field and
the functional has to depend both on the electron density
ρ and the current density j, induced by the magnetic
field.[29] The functionals used here, including B3LYP, do
not, and this may explain why DFT would be less reliable
for compounds or specific fragments of compounds in
which the effect of the current density J is logically larger,
such as the aromatic ring in these 1,3λ4δ2,2,4-benzodithiad-
iazines. Only for the parent compound 2 and its 6-fluoro
and 7-fluoro derivatives are the 13C chemical shifts repro-
duced perfectly by both methods. For both 4 and 6 the two
methods agree on the chemical shifts and the previously
uncertain experimental assignments can be confidently
made. For the 5-fluoro derivative of 2, DFT switches C-4a
and C-7 and incorrectly reproduces the shifts of C-6, C-8
and C-8a Ϫ the small difference calculated for C-8 and C-
6 might explain this, even though the HF calculation dis-
plays the same small difference yet makes a correct predic-
tion. In the case of compound 5, DFT switches C-5 and C-
6 but the small calculated difference may again be the cause.
For 7 the assignments of C-5 and C-8 can now be made
unambiguously, from the results of both methods, but those
of C-6 and C-7 remain doubtful. Compound 3 fares the
worst where HF is concerned, since only C-4a is correctly
2.5. Antiaromaticity
The aromaticity or antiaromaticity of π-systems is char-
acterised by a combination of energetic, geometric and
magnetic criteria. Application of these criteria usually yields
different, and very frequently contradictory, results.[5] In
these situations the Nucleus-Independent Chemical Shift
(NICS) concept[15] appears to be a useful criterion for the
quantification of (anti)aromaticity. For all 15 possible (po-
ly)fluorinated compounds, and also the parent, the NICS
values both of the carbocycle and of the heterocycle were
calculated at the HF level of theory, together with a selected
number at the DFT/B3LYP level, and studied as a function
of the number and the position of fluorine atoms.
The heterocyclic ring of the parent compound 2 has a
positive NICS value and is antiaromatic, whilst the car-
bocyclic ring has a negative value and is aromatic (Table 3).
In comparison, at the same level of theory, but with the
B3LYP/6-31G* geometry, the NICS values for benzene,
naphthalene and D4h-cyclooctatetraene are δ ϭ Ϫ9.7, Ϫ9.9
and 30.1 ppm[15] (for a minireview on the antiaromaticity
of planar cyclooctatetraene, see ref.[30]). For 1,3λ4δ2,5,2,4-
trithiadiazepine (10π-electron sulfurϪnitrogen heterocycle)
this value is δ ϭ Ϫ9.3 ppm.[31] The NICS values (Table 3)
also show that fluorination increases the aromaticity of the
carbocyclic moieties of the compounds in series 1. These
predicted and the rest is switched; even though DFT per- effects have been discussed previously with respect to fluor-
Eur. J. Inorg. Chem. 2003, 77Ϫ88
83

F. Blockhuys, A. V. Zibarev et al.
FULL PAPER
Table 3. Calculated NICS values (ppm) for all possible fluoro-substituted 1,3λ4δ2,2,4-benzodithiadiazines and the parent compound 2 in
their minimum energy conformations [B3LYP/6-311ϩG* geometries and energies, except for C1-7 for which the MP2/6-31G* geometry
was used (see ref.[9] for details)]; values at HF/6-31ϩG* and B3LYP/6-31ϩG* levels and relative energies ∆E [kJ·molϪ1] are given for
each isomer
Compound
∆E
Conformation
HF/6-31ϩG*
B3LYP/6-31ϩG*
Carbocycle Heterocycle
Carbocycle
Heterocycle
Parent (2)
0.00
C1
Cs (TS)
C1
C1
C1
Cs
C1
C1
Cs
C1
Cs
Cs
C1
Cs
Cs
Cs
Cs
Ϫ6.70
Ϫ5.21
Ϫ8.14
Ϫ8.55
Ϫ8.24
Ϫ6.99
Ϫ9.97
Ϫ9.56
Ϫ8.57
Ϫ10.30
Ϫ8.62
Ϫ8.75
Ϫ11.58
Ϫ10.13
Ϫ10.19
Ϫ10.41
Ϫ11.77
Ϫ14.16
10.74
16.48
11.02
10.41
11.12
16.02
10.66
11.31
15.81
10.64
15.67
15.37
10.85
15.36
15.09
15.32
14.94
5.90
Ϫ1.64
14.99
1.72
8.46
0.00
0.08
3.39
19.46
4.86
2.25
9.42
0.00
14.57
10.88
0.04
0.00
6.28
Ϫ
5-Fluoro
6-Fluoro
7-Fluoro
8-Fluoro
Ϫ2.80
Ϫ3.42
Ϫ3.00
15.32
14.63
15.32
5,6-Difluoro
5,7-Difluoro
5,8-Difluoro
6,7-Difluoro (4)
6,8-Difluoro (3)
7,8-Difluoro
5,6,7-Trifluoro (5)
5,6,8-Trifluoro (6)
5,7,8-Trifluoro
6,7,8-Trifluoro
5,6,7,8-Tetrafluoro (7)
Ϫ4.96
Ϫ2.49
14.77
21.68
Ϫ6.03
Ϫ3.63
15.01
21.05
Ϫ5.24
20.35
Ϫ
C1 (TS)
inated pyridines and pyridones and were explained in terms rect. Apart from the difference in scale, this is the only item
of the increased π-electron density in the ring system due that distinguishes the two methods.
