3,4,5,6-tetrafluorophenylnitren-2-yl: A ground-state quartet triradical

Article abstract of DOI:10.1002/chem.200903285

The photochemistry of 2iodo-3,4,5,6-tetrafluorophenyl azide (7d) has been investigated in argon and neon matrices at 4 K, and the products characterized by IR and EPR spectroscopy. The primary photochemical step is loss of a nitrogen molecule and formation of phenyl nitrene 1d. Further irradiation with UV or visible light results in mixtures of Id with azirine 5d′ , ketenimine 6d′, nitreno radical 2d, and azirinyl radical 9. The relative amounts of these products strongly depend on the matrix and on the irradiation conditions. Nitreno radical 2d with a quartet ground state was characterized by EPR spectroscopy. Electronic structure calculations in combination with the experimental results allow for a detailed understanding of the properties of this unusual new type of organic high-spin molecules.

Full text of DOI:10.1002/chem.200903285

DOI: 10.1002/chem.200903285
3,4,5,6-Tetrafluorophenylnitren-2-yl: A Ground-State Quartet Triradical
Dirk Grote,[a] Christopher Finke,[a] Simone Kossmann,[b] Frank Neese,[b] and
Wolfram Sander*[a]
Abstract: The photochemistry of 2-
iodo-3,4,5,6-tetrafluorophenyl azide
light results in mixtures of 1d with azir-
ine 5d’, ketenimine 6d’, nitreno radical
2d, and azirinyl radical 9. The relative
amounts of these products strongly
depend on the matrix and on the irra-
diation conditions. Nitreno radical 2d
with a quartet ground state was charac-
terized by EPR spectroscopy. Electron-
ic structure calculations in combination
with the experimental results allow for
a detailed understanding of the proper-
ties of this unusual new type of organic
high-spin molecules.
(7d) has been investigated in argon
and neon matrices at 4 K, and the
products characterized by IR and EPR
spectroscopy. The primary photochemi-
cal step is loss of a nitrogen molecule
and formation of phenyl nitrene 1d.
Further irradiation with UV or visible
Keywords: EPR spectroscopy · IR
spectroscopy · matrix isolation ·
photochemistry · radicals
Introduction
cyclohexadienylidenes have robust triplet ground states, a
quartet state for 2a and 4a is energetically favorable, since
in this case local high-spin triplet configurations at the
formal nitrene and carbene centers are maintained.
The introduction of a radical center into phenyl nitrene 1
can lead to three isomeric nitreno radicals 2a, 3a, and 4a
with the radical centers in ortho, meta, and para position, re-
spectively. The electronic structure of these nitreno radicals
is best described as a s,s,p triradical with one unpaired elec-
tron located at the nitrogen atom in the s plane, one at the
radical center of the phenyl ring (also in the s plane), and
the third unpaired electron delocalized over the p system
and interacting with both unpaired s electrons. Density
functional calculations predict for 2a and 4a high-spin quar-
tet (Q) ground states, while for 3a a low-spin doublet (D)
ground state is calculated.[1] For 2a and 4a quinoid reso-
nance structures 2a’ and 4a’ can be formulated, which sug-
gest that these triradicals also have some carbene (cyclohex-
adienylidene) character. In contrast, for 3a no such reso-
nance structure is possible. Since both phenyl nitrenes and
A synthetic route to nitreno radicals starts from iodo-
phenyl azides, which on photolysis in low-temperature ma-
trices split off N2 to give the corresponding phenyl nitrene[2]
and subsequently iodine atoms to give the nitreno radical.[3]
However, phenyl nitrenes are photolabile[4] and easily rear-
range to azirines[2] and ketenimines,[5] and cleavage of the
ꢀ
C I bond to form an additional radical center competes
[a] Dr. D. Grote, C. Finke, Prof. Dr. W. Sander
Lehrstuhl fꢀr Organische Chemie II
Ruhr-Universitꢁt Bochum
with these rearrangements. Thus, photolysis of matrix-isolat-
ed (argon, 10 K) phenyl azide 7a produces mixtures of
phenyl nitrene 1, azirine 5a, and ketenimine 6a in photosta-
44780 Bochum (Germany)
Fax : (+49)234-321-4353
E-mail: wolfram.sander@rub.de
tionary equilibria, and the yield of nitrene
1 is low
(Scheme 1).[6] Fluorine substituents in the ortho positions of
1 decrease the tendency of these rearrangements,[7] and
therefore ortho-fluorinated phenyl nitrenes are obtained in
higher yields under the same conditions.[5–6,8]
[b] S. Kossmann, Prof. Dr. F. Neese
Lehrstuhl fꢀr Theoretische Chemie
Universitꢁt Bonn, 53115 Bonn (Germany)
Supporting information for this article (geometries, total energies,
and IR spectroscopic data of all photoproducts) is available on the
WWW under http://dx.doi.org/10.1002/chem.200903285.
The only nitreno radicals 4 that could be matrix-isolated
and spectroscopically characterized are therefore those bear-
4496
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 4496 – 4506

