Isolation of a labile homoleptic diazenium cation

Article abstract of DOI:10.1002/anie.201310186

Following our interest in nitrogen chemistry, we now describe the synthesis, structure, and bonding of labile disilylated diazene, its GaCl ₃ adduct, and the intriguing trisilylated diazenium ion [(Me ₃Si)₂Ni=N-SiMe₃]⁺, a dark blue and highly labile (T_decomp>-30 °C) homoleptic cation of the type [R₃N₂]⁺. Although direct silylation of Me ₃Si-Ni=N-SiMe₃ failed, the [(Me₃Si) ₂Ni=N-SiMe₃]⁺ ion was generated in a straightforward two-electron oxidation reaction from mercury(II) dihydrazide and Ag[GaCl₄]. Moreover, previous structure data of Me ₃Si-Ni=N-SiMe₃ were revised on the basis of new data. Cold, blue, and positive: The first homoleptic diazenium cation has been formed by treating mercury(II) dihydrazide with Ag[GaCl₄] at low temperature. While numerous attempts at the direct silylation of diazene failed, this alternative new two-electron oxidation of the trisilylated hydrazide ion affords blue, highly labile [(Me₃Si)₂Ni=N-SiMe ₃]⁺ ions in almost quantitative yield.

Full text of DOI:10.1002/anie.201310186

.
Angewandte
Communications
DOI: 10.1002/anie.201310186
Diazenium Ions
Isolation of a Labile Homoleptic Diazenium Cation**
Wolfgang Baumann, Dirk Michalik, Fabian Reiß, Axel Schulz,* and Alexander Villinger*
Abstract: Following our interest in nitrogen chemistry, we now
describe the synthesis, structure, and bonding of labile
range for known trans diazenes (1.219–1.254 ꢁ; cf. ꢀrcov-
(N=N) = 1.20 ꢁ),[6] and therefore provoked controversial
disilylated diazene, its GaCl3 adduct, and the intriguing
remarks in the literature. Moreover, computational studies
+
=
ꢀ
trisilylated diazenium ion [(Me3Si)2N N-SiMe3] , a dark
predicted longer N N bonds, which were consistent with
blue and highly labile (Tdecomp > ꢀ308C) homoleptic cation of
experimental and theoretical trends found for other di-
+
the type [R3N2] . Although direct silylation of Me3Si-N N-
azenes.[5,7]
=
+
=
=
SiMe3 failed, the [(Me3Si)2N N-SiMe3] ion was generated in
Reports on the chemistry of diazenium cations [R2N N-
a straightforward two-electron oxidation reaction from mer-
cury(II) dihydrazide and Ag[GaCl4]. Moreover, previous
R]+ (R = alkyl, aryl) are restricted to only a few examples or
are even unknown (R = SiMe3) in the literature.[8–10] Herein,
we focus on the synthesis of highly labile silylated diazenium
=
structure data of Me3Si-N N-SiMe3 were revised on the basis
of new data.
species. Interestingly, the formation of mostly transient
+[11]
=
diazenium ions such as [Ph2N N-H]
was previously
T
he rise of the synthetic dyestuffs industry during the 19th
discussed in Diels–Alder, thermolysis, and solvolysis reac-
tions, or electrochemically induced redox processes, often
without any characterization.[12–15]
As shown by Wiberg, the reaction of tosylazide with 1 at
low temperatures provides ready access to light blue,
disilylated trans diazene 2 in approximately 65% yield
(Scheme 1, Figure 1 left).[4] The major challenge when dealing
with 2 arises from its very low decomposition point Tdecomp
century began with the discovery of diazonium salts, for
+
ꢀ
ꢁ
example, [R-N N] Cl , and diazotation reactions by Griess
as early as 1858.[1,2] Azo dyes, compounds that usually contain
=
two aromatic fragments connected by a N N bond, still have
a large impact and versatile scientific, medical, ecological, and
technical importance.[3] Although aryl-substituted diazenes
have been extensively investigated and represent a class of
stable compounds, much less is known about alkyl- and silyl-
=
substituted diazenes R-N N-R (also known as diimines). It
was not until 1968 that Wiberg et al. succeeded in the isolation
of pale blue N,N’-bis(trimethylsilyl)diazene (2) from the
reaction of p-tosylazide with lithium tris(trimethylsilyl)hy-
drazide (1) at low temperatures (Scheme 1).[4] In 1974 Veith
and Bꢀrnighausen published a single-crystal structure deter-
mination of 2, which confirmed the postulated trans config-
uration but revealed an extraordinary short N–N distance of
1.171(7) ꢁ.[5] This value, however, lies outside the reported
Figure 1. ORTEP representation of the molecular structure of 2 (left)
and 3 (right) in the crystal. Thermal ellipsoids correspond to 50%
probability at 173 K.