to the positive mesomeric effect from the fluorine substitu-
Table 3 also lists data for two transition states (TS): the
ents.[32] The antiaromaticity of the heterocyclic moieties of planar conformation of 2 and the nonplanar one of 7. Even
the compounds in series 1 also depends on the number of though the differences between TSs and ground states are
fluorine substituents, but in a more complicated way. In not extreme, the values for the TSs Ϫ especially those for
some cases fluorine substituents increase the antiaromatic- the heterocycle Ϫ are clearly located outside the set of the
ity (Table 3), whereas in some other cases (cf. 2, 4 and 5) values for the ground states; this may be seen as an addi-
fluorination does not affect the NICS value for the hetero- tional criterion for the assignment of the electronic states
cycle.
The positions of the fluorine substituents seem to be of
of a given conformer.
From the NICS values (Table 3 and ref.[15]), one can con-
importance, especially the presence of a fluorine atom in clude that the antiaromaticity of the heterocycle moiety in
the 8-position (cf. 2 and its 8-F, 5,8-F2 and 6,7,8-F3 derivat- compound 2 is about 1/3 of the antiaromaticity of D4h
ives 3, 6 and 7 in Table 3): the very position responsible for cyclooctatetraene. According to recent estimations of reson-
the flattering of the heterocyclic ring on going from 2 to ance energy, however, the destabilization of D4h cyclooctate-
7.[9] It is clear from the results of the HF calculations in traene due to the cyclic interaction of the π bonds (i.e., the
Table 3 that the presence of the fluorine atom in the 8-posi- antiaromaticity) is negligible.[30] Taking these findings into
tion and the resulting increase in symmetry to Cs reduces account, along with the fact that only the heterocyclic moi-
the aromaticity of the carbocycle and increases the antiar- eties of the compounds in series 1 are antiaromatic, whereas
omaticity of the heterocycle, roughly by about 1 and 5 ppm, the carbocyclic moieties are aromatic (and the C4aϪC8a)
respectively. This effect can also be seen in the DFT calcula- bond belongs to both moieties at the same time), as well as
tions, but the shifts in NICS values are larger there. There other available data on these compounds’ molecular and
is no relation between NICS values and energetic stabilis- electronic structures[1Ϫ4,6Ϫ9,11] and thermal stabilities,[10,12]
ation.
one can speculate that, in terms of the aromaticity/antiar-
The NICS values of the carbocycle at the DFT level seem omaticity concept,[5] 1,3λ4δ2,2,4-benzodithiadiazines could
unexpectedly low, at least in comparison with the data of reasonably be classified on the whole as conjugated non-
other compounds described above, while those at the HF aromatic substances.
level ‘‘fit’’ better. However, from the discussion of the per-
formance of both levels of theory in the prediction of NMR
shifts it is difficult to assess which of the two methods gen-
erates the more ‘‘reasonable’’ values. On the other hand, if
the value for 2 at the DFT level is taken at face value, one
3. Conclusions
Novel fluorinated derivatives of 1,3λ4δ2,2,4-benzodithia-
might be inclined to assume the compound’s carbocycle is diazines (series 1) have been prepared and characterised
non-aromatic (NICS close to zero), which is clearly incor- structurally. They are interesting not only from structural
84
Eur. J. Inorg. Chem. 2003, 77Ϫ88

Fluorinated 1,3λ4δ2,2,4-Benzodithiadiazines
FULL PAPER
law of twinning is 100, 0Ϫ10, 00Ϫ1). Hydrogen atoms were located
on difference Fourier maps for 5 and 14, or geometrically for 3 and
6. Atomic coordinates, thermal parameters, bond lengths and bond
angles; CCDC-178645 (3), -178646 (5), -178647 (6), and -178648
(14) contain the supplementary crystallographic data for this paper.
These data can be obtained free of charge at www.ccdc.cam.ac.uk/
conts/retrieving.html or from the Cambridge Crystallographic
Data Centre, 12, Union Road, Cambridge CB2 1EZ, UK [Fax:
(internat.) ϩ 44-1223/336-033; E-mail: deposit@ccdc.cam.ac.uk].
and theoretical points of view, the latter including an exten-
sion of the aromaticity/antiaromaticity concept to non-tra-
ditional compounds, but also for applications in synthesis.
As noted above, hydrolysis of the compounds in series 1
produces 2,2Ј-diaminodiphenyl disulfides.[3] The latter are
well-known starting materials in the preparation of 1,5-
benzothiadiazepines with anti-anginal and anti-hypertens-
ive activities.[33] In this context one can believe that these
fluorinated derivatives (from this work and refs.[2Ϫ4]) have
some prospects in the synthesis of currently unknown fluor-
inated 1,5-benzothiadiazepines with potentially useful phar-
macological properties.