FULL PAPER
meta nitreno radical 3c was not found, but a product of its
ring opening and re-addition of iodine, namely, allene 8, was
observed. The radical center in meta position of the phenyl
nitrene results in a 1,4-diradicaloid structure that has a ten-
dency for b-cleavage. The cleavage is even more pro-
nounced when two radical centers are formed in the two
meta positions.
Here we describe the photochemistry of 2-iodo-3,4,5,6-tet-
rafluorophenyl azide (7d), a precursor of 3,4,5,6-tetrafluoro-
phenylnitren-2-yl (2d).
Results and Discussion
Scheme 1. Photochemistry of phenyl azide 7a.
Matrix IR spectroscopy: UV irradiation (254 nm) of matrix-
isolated (argon or neon at 3 K) azide 7d results in the rapid
decrease of all IR absorptions assigned to the azide and for-
mation of strong IR bands at 997.3, 1412.8, and 1480.2 cmꢀ1
assigned to triplet phenyl nitrene 1d. The experimental IR
spectrum of 1d is in excellent agreement with calculations
ing ortho-fluoro substituents.[3a,c] UV photolysis of azide 7b
produces nitrene 1b together with ketenimine 6b and nitre-
no radical 4b (Scheme 2). Although the yield of 4b is low, it
could be characterized by IR und EPR spectroscopy, which
also confirmed its high-spin quartet ground state.
at the B3LYP/6-311GACTHNUTRGNE(NUG d,p) level of theory (Figure 1). Pro-
longed UV irradiation produces ketenimine 6d’ with the
characteristic broad C=C=N stretching vibration[9] at
1878.1 cmꢀ1. The IR spectra calculated for isomeric keteni-
mines 6d and 6d’ are very similar. The major difference in
the calculated spectra is that the most intense band appears
at 1384 cmꢀ1 in 6d and at 1328 cmꢀ1 in 6d’. Careful compari-
son of the experimental and calculated IR spectra suggests
that 6d’ is the major photoproduct of 1d, and its isomer 6d
is formed as a minor constituent. However, since many
bands of 6d and 6d’ overlap and only the strongest band of
6d can be assigned, the identification of 6d is only tentative.
During the photolysis of 7d with UV light (254 or
320 nm) another set of IR bands at 1456.7, 1080.5, 1018.0,
956.4, and 775.9 cmꢀ1 is formed which does not belong to ni-
trene 1d, azirines 5d, or ketenimines 6d. This new species is
selectively produced by prolonged irradiation at 320 nm
(Figure 2), and decreases during annealing of the matrix
from 3 to 8 K in neon or to 30 K in argon. A thermal reac-
tion during annealing a matrix is typical of a radical pair
that recombines when the matrix becomes soft enough to
allow diffusion. By comparison
Scheme 2. Photochemistry of aryl azide 7b.
In the photolysis mixture of azide 7c, nitrene 1c, and its
rearranged products, the isomeric azirines 5c and 5c’ and
ketenimines 6c and 6c’, were identified (Scheme 3).[3b] The
with DFT calculations, the new
species is assigned to azirinyl
radical 9 (Table 1). Radical 9
could be formed either by rear-
rangement of nitreno radical 2d
or by loss of an iodine atom
from azirine 5d (Scheme 4). Ni-
treno radical 2d can be ob-
served in low concentrations by
EPR spectroscopy (see below)
but not by the less sensitive and
less selective IR spectroscopy.
Irradiation of an argon
matrix containing ketenimines
Scheme 3. Photochemistry of aryl azide 7c.
6d and 6d’ and azirinyl radical
Chem. Eur. J. 2010, 16, 4496 – 4506
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
4497