=
=
< ꢀ408C (cf. tBu-N N-tBu > 2008C, Ph-N N-Ph > 6008C),
which even lies below its melting point of about ꢀ38C.[16,17]
Since diazene 2 decomposes at ambient temperatures with
a half-time of 7.6 h (zero-order reaction, in a 5:1 mixture of n-
pentane/toluene) it must be kept at temperatures below
ꢀ508C in solution or in the solid state (see Figure S14 in the
Supporting Information).[18] A suitable low-temperature syn-
Scheme 1. Synthesis of N,N’-bis(trimethylsilyl) diazene (2); Tos=p-
MeC6H4SO2.
[*] Dr. W. Baumann, Dr. D. Michalik, Prof. Dr. A. Schulz, Dr. A. Villinger
Universitꢀt Rostock, Institut fꢁr Chemie
Albert-Einstein-Strasse 3a, 18059 Rostock (Deutschland)
and
Leibniz-Institut fꢁr Katalyse e.V. an der Universitꢀt Rostock
Albert-Einstein-Strasse 29a, 18059 Rostock (Germany)
E-mail: axel.schulz@uni-rostock.de
thesis had to be found to generate the tris(trimethylsilyl)dia-
+
zenium ion [(Me3Si)2N N-SiMe3] from diazene 2.[19] We
=
=
thought that the obvious route to the desired [(Me3Si)2N N-
SiMe3]+ cation was the direct silylation of diazene 2 with
salts of the type [Me3Si·LB]+[B(C6F5)4]ꢀ or [Me3Si]+[CB]ꢀ at
low temperatures (Scheme 2, route A; LB = weak Lewis
base = toluene, Me3Si-H, Me3Si-X with X = Cl, F; CB =
[CHB11H5X6], X = Br, Cl). The calculated gas-phase reac-
tion (1)[18g] with DH298 = ꢀ57.1 kcalmolꢀ1 also supported
the idea of direct silylation (cf. ꢀ8.9N2, ꢀ30.4toluene,
alexander.villinger@uni-rostock.de
Homepage: http://www.chemie.uni-rostock.de/ac/schulz
[**] We are indebted to N. Wiberg, G. Fischer, and H. Brand (Ludwig-
Maximilians-Universitꢀt Mꢁnchen).
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201310186.
3250
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2014, 53, 3250 –3253

Angewandte
Chemie
be promising candidates, since the 100% signal in the CI+
mass spectra of 5[GaCl4] and 6 could be assigned to
diazenium cation 4+, thus indicating their potential as
diazenium sources. Indeed, the reaction of 5[GaCl4] with
GaCl3 and elemental chlorine or sulfuryl chloride as the
oxidizer in dichloromethane at temperatures below ꢀ508C
gave, after filtration to remove precipitates, concentration in
vacuum, and storage at ꢀ808C, low yields of blue, highly
temperature-sensitive crystals, which were identified as 4-
[GaCl4] by X-ray studies (Scheme 3, top).[18] However, the
decomposition of the bismuth salt 5[GaCl4] to 4[GaCl4] and
bismuth chlorides was difficult to carry out because of the
high reactivity of Cl2 and SO2Cl2 leading to many side
reactions and low yields.
Finally, we found that treatment of mercury salt 6 with two
equivalents of Ag[GaCl4] at ꢀ808C led in a clean, almost
quantitative reaction to pure, crystalline 4[GaCl4] as well as
elemental mercury and silver. The latter form an amalgam
which can be easily removed by low-temperature filtration
(Scheme 3, bottom).[18h] It is interesting to note, that the
Scheme 2. Unsuccessful attempts for the synthesis of salts bearing the
diazenium ion 4+ (A: [wca]=[B(C6F5)4], [CHB11H6X6], X=Br, Cl; B/C:
[wca]=[B(C6F5)4], [SbCl6], [AsF6], [GaCl4]).