Theoretical calculations were performed with Gaussian98,[34] by
application of standard gradient techniques at the DFT/B3LYP
level of theory using the 6-31G* and the 6-311ϩG* basis sets on
all atoms; all basis sets were used as implemented in the program.
Initially, the geometries of all compounds were optimised at the
B3LYP/6-31G* level in Cs symmetry and force field calculations
were used to ascertain whether the resulting structures were energy
minima. When necessary, the symmetry was reduced in the sub-
sequent B3LYP/6-311ϩG* calculation. Chemical shielding factors
were calculated at all atomic positions and at both ring centres
(non-weighted mean of the heavy-atom coordinates) for the NICS
values, both at the HF/6-31ϩG* and B3LYP/6-31ϩG* levels, at
the B3LYP/6-311ϩG* geometry, by use of the GIAO method im-
plemented in Gaussian 98. Chemical shifts for the carbon and hy-
drogen atoms were obtained by subtracting the chemical shielding
values of these atoms from those calculated for tetramethylsilane
(TMS), which are δ ϭ 200.7499 and 32.6074 ppm, respectively, at
the HF/6-31ϩG* level, and δ ϭ 191.6496 and 32.1776 ppm, re-
spectively, at the B3LYP/6-31ϩG* level, based on a B3LYP/6-
311ϩG* geometry. Chemical shifts for the fluorine atoms were ob-
tained by subtracting the chemical shielding values of these atoms
from that calculated for hexafluorobenzene (C6F6), which is δ ϭ
388.5052 ppm at the HF/6-31ϩG* level and 347.9219 ppm at the
B3LYP/6-31ϩG* level, based on a B3LYP/6-311ϩG* geometry.
Chemical shifts for the nitrogen atoms were obtained by sub-
tracting the chemical shielding values of these atoms from that cal-
Experimental Section
General: The 1H, 13C and 15N NMR spectra were measured in
CDCl3 (unless otherwise indicated) with a Bruker DRX 500 spec-
trometer at frequencies of 500.13, 125.76 and 50.68 MHz, respect-
ively, with TMS and NH3 (liq.) as standards; the 19F NMR spectra
were recorded with a Bruker AC 200 machine at a frequency of
188.28 MHz with C6F6 as the standard. The high-resolution mass
spectra (EI, 70 eV) were collected with a Finnigan MAT MS-8200
instrument, and the UV/Vis spectra were recorded with Specord
M40 and HP 8453 spectrophotometers. The GC-MS measurements
of solutions in CH2Cl2 were performed with a HewlettϪPackard
G1800A GCD apparatus.
The X-ray structure determinations (Table 4) were carried out with
a Bruker P4 diffractometer with the use of Mo-Kα radiation with
a graphite monochromator, at Ϫ30 °C for 5 and 6, and at 20 °C
for 3 and 14. The structures were solved by direct methods by use
of the SHELXS-97 program and refined by the least-squares
method in the full-matrix anisotropic (isotropic for H atoms)
approximation by use of the SHELXL-93 and SHELXL-97 pro-
grams. The structure of 6 was refined as a merohedric twin (the culated for ammonia (NH3), which is 266.1027 ppm at the HF/6-
Table 4. Crystal and refinement data of compounds 3, 5, 6 and 14
3
5
6
14
Empirical formula
Formula mass
Crystal system
Space group
C6H2F2N2S2
204.22
monoclinic
P21/c
13.521(4)
7.7464(19)
7.202(2)
99.99(3)
742.9(3)
4
C6HF3N2S2
222.21
monoclinic
P21/c
3.8936(8)
5.8300(13)
32.738(6)
90.60(2)
743.1(3)
4
C6HF3N2S2
222.21
monoclinic
P21/n
3.8743(16)
7.1529(18)
27.115(8)
90.59(2)
751.4(4)
4
C9H10F4N2S2Si
314.40
monoclinic
P21/c
11.6965(12)
17.253(2)
7.1838(7)
99.768(8)
1428.7(3)
4
˚
a [A]
˚
b [A]
˚
c [A]
β [°]
3
˚
V [A ]
Z
Dc [g cmϪ3
]
1.826
0.687
408
1.986
0.714
440
1.964
0.707
440
1.462
0.484
640
µ [mmϪ1
F(000)
]
Crystal size [mm]
Scan mode
2θ range [°]
Measured reflections
Fo Ͼ 4σF
Transmission
R (obsd.)
3.20 ϫ 0.09 ϫ 0.07
θ-2θ
Ͻ 50
1315
712
0.2171Ϫ0.9535
0.0700
1.40 ϫ 0.20 ϫ 0.04
θ-2θ
Ͻ 50
1195
843
0.8668Ϫ0.9717
0.0497
5.00 ϫ 0.24 ϫ 0.06
ω
Ͻ 52.38
1173
909
0.844Ϫ0.958
0.0615
0.90 ϫ 0.14 ϫ 0.09
θ-2θ
Ͻ 50
2500
1439
0.9258Ϫ0.9631
0.0463
wR2 (all)
S
0.1824
0.876
0.1571
1.065
0.2183
1.034
0.1271
1.006
Eur. J. Inorg. Chem. 2003, 77Ϫ88
85

F. Blockhuys, A. V. Zibarev et al.
FULL PAPER
31ϩG* level and 263.9129 ppm at the B3LYP/6-31ϩG* level,
based on a B3LYP/6-311ϩG* geometry.
1-Fluoroaryl-3-trimethylsilyl-1,3-diaza-2-thiaallenes 8؊10: A solu-
tion of the corresponding ArNSO compound 11Ϫ13 (0.060 mol)
in hexane (25 mL) was added at Ϫ30 °C to a suspension of LiN-
(SiMe3)2 (10.00 g, 0.060 mol) in hexane (50 mL). The temperature
was allowed to rise to 20 °C over 2 h, and Me3SiCl (6.60 g, 0.060
mol) in hexane (10 mL) was added. The precipitate was filtered off,
the solvent was distilled off, and the residue was distilled under
vacuum. Compounds 8Ϫ10 were obtained as orange oils.
The ArFSCl compounds 28 (ArF ϭ 6-HC6F4)[35] and 30 (ArF ϭ 4-
HC6F4)[36] were prepared by known methods. The syntheses de-
scribed below were carried out under argon (except for 29, 31 and
32) in absolute solvents with stirring. The reagents were added
dropwise, and the solvents were distilled off at reduced pressure.
CsF was calcinated and SCl2 was distilled directly before use.
Tables 5 and 6 list the physical and analytical data for the com-
pounds synthesised.
1-Polyfluoroaryl-4-trimethylsilyl-2,4-diaza-1,3-dithia-2,3-butadienes
14؊18: A solution of the corresponding ArFSCl compound 29Ϫ33
(0.025 mol) in CH2Cl2 (10 mL) was added at 20 °C (this temper-
ature was Ϫ30 °C in the syntheses of 17 and 18, followed by
slow warming to 20 °C) over a period of 1 h to a solution of
Me3SiNϭ)2S[19] (5.15 g, 0.025 mol) in CH2Cl2 (25 mL). After an
additional 16 h, the solvent was distilled off, the residue was dis-
tilled under vacuum (for 15 and 16) or recrystallised from hexane
(for 14, 17 and 18), followed (for 18) by vacuum sublimation. Com-
pounds 15 and 16 were obtained as red oils, and compounds 14,
17 and 18 as transparent yellow crystals.
6,8-Difluoro- (3), 6,7-Difluoro- (4) and 5,6,7-Trifluoro-1,3λ4δ2,2,4-
benzodithiadiazine (5): Solutions of the corresponding ArϪNϭSϭ
NϪSiMe3 compounds (8Ϫ10, 0.010 mol) and SCl2 (0.77 g,
7.50 mmol) in CH2Cl2 (30 mL) were mixed by adding them to 300
mL (600 mL in the preparation of 5) of CH2Cl2 at 40 °C over a
period of 1 h. After an additional hour, the reaction mixture was
cooled to 20 °C and filtered, the solvent was distilled off, and the
residue was sublimed under vacuum and recrystallised from hexane
(to purify 5 at least three sublimation-recrystallisation cycles were
necessary). Compounds 3Ϫ5 were obtained as black crystals.
Transformations of Compounds 14 and 16؊18 by Treatment with
CsF: A solution of the corresponding compound 14 or 16Ϫ18
(0.050 mol) in MeCN (20 mL) was added over a period of 1 h to
a refluxing suspension of CsF (0.76 g, 0.005 mol) in MeCN (80
5,6,8-Trifluoro-1,3λ4δ2,2,4-benzodithiadiazine (6): A solution of
compound 15 (1.57 g, 0.005 mol) in MeCN (20 mL) was added
over 1 h to a refluxing suspension of CsF (0.76 g, 0.005 mol) in mL). After an additional 1/2 h, the reaction mixture was cooled to
MeCN (80 mL). After an additional 1/2 h, the reaction mixture
was cooled to 20 °C and filtered, the solvent was distilled off, and
the residue was sublimed under vacuum and recrystallised from
hexane. Compound 6 was obtained as black crystals.