ꢀ31.3Me3Si-H,
ꢀ34.8Me3Si-F,
ꢀ54.4Me3Si-CN,
and
[19]
ꢀ71.7 kcalmolꢀ1 for Me3Si-N C N-SiMe3).
= =
Me3Si-N¼N-SiMe3ðgÞ þ ½Me3Siꢂþ ! ½ðMe3SiÞ2N¼N-SiMe3ꢂþ
ð1Þ
ðgÞ
ðgÞ
Reactions of all the utilized silylating agents (Scheme 2)
with an excess of 2 in the absence of solvent led to violent
decomposition even at very low temperatures (< ꢀ1508C),
but the reactions in all the applied solvents were also not
successful due to decomposition of the reactants and the
[B(C6F5)4]ꢀ ion. In contrast, [Me3Si]+[CHB11H5X6]ꢀ (X = Cl,
Br) did not decompose under these reaction conditions;
however, because of the limited solubility of
[Me3Si]+[CHB11H5X6]ꢀ no reaction with 2 below its decom-
position temperature was observed. Since all these attempts
failed, we then treated 2 with Me3Si-X (X = Cl, I) in the
presence of Ag[wca] (wca = weakly coordinating anion) or
strong Lewis acids (LA) such as GaCl3 and SbCl5 (Scheme 2,
routes B and C; [wca] = [AsF6], [SbCl6], [B(C6F5)4], [GaCl4]).
Again, only decomposition or undesired side reactions were
observed. Although the initially precipitation of AgX (X = I,
Cl) was observed, all the reactions led to the complete
decomposition of 2 into N2 and Me3Si-X, along with the
formation of E(III)X3 (E = As, Sb, Ga; X = F, Cl).[18] This is in
accord with the observation by Wiberg that 2 reduces main
group element chlorides such as GeCl4 with release of N2 and
Me3Si-Cl.[16c] However, the reaction of 2 with GaCl3 as the
Lewis acid or with Ag[GaCl4]/Me3Si-I in dichloromethane at
low temperatures resulted in the formation of the diazene
Scheme 3. Synthesis of 4[GaCl4].
analogous reaction of 6 with other silver salts such as
Ag(toluene)3[B(C6F5)4], Ag[CF3SO3], Ag[CHB11H5X6], or
Ag[AsF6] was not successful.[18] The overall process repre-
sents a remarkable two-electron oxidation of the [(Me3Si)2N-
N-SiMe3]ꢀ ion according to the following formal Equa-
tions (2) and (3).[18h]
2 ½ðMe3SiÞ2N-N-SiMe3ꢂꢀ ! 2 ½ðMe3SiÞ2N¼N-SiMe3ꢂþ þ 4 eꢀ
Hg2þ þ 2 Agþ þ 4 eꢀ ! HgAg2
ð2Þ
ð3Þ
=
adduct Me3Si-N N-SiMe3·GaCl3 (3, Figure 1, right) in the
form of dark blue crystals (yield ca. 95%), which were
thermally stable up to ꢀ208C (Scheme 2, routes B and C).