20 °C and filtered, and the solvent was distilled off. The residue
was worked up as follows:
Starting Compound 14: The residue was chromatographed on a sil-
ica column with hexane as the eluent. 2,2Ј,3,3Ј,4,4Ј,5,5Ј-Octafluoro-
Table 5. Spectroscopic data of the compounds
NMR (δ[a] [ppm])
UV/Vis
Compound 1H
19F
13C
15N (J [Hz])
λmax [nm] (log ε)[b]
3
4
5
6
6.23, 5.53
53.6, 47.4
29.7, 25.4
30.1, 16.3, 3.3
41.6, 32.0, 10.7
52.3
163.8, 155.5, 139.0, 107.8, 265.4 (t, 2), 250.1
107.1, 97.6 (ddd, 2.5, 1.5)
152.5, 150.0, 135.0, 113.6, 265.0 (t, 2),
113.2, 111.0 258.8 (q, 2.5)
152.8, 142.2, 140.6, 124.5, 265.1, 240.9
110.8, 108.1
632 (2.70), 365 (3.37),
277 (4.23), 264 (4.03)
625 (2.60), 291 (4.18),
284 (4.18), 247 (3.81)
620 (2.58), 369 (3.12),
283 (4.11), 246 (3.79)
631 (2.61), 377 (3.45),
265 (4.10)
334 (3.81)
344 (4.13)
340 (3.92)
309 (4.13)
5.86, 5.72
5.51
6.29
152.4, 150.4, 138.7, 127.2, 264.2, 232.8
107.7, 97.5
8
9
7.21, 6.56, 0.29
7.75, 7.36, 7.02, 0.24 26.4, 25.5
10
11
12
13
14
15
16
17
18
27
7.55, 6.89, 0.23
7.60, 7.09
7.46, 7.30, 6.89
8.08, 6.99
7.41, 0.33
6.85, 0.29
7.14, 0.32
0.39
26.0, 25.0, 2.7
58.0
33.5, 30.4
36.1, 30.3, 6.9
24.3, 22.6, 7.1, 4.4
54.8, 38.0, 34.5, Ϫ0.9
29.1, 24.2
105.1, 30.7, 22.2
32.5, 15.4
316 (3.95)
376 (4.12)
347 (4.06)
349 (4.11)
352 (4.18)
362 (4.24)
306 (4.04)
0.34
7.63, 7.42
24.9, 15.9
152.4, 148.7, 145.9, 139.3, 333.3, 323.4
121.2, 116.9
(dd, 6.5, 2.7)
29
31
32
33
34
35
36
6.89
58.7, 42.5, 39.5, 0.6
105.2, 35.3, 23.0
37.2, 16.9
50.0, 33.2, 26.4, Ϫ1.5
53.5
27.3, 25.4
27.8, 24.9, 3.6
6.76, 3.48
7.01, 6.69
7.46, 7.17Ϫ7.04
7.20, 6.89
411 (4.03)
418 (4.09)
415 (3.91)
[a]
[b]
Solvents: CCl4 (8Ϫ18, 32, 33, 36Ϫ38), neat liquid (30, 35). In heptane.
86
Eur. J. Inorg. Chem. 2003, 77Ϫ88

Fluorinated 1,3λ4δ2,2,4-Benzodithiadiazines
FULL PAPER
Table 6. Characterisation of the compounds
Compound
M.p. [°C],
b.p. [°C/Torr]
Yield (%)
Empirical formula
MS: m/z [Mϩ]
found (calcd.)
3
4
59Ϫ60
73Ϫ74
60Ϫ61
58Ϫ59
65Ϫ67/1
98Ϫ100/5
90Ϫ91/4
73Ϫ75/1, 38Ϫ39
55Ϫ56/2
64Ϫ65/10
97Ϫ99/1, 57Ϫ58
107Ϫ108/1
115Ϫ117/1
111Ϫ113/2, 36Ϫ38
52Ϫ53
70Ϫ72/11
71Ϫ72/15
86Ϫ87
32
29
9
C6H2F2N2S2
C6H2F2N2S2
C6HF3N2S2
203.9625 (203.9628)
203.9636 (203.9628)
221.9579 (221.9533)
221.9535 (221.9533)
246.0458 (246.0458)
246.0459 (246.0458)
264.0363 (264.0364)
174.9913 (174.9915)
174.9918 (174.9915)
192.9809 (192.9809)
313.9986 (313.9991)
313.9994 (313.9991)
313.9985 (313.9991)
381.9868 (381.9865)
171.9904 (171.9907)
5
6
8
9
55
40
50
56
90
90
92
70
93
73
84
58
81
90
90
57
C6HF3N2S2
C9H12F2N2SSi
C9H12F2N2SSi
C9H11F3N2SSi
C6H3F2NOS
C6H3F2NOS
C6H2F3NOS
C9H10F4N2S2Si
C9H10F4N2S2Si
C9H10F4N2S2Si
C10H9F7N2S2Si
C6H2F2N2S
10
11
12
13
14
15
16
17
27
29
31
32
33
34
35
36
C6HClF4S[a]
C7ClF7S[a]
C6ClF4NO2S[a]
C6H2F4S
41Ϫ43/12
99Ϫ101
84Ϫ85
181.9808 (181.9808)
286.0179 (286.0188)
286.0188 (286.0188)
322.0003 (321.9999)
C12H6F4N2S
C12H6F4N2S
C12H4F6N2S
88Ϫ89
Found (calcd.) % Cl: 29: 16.22 (16.40); 31: 12.25 (12.48); 32: 13.70 (13.58).
1
diphenyl disulfide (25,[37] identified by high-resolution MS and H
cess of dry Cl2 was slowly passed through a solution of the corres-
ponding ArFSH [ArF ϭ
5-HC6F4 (33), 4-F3CC6F4,[41,43] 4-
and 19F NMR) was obtained as yellow oil in a 10% yield.
O2NC6F4 [44]] in CCl4 (40 mL) at ambient temperature. The solvent
was distilled off and the residue was distilled (for 29 and 31) or
sublimed (for 32) under vacuum. Compounds 29 and 31 were ob-
tained as red oils, and compound 32 as yellow crystals.
Starting Compound 16: The residue was distilled under vacuum,
yielding, according to GC-MS and 19F NMR, 0.10 g of a 1:3:0.1
mixture
of
4,5,7-trifluoro-2,1,3-benzothiadiazole
(19)[25]
2,2Ј,3,3Ј,5,5Ј,6,6Ј-octafluorodiphenyl disulfide (21)[38] and 2,3,5,6-
tetrafluoroaniline (23).[39]
2,3,4,6-Tetrafluorobenzenethiol (33): A solution of BuLi (4.78 g,
0.075 mol) in hexane (50 mL) was added at Ϫ80 °C to a solution
of 1,2,3,5-tetrafluorobenzene (11.32 g, 0.075 mol) in Et2O (70 mL).
After 1 h, finely powdered sulfur (2.40 g, 0.075 mol) was added to
the reaction solution at the same temperature. The temperature was
allowed to rise to 20 °C, and the reaction mixture was quenched
with 1:2 diluted HCl (25 mL). The organic layer was separated,
washed with H2O (20 mL) and dried with CaCl2. The solvent was
distilled off and the residue was distilled under vacuum. Compound
33 was obtained as a colourless liquid.
Starting Compound 17: The residue was distilled under vacuum,
yielding, according to GC-MS and 19F NMR, 0.35 g of a 2:1:6
mixture of 4,6,7-trifluoro-5-trifluoromethyl-2,1,3-benzothiadiazole
(20),[25]
4,4Ј-bis(trifluoromethyl)-2,2Ј,3,3Ј,5,5Ј,6,6Ј-octafluorodi-
phenyl disulfide (22),[40] and 2,3,5,6-tetrafluoro-4-(trifluoromethyl)-
aniline (24).[41]
Starting Compound 18: The residue was sublimed under vacuum
and recrystallised from hexane/CHCl3. 2,3,5,6-Tetrafluoro-4-nitro-
aniline (26,[42] identified by comparison of its m.p. and H and 19
F
1
Fluorinated 1,3-Diaryl-1,3-diaza-2-thiaallenes (34؊36): The
(ArϪNϭ)2S compounds [Ar ϭ 3,5-F2C6H3 (34), 3,4-F2C6H3 (35),
2,3,4-F3C6H2 (36)] were obtained as by-products in the prepara-
tions of compounds 3Ϫ5, respectively, and were isolated as orange-
yellow crystals by fractional sublimation followed by recrystallis-
ation from hexane.
NMR spectra with those of an authentic[42] sample) was obtained
as yellow crystals in 74% yield.
4,5-Difluoro-2,1,3-benzothiadiazole (27): A solution of compound
10 (1.32 g, 0.005 mol) in MeCN (20 mL) was added over 1 h to a
refluxing suspension of CsF (0.76 g, 0.005 mol) in MeCN (80 mL).
After an additional hour, the reaction mixture was cooled to 20
°C and filtered, the solvent was distilled off, and the residue was
chromatographed on a silica column (with CHCl3 as the eluent)
and sublimed under vacuum followed by recrystallisation from hex-
ane. Compound 27 was obtained as colourless crystals.
Acknowledgments
The Novosibirsk authors are grateful to Mrs. Nina F. Yakusheva
for her assistance in the preparations, and to the Russian Founda-
tion for Basic Research for financial support of their work (grants
99-03-33115 and 01-03-06190) including instrumental facilities at
the Collective Analytical Center of the Siberian Division of the
Russian Academy of Sciences (grant 00-03-40135), and for access
to the Cambridge Structural Database (grant 99-07-90133) and to
the STN International databases via the STN Center at their Insti-
Polyfluorinated N-Sulfinylarylamines 11؊13: The ArϪNϭSϭO
compounds were prepared by Michaelis reaction from the corres-
ponding ArNH2 (Aldrich) and SOCl2 in benzene. After the usual
workup, compound 11 was obtained as yellow crystals (from hex-
ane), and compounds 12 and 13 as yellow oils.
2,3,4,6-Tetrafluoro- (29), 2,3,5,6-Tetrafluoro-4-trifluoromethyl- (31) tute (grant 00-03-32721). A.Yu. M. is also grateful to the Siberian
and 2,3,5,6-Tetrafluoro-4-nitrophenylsulfenyl Chloride (32): An ex-
Division of the Russian Academy of Sciences for the N. N. Vo-
Eur. J. Inorg. Chem. 2003, 77Ϫ88
87

F. Blockhuys, A. V. Zibarev et al.
G. M. Brooke, J. Fluorine Chem. 1997, 86, 1Ϫ76.
A. V. Zibarev, E. Lork, R. Mews, Chem. Commun. 1998,
991Ϫ992.