Although gas-phase computations predict an almost
barrier-free approach of the [Me3Si]+ ion to diazene 2 (see
Figure S16 in the Supporting Information), all the synthetic
routes used did not succeed in preparing the diazenium ion.[18]
We therefore decided to attempt the preparation of the
Blue 4[GaCl4] is extremely air and moisture sensitive, and
decomposes at temperatures above ꢀ308C. In solution it also
decomposes rapidly at temperatures above ꢀ308C. Nonethe-
less, 4[GaCl4] could be fully characterized by low-temper-
1
ature H, 13C, 14/15N, 29Si NMR, and Raman spectroscopy as
well as by single-crystal X-ray diffraction (Table 1).15N NMR
spectroscopy (ꢀ608C, CD2Cl2) is particularly well suited to
distinguish between two-coordinate and the three-coordinate
nitrogen atoms found in diazene species (2: d = 605 (N1/N2)
versus 4+: 612 (N1) and 214 ppm (N2)). Silylation at N2
results in the resonance for N2 being dramatically upfield
shifted by more than Dd = 404 ppm, while the resonance for
+
ꢀ
=
diazenium salt [(Me3Si)2N N-SiMe3] [GaCl4] (4[GaCl4]) by
redox reactions starting from heavy metal hydrazide deriva-
tives already bearing a [(Me3Si)2N-N-SiMe3]ꢀ unit. For
example, the dihydrazinobismuth cation in [Bi{N(SiMe3)N-
(SiMe3)2}2]+[GaCl4]ꢀ (5[GaCl4])[20] and the neutral mercu-
ry(II) dihydrazide Hg[N(SiMe3)N(SiMe3)2]2 (6)[21] seemed to
Angew. Chem. Int. Ed. 2014, 53, 3250 –3253
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3251

.
Angewandte
Communications
Table 1: Spectroscopic data of diazene species 2, 3, 4[GaCl4] and for
comparison mercury(II) dihydrazide 6.
Species[a]
Mp/Tdecomp/8C
2
3
4[GaCl4]
–/ꢀ30
6
ꢀ3/ꢀ30[d]
36/ꢀ20
–/156
NMR[a]
15N, N1
605
605
5.2
5.2
ꢀ4.0
ꢀ4.0
0.17
596
248
23.8
45.6
612
214
24.26
44.99
ꢀ269
15N, N2
ꢀ309
29Si, N1
9.77
29Si, N2
5.18
3.09
2.63
0.277
13C, N1
ꢀ1.6
ꢀ0.9
13C, N2
ꢀ1.8
ꢀ0.4
Figure 2. Left: ORTEP representation of the molecular structure of 4+
in the crystal. Thermal ellipsoids correspond to 50% probability at
173 K. Right: Three-dimensional representation of the ELF at 0.82.
1H, N1
0.64
0.70
1H, N2
0.17
0.46
0.55
0.284
nNN,Raman/cmꢀ1
N1-N2/ꢂ
N1-Si1/ꢂ
N2-Si2/ꢂ
N2-Si3/ꢂ
Si1-N1-N2/8
Si1-N1-N2-Si2/8
q(N1)/e
1555[b]
1.227(2)[e]
1.807(3)[e]
1.807(3)[e]
–
1568[b]
1.243(2)
1.826(2)
1.884(2)
–
134.6(1)
175.5(1)
ꢀ0.43
ꢀ0.59
0.24
1570[b]
1.254(2)
1.829(1)
1.894(1)
1.899(1)
140.5(1)
175.3(1)
ꢀ0.40
ꢀ0.59
0.33
1062
1.462(4)
1.725(3)
1.735(3)
1.737(3)
123.8(2)
ꢀ98.0(3)[g]
ꢀ1.15
ꢀ1.21
0.48
mentioned that 2, 3, and 4[GaCl4] quickly decompose in the
laser beam to form molecular nitrogen and the hydrazine
derivative (Me3Si)4N2, which is in accord with observations by
Wiberg et al.[4]
X-ray diffraction analysis (Figure 2) revealed unequivo-
cally the existence of diazenium cation 4+ with an almost
planar N2Si3 skeleton (deviation from planarity less than 58).
There are neither significant cation–anion nor anion–anion
contacts. Both N atoms are in a planar environment with two
117.3(2)[e]
180
ꢀ0.55
ꢀ0.55
–
q(N2)/e
Q
CT/e[f]
[a] d [ppm], ꢀ758C, CD2Cl2. [b] The sample decomposes in the laser
beam. [c] No signals were detected. [d] Slow decomposition starts
earlier. [e] Only structural parameters of part A with the larger occupancy
of the disordered molecule are listed. [f] QCT =total charge transfer from
the diazene to the Lewis acid. [g] The hydrazide group adopts an almost
perpendicular arrangement as a result of Pauli repulsion between the
lone pairs of electrons localized on the two N atoms (see Figure S1).