T. Borrmann, A. V. Zibarev, E. Lork, G. Knitter, S. J. Chen, P.
G. Watson, E. Cutin, M. M. Shakirov, W. D. Stohrer, R. Mews,
Inorg. Chem. 2000, 39, 3999Ϫ4005.
FULL PAPER
rozhtsov Award for young scientists. F. B. and C. V. A. gratefully
acknowledge support by the University of Antwerp under grant
GOA-BOF-UA No. 23.
[22]
[23]
[24]
[25]
[1]
A. W. Cordes, M. Hojo, H. Koenig, M. C. Noble, R. T. Oakley,
W. T. Pennington, Inorg. Chem. 1986, 25, 1137Ϫ1145.
A. V. Zibarev, A. O. Miller, J. Fluorine Chem. 1990, 50,
359Ϫ363.
[2]
A. V. Zibarev, Yu. V. Gatilov, A. O. Miller, Polyhedron 1992,
[26] [26a]
11, 1137Ϫ1141.
I. Yu. Bagryanskaya, Yu. V. Gatilov, A. Yu. Makarov, A. M.
S. Doraiswamy, S. D. Sharma, J. Mol. Struct. 1983, 102,
[3]
[26b]
81Ϫ92.
L. Nygaard, E. R. Hansen, R. L. Hansen, J.
Maksimov, A. O. Miller, M. M. Shakirov, A. V. Zibarev, Het-
eroatom Chem. 1999, 10, 113Ϫ124.
Rastrup-Andersen, G. O. Soerensen, Spectrochim. Acta A 1967,
23, 2813Ϫ2819.
[4]
[27]
[28]
A. Yu. Makarov, I. Yu. Bagryanskaya, Yu. V. Gatilov, T. V.
M. M. Shakirov, I. V. Beregovaya, A. V. Zibarev, Khim. Geter-
otsikl. Soedin. 1995, 262Ϫ265 [Chem. Abstr. 1995, 123, 55291j].
I. Yu. Bagryanskaya, H. Bock, Yu. V. Gatilov, A. Haas, M. M.
Shakirov, B. Solouki, A. V. Zibarev, Chem. Ber./Recueil 1997,
130, 247Ϫ253.
W. Koch, M. C. Holthausen, A Chemist’s Guide to Density
Functional Theory, Wiley-VCH, New York, 2001, chapter 11.
F. G. Klaerner, Angew. Chem. Int. Ed. 2001, 40, 3977Ϫ3981.
S. Martin-Santamaria, B. Lavan, H. S. Rzepa, Chem. Commun.
2000, 1089Ϫ1090.
P. A. Hunt, T. Fisher, P. Schwerdtfeger, J. Org. Chem. 1997,
62, 8063Ϫ8070.
A. R. Katritzky, B. V. Rogovoy, C. Chassaing, V. Vvedensky,
B. Forood, B. Flatt, H. Nakai, J. Heterocycl. Chem. 2000, 37,
1655Ϫ1658.
M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M.
A. Robb, J. R. Cheeseman, V. G. Zakrzewski, J. A. Montgom-
ery, Jr., R. E. Stratmann, J. C. Burrant, S. Dapprich, J. M.
Millam, A. D. Daniels, K. N. Kudin, M. C. Strain, O. Farkas,
J. Tomasi, V. Barone, M. Cossi, R. Cammi, B. Mennucci, C.
Pomelli, C. Adamo, S. Clifford, J. Ochterski, G. A. Petersson,
P. Y. Ayala, Q. Cui, K. Morokuma, D. K. Malick, A. D. Rab-
uck, K. Raghavachari, J. B. Foresman, J. Cioslowski, J. V. Or-
tiz, A. G. Baboul, B. B. Stefanov, G. Liu, A. Liashenko, P.
Piskorz, I. Komaromi, R. Gomperts, R. L. Martin, D. J. Fox,
T. Keith, M. A. Al-Laham. C. Y. Peng, A. Nanayakkara, C.
Gonzales, M. Challacombe, P. M. W. Gill, B. Johnson, W.
Chen, M. W. Wong, J. L. Andres, C. Gonzalez, M. Head-Gor-
don, E. S. Replogle, J. A. Pople, Gaussian 98, Revision A.7,
Gaussian Inc., Pittsburgh PA, 1998.
T. P. Culberton, J. Heterocycl. Chem. 1991, 28, 1701Ϫ1703.
P. F. Butler, M. E. Peach, J. Fluorine Chem. 1987, 35, 489Ϫ496.
G. M. Brooke, M. A. Quasem, J. Chem. Soc. C 1967, 865Ϫ869.
P. Robson, M. Stacey, R. Stephens, J. C. Tatlow, J. Chem. Soc.
1960, 4754Ϫ4760.
G. M. Brooke, J. Burdon, M. Stacey, J. C. Tatlow, J. Chem.
Soc. 1960, 1768Ϫ1771.
G. G. Furin, T. V. Terentieva, G. G. Yakobson, Izv. Sib. Otd.
Akad. Nauk SSSR Ser. Khim. Nauk 1972, 6, 78Ϫ88 (in Rus-
sian).
D. J. Alsop, J. Burdon, J. C. Tatlow, J. Chem. Soc. 1962,
1801Ϫ1805.
G. M. Brooke, J. Burdon, J. C. Tatlow, J. Chem. Soc. 1961,
802Ϫ807.
A. M. Maksimov, V. E. Platonov, Fluorine Notes 1999, 4.
www.fluorine.ru\Notes\archive.html
Mikhalina, M. M. Shakirov, L. N. Shchegoleva, A. V. Zibarev,
Heteroatom Chem. 2001, 12, 563Ϫ576.
[5] [5a]
V. I. Minkin, M. N. Glukhovtsev, B. Ya. Simkin, Aro-
maticity and Antiaromaticity. Electronic and Structural Aspects,
[5b]
[29]
Wiley, New York, 1994, 336 pp.
K. B. Wiberg, Chem. Rev.
2001, 101, 1317Ϫ1331 (and other articles of this special issue).
[5c]
[30]
[31]
M. K. Cyranski, T. M. Krygowsky, A. R. Katritzky, P. von
R. Schleyer, J. Org. Chem. 2002, 67, 1333Ϫ1338.
A. P. Hitchcock, R. S. DeWitte, J. M. Van Esbroek, P. Aebi,
C. L. French, R. T. Oakley, N. P. C. Westwood, J. Electron
Spectrosc. Relat. Phenom. 1991, 57, 165Ϫ187.
N. E. Petrachenko, Yu. V. Gatilov, A. V. Zibarev, J. Electron
Spectrosc. Relat. Phenom. 1994, 67, 489Ϫ492.
A. V. Manaev, A. Yu. Makarov, Yu. V. Gatilov, J. N. Latosin-
ska, V. V. Shcherbukhin, V. F. Traven, A. V. Zibarev, J. Electron
Spectrosc. Relat. Phenom. 2000, 107, 33Ϫ38.
F. Blockhuys, S. L. Hinchley, A. Yu. Makarov, Yu. V. Gatilov,
A. V. Zibarev, J. D. Woollins, D. W. H. Rankin, Chem. Eur. J.
2001, 7, 3592Ϫ3602.
I. V. Vlasyuk, V. A. Bagryansky, N. P. Gritsan, Yu. N. Molin,
A. Yu. Makarov, Yu. V. Gatilov, V. V. Shcherbukhin, A. V.
Zibarev, Phys. Chem. Chem. Phys. 2001, 3, 409Ϫ415.
M. Schettino, R. M. Romano, C. O. Della Vedova, A. Yu.
Makarov, A. V. Zibarev, Phys. Chem. Chem. Phys. 2001, 3,
1411Ϫ1418.
V. A. Bagryansky, I. V. Vlasyuk, Yu. V. Gatilov, A. Yu. Maka-
rov, Yu. N. Molin, V. V. Shcherbukhin, A. V. Zibarev, Mende-
leev Commun. 2000, 5Ϫ7.
N. P. Gritsan, V. A. Bagryansky, I. V. Vlasyuk, Yu. N. Molin,
A. Yu. Makarov, M. S. Platz, A. V. Zibarev, Russ. Chem. Bull.,
Int. Ed. 2001, 50, 2064Ϫ2070.
A. V. Zibarev, Yu. V. Gatilov, I. Yu. Bagryanskaya, A. M. Mak-
simov, A. O. Miller, J. Chem. Soc., Chem. Commun. 1993,
298Ϫ299.
P. von R. Schleyer, C. Maerker, A. Dransfeld, H. Jiao, N. J. R.
van Eikema Hommes, J. Am. Chem. Soc. 1996, 118,
6317Ϫ6318.
E. Lork, R. Mews, M. M. Shakirov, P. G. Watson, A. V. Ziba-
rev, Eur. J. Inorg. Chem. 2001, 2123Ϫ2134.
T. Borrmann, E. Lork, R. Mews, W. D. Stohrer, P. G. Watson,
A. V. Zibarev, Chem. Eur. J. 2001, 7, 3504Ϫ3510.
A. V. Zibarev, A. O. Miller, Yu. V. Gatilov, G. G. Furin, Het-
eroatom Chem. 1990, 1, 443Ϫ453.
I. Yu. Bagryanskaya, Yu. V. Gatilov, A. O. Miller, M. M.
Shakirov, A. V. Zibarev, Heteroatom Chem. 1994, 5, 561Ϫ565.
T. Chivers, Chem. Rev. 1985, 85, 314Ϫ365.
[6]
[32]
[33]
[7]
[8]
[34]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[35]
[36]
[37]
[38]
[39]
[40]
[16]
[17]
[18]
[19]
[41]
[42]
[43]
[44]
[20]
[21]
D. J. Hupe, E. R. Pohl, Isr. J. Chem. 1985, 26, 395Ϫ399.
A. V. Zibarev, Yu. V. Gatilov, I. Yu. Bagryanskaya, S. N. Kon-
chenko, Polyhedron 1992, 11, 2787Ϫ2793.
Received May 3, 2002
[I02284]
88
Eur. J. Inorg. Chem. 2003, 77Ϫ88