ꢀ
slightly different sets of N Si bond lengths (N1-Si1 1.829(1),
N2-Si2 1.894(1), and N2-Si3 1.899(1); cf. ꢀrcov(Si-N) =
1.87 ꢁ),[6] which can be attributed to hyperconjugation of
the lone pair of electrons localized at N1 into the N2-Si and
Si1-C s* orbitals, respectively. The most striking feature is the
ꢀ
N N bond length of 1.254(2) ꢁ, which lies in the typical range
=
=
ꢁ
dicoordinated N1 is only shifted by about d = 8 ppm to
a higher field (cf. d = 357/ꢀ11 ppm by the addition of GaCl3 to
3). These values display an extreme nuclear magnetic
deshielding of both nitrogen atoms in 2 and N1 in 3 and 4+,
a fact which was already observed and studied by Wrack-
of a N N bond (cf. ꢀrcov(N-N) = 1.42, N N 1.20, N N 1.08 ꢁ;
1.243(3) ꢁ in azobenzene)[6,24] and in accord with those found
in tetracyclic and hexacyclic formal bis(trialkyldiazenium)
bis(tetrafluoroborate) salts reported by Nelsen et al.
(1.244(5)–1.261(5) ꢁ), where the N2 units are, however, part
of a six-membered ring.[10b] Comparison of the structural data
of 4+ with those of 2 and 3 reveal only small differences
(Table 1). For all species, the N-Si skeleton is planar and two
meyer and co-workers in a series of diazenes (cf. 2: d = 618–
[18j]
=
= =
630, Ph-N N-SiMe3: 251,
Ph-N N-Ph: 131, Et-N N-Et:
151 ppm).[17a,22] Wrackmeyer and co-workers found that the
remarkable downfield shift of d[14N] upon substituting one or
slightly different Si N bond lengths are observed, except
for the C2-symmetric diazene (trans configuration). The
ꢀ
=
both alkyl or phenyl groups in R-N N-R by Me3Si can be
ꢀ
attributed to a decrease in the energy gap for magnetic dipole
allowed electronic transitions, in particular of the n!p* type,
and thus reflects the unique influence of silyl substituents on
the electronic structures, and is also responsible for the blue
color.[22,23] Thus, the 15N resonance of azo compounds of the
shortest Si N bonds are found in 2 because of stronger
hyperconjugative effects. The two Me3Si groups in 2 and 3
adopt a trans arrangement. The Ga-N donor–acceptor bonds
in 3 at 2.036(2) ꢁ[25] are in the expected range (cf. 2.011(4)
in NH(SiMe3)2·GaCl3, 1.965(2) ꢁ in Me3Si-N = S = N-
SiMe3·GaCl3).[26,27]
=
type R-N N-SiMe3 was observed approximately in the
middle of the 15N resonances of the symmetrically substituted
compounds. A similar effect can be discussed when 2 is
silylated to yield 4+. The 29Si NMR data illustrate that when
silylation occurs both 29Si resonances are shifted downfield
with respect to those of diazene 2 by Dd = 19 (N1) and 39 ppm
(N2) (2: d = 5.2 (N1/N2) versus 4+: 24.3 (N1) and 45.0 ppm
(N2)). The Raman data of all the considered diazene species
(Table 1) show sharp bands in the expected region between
1560 and 1600 cmꢀ1, which can be assigned to the nNN
Finally, it is interesting to note that the previous structure
determination of diazene 2 needs to be revised on the basis of
our data.[5] Although the cell data are in agreement (space
group P21/c, Z = 2), the previously reported structure of 2 was
refined without taking into account that the whole molecule is
disordered, which results for example in an unrealistically
ꢀ
short N N bond length of 1.171(7) ꢁ and a large N-N-Si angle
of 120.0(4)8. However, if the molecule is refined as disordered
over two positions then a significantly longer N-N distance of
stretching frequency (cf. 1589 cmꢀ1 in Ph-N N-Ph). The
1.227(2) ꢁ (cf. ꢀrcov(N N) = 1.20 ꢁ) and a smaller N-N-Si
=
=
coordination of [Me3Si]+ and GaCl3 to diazene 2 causes no
significant shift in the band to lower wave numbers, thus
indicating an almost pure nonbonding character of the lone
pair of electrons at the nitrogen donor center. It should be
angle of 117.3(2)8 in accordance with the computations are
obtained (e.g. 1.247 from B3LYP/6-311 + G(d,p),[22] 1.242 ꢁ
from M06-2X/aug-cc-pvTZ), which now lie in the expected
range for diazenes.[18]
3252
www.angewandte.org
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2014, 53, 3250 –3253

Angewandte
Chemie
[2] R. Wizinger-Aust, Angew. Chem. 1958, 70, 199.
[3] A. Bafana, S. Devi, T. Chakrabarti, Environ. Rev. 2011, 19, 350.
[4] N. Wiberg, W.-C. Joo, W. Uhlenbrock, Angew. Chem. 1968, 80,
661.
[5] M. Veith, H. Bꢀrnighausen, Acta Crystallogr. 1974, B30, 1806.
[6] P. Pyykkç, M. Atsumi, Chem. Eur. J. 2009, 15, 12770.
[7] A. Modelli, F. Scagnolari, D. Jones, G. Distefano, J. Phys. Chem.
A 1998, 102, 9675.
[8] H. Wieland, E. Wecker, Chem. Ber. 1910, 43, 3260.
[9] U. Kuhlmann, DE1419789B1, 1969.
[10] a) S. F. Nelsen, R. T. Landis II, J. Am. Chem. Soc. 1974, 96, 1788;
b) S. F. Nelsen, H. Chang, J. J. Wolff, D. R. Powell, J. Org. Chem.
1994, 59, 6558.
[11] S. Hꢂnig, H. Balli, E. Breither, F. Brꢂhne, H. Geiger, E. Grigat,
F. Mꢂller, H. Quast, Angew. Chem. 1962, 74, 818; Angew. Chem.
Int. Ed. Engl. 1962, 1, 640.
[12] See, for example: a) G. Cauquis, M. Genies, Tetrahedron Lett.
1968, 32, 3537; b) G. Cauquis, M. Genies, Tetrahedron Lett. 1971,
48, 4677.
[13] A. Ferguson, Tetrahedron Lett. 1973, 30, 2889.
[14] a) W. R. McBride, H. W. Kruse, J. Am. Chem. Soc. 1957, 79, 572;
b) W. R. McBride, E. M. Bens, J. Am. Chem. Soc. 1959, 81, 5546.
[15] E. Allred, J. Am. Chem. Soc. 1982, 104, 5422.
[16] a) N. Wiberg, W. Uhlenbrock, W. Baumeister, J. Organomet.
Chem. 1974, 70, 259; b) N. Wiberg, W. Uhlenbrock, J. Organo-
met. Chem. 1974, 70, 239; c) N. Wiberg, W.-C. Joo, G. Schwenk,
W. Uhlenbrock, M. Veith, Angew. Chem. 1971, 83, 379; Angew.
Chem. Int. Ed. Engl. 1971, 10, 374.
[17] a) J. Kroner, W. Schneid, N. Wiberg, B. Wrackmeyer, G.
Ziegleder, J. Chem. Soc. Faraday Trans. 2 1978, 74, 1909; b) H.
Bock, K. Wittel, M. Veith, N. Wiberg, J. Am. Chem. Soc. 1976,
98, 109.
[18] Supporting information: a) Synthesis details, b) details of X-ray
structure elucidation, c) kinetics of the decomposition of
diazene, d) IR, Raman, NMR spectra, e) 1H,15N HMBC spec-
tra, and f) computational details, g) applied level theory:
pbe1pbe/aug-cc-pwCVDZ, h) the precipitate could be identi-
fied as a mixture of amalgam Ag11Hg9 along with traces of AgCl
by a Rietveld refinement of the powder diffraction data,
i) diazene 2: The asymmetric unit was split in two parts and
the occupancy of each part was refined freely (0.652(12)/
0.348(12)); j) we observed two resonances at d = 310 (NSi) and
255 ppm (NPh).
Figure 3. Left: Lewis representations of 4+ according to NBO/NRT
analysis. Right: MO40 (p), HOMO (n/s), and LUMO (p*).
DFT calculations of the electronic structure as well as
ELF, MO, and NBO/NRT analyses (ELF = electron local-
ization function,[28] NBO = natural bond orbital, NRT natural
resonance theory)[29] were carried out to gain insight into the
structure and bonding of diazenium ion 4+. MO and NBO
=
calculations show a localized N N bond and a lone pair of
electrons that according to NRTanalysis is mainly localized at
N1 (Figure 3). This is in accordance with the ELF of 0.82 that
exhibits a small dumbbell-shaped region of localized elec-
trons between the two N atoms, which is a good indicator of
a classical double bond, along with a monosynaptic valence
basin for the lone pair of electrons localized at N1 (Figure 2,
right). The calculated natural atomic orbital population
(NAO) net charges are q(N1) = ꢀ0.40 and ꢀ0.59 e (N2),
which means that N1 becomes slightly less negative and N2
even more negative upon formal [Me3Si]+ addition, that is, the
overall charge along the N1-N2 unit remains almost constant
(cf. 2: q(N1) + q(N2) = ꢀ1.10, 3: ꢀ1.02, 4+: ꢀ0.99 e, Table 1),
in accordance with the charge distribution in the adduct 3.
According to TD-DFT calculations,[18] the blue color arises
from the weak n/s!p* HOMO–LUMO electronic transition
at lmax,calcd = 731 nm in cation 4+ (Figure 3, right).[17b]
In summary, a highly labile salt bearing the trisilylated
homoleptic diazenium ion 4+ is presented for the first time.
Although the direct silylation of 2 failed because of side
reactions or low solubility, diazenium ion 4+ was formed in
a straightforward two-electron oxidation process starting
from mercury(II) dihydrazide and Ag[GaCl4]. Structural
and spectroscopic data are in accordance with those of
diazene 2 and the hitherto unknown diazene·GaCl3 adduct 3.
Furthermore, we found that the previous structural data of
diazene 2 needed to be revised on the basis of a new single-
crystal structure determination, thus solving an old problem
in diazene chemistry.
[19] a) M. Lehmann, A. Schulz, A. Villinger, Angew. Chem. 2009,
121, 7580; Angew. Chem. Int. Ed. 2009, 48, 7444; b) A. Schulz, A.
Villinger, Chem. Eur. J. 2010, 16, 7276; c) M. F. Ibad, P. Langer,
F. Reiß, A. Schulz, A. Villinger, J. Am. Chem. Soc. 2012, 134,
17757.
[20] W. Baumann, A. Schulz, A. Villinger, Angew. Chem. 2008, 120,
9672; Angew. Chem. Int. Ed. 2008, 47, 9530.
[21] K. Seppelt, W. Sundermeyer, Chem. Ber. 1970, 103, 3939.
[22] B. Wrackmeyer, C. Kçhler, Heteroat. Chem. 2005, 16, 84.
[23] H. Seidl, H. Bock, N. Wiberg, M. Veith, Angew. Chem. 1970, 82,
42; Angew. Chem. Int. Ed. Engl. 1970, 9, 69.
[24] C. J. Brown, Acta Crystallogr. 1966, 21, 146.
[25] E. I. Davydova, T. N. Sevastianova, A. V. Suvorov, A. Y. Tim-
oshkin, Coord. Chem. Rev. 2010, 254, 2031.
[26] C. J. Carmalt, J. D. Mileham, A. J. P. White, D. J. Williams,
Dalton Trans. 2003, 4255.
[27] C. Hubrich, A. Schulz, A. Villinger, Z. Anorg. Allg. Chem. 2007,
633, 2362.
[28] a) B. Silvi, A. Savin, Nature 1994, 371, 683; b) H. Grꢂtzmacher,
T. F. Fꢀssler, Chem. Eur. J. 2000, 6, 2317.
[29] F. Weinhold, C. Landis, Valency and Bonding, A Natural Bond
Orbital Donor-Acceptor Perspective, Cambridge University
Press, Cambridge, 2005 and references therein.
Received: November 23, 2013
Published online: February 14, 2014
Keywords: diazenes · diazenium ions · oxidation · silylium ions ·
.
structure elucidation
[1] P. Griess, Liebigs Ann. Chem. 1858, 106, 123.
Angew. Chem. Int. Ed. 2014, 53, 3250 –3253
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3253