W. Sander et al.
Table 1. Spectroscopic data of azirinyl radical 9.
[a,b]
[b,c]
~
nexptl
~
ncalcd
Mode
Symmetry
Irel,exptl
Irel,calcd.
[a]
[c]
A
]
A
]
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
A2
B1
A1
A2
B2
B2
A1
B2
A1
B2
A1
A1
B2
A1
B2
A1
567.3
585.3
626.8
655.8
675.9
789.1
957.4
1033.2
1095.9
1258.4
1342.9
1436.4
1501.7
1502.3
1514.8
1810.2
0.00
0.00
0.01
0.00
0.00
0.06
0.06
0.27
0.11
0.01
0.00
0.20
1.00[d]
0.32
0.05
0.02
775.9
956.4
1018.0
1080.5
0.14
0.09
0.38
0.28
1456.7
1.00
Figure 1. IR spectra showing the 254 nm photochemistry of 1d in argon
at 3.5 K after 10 min irradiation. a) Calculated IR spectrum of triplet ni-
trene 1d. b) Difference IR spectrum: Bands pointing downward disap-
pear during irradiation and belong to 1d; bands pointing upward appear
and are assigned to 6d’. Some additional bands can be assigned to keteni-
mine 6d, but since these bands are relatively weak, this assignment is ten-
tative. Bands marked with ꢃ appear during photolysis and can be as-
signed to the azirinyl radical 9 (Figure 2). c) Calculated IR spectrum of
ketenimine 6d’, d) calculated spectrum of ketenimine 6d. All calculations
[a] Ar matrix, 3 K. [b] Relative intensity based on the strongest absorp-
tion. [c] B3LYP/6-311G(d,p), unscaled. [d] Calculated absolute intensity
of the strongest absorption: 598.6 kmmolꢀ1
AHCTUNGTRENNUNG
.
reveals that the isomeric azirine 5d is also formed, but in
much lower yields. The calculated IR spectra of 5d and 5d’
are more different than those of 6d and 6d’, and this makes
the assignment of 5d more reliable than that of 6d. The
photochemical selectivity leads to 6d’ as major isomer upon
irradiation with UV light and to 5d’ upon irradiation with
visible light.
at the (U)B3LYP/6-311GACTHNUTRGNEUGN(d,p) level of theory.
Matrix EPR spectroscopy: To obtain EPR spectra, an argon
matrix doped with small amounts of aryl azide 7d was de-
posited on top of a copper rod at 4 K. The copper rod was
then positioned inside a quartz tube (within the high-
vacuum system) to allow irradiation with UV light and re-
cording of EPR spectra of the products formed upon irradi-
ation.
Irradiation of matrix-isolated aryl azide 7d with an XeCl
excimer laser (308 nm) results in the formation of several
EPR signals (Figure 4). The most intense signal is found
around g=2, characteristic of doublet radicals. The width of
this signal indicates that a mixture of several radical species
is formed, and hyperfine couplings (hfc) could therefore not
be resolved.
Additional weak signals around 7000 G are assigned to
the x2, y2, and z1 transitions of a randomly oriented triplet
species, and thus indicate formation of nitrene 1d. A simula-
tion with the parameters g=2.003, jD/hcj=0.982 cmꢀ1 and
jE/hcj=0.0071 cmꢀ1 nicely reproduces the experimental
spectrum. The size of the zero field splitting (ZFS) parame-
ter D is characteristic of triplet aryl nitrenes and resembles
the D value of 4-iodo-2,3,5,6-tetrafluorphenylnitrene (1b),[3c]
which has jD/hcj=1.108 cmꢀ1. The E value of 1d, though
significantly smaller than that of 1b (jE/hcj=0.012 cmꢀ1), is
still larger than that of most aryl nitrenes.[10] In our discus-
sion of the ZFS parameters of 1b, we argued that D and E
values are significantly influenced by spin–orbit coupling
(SOC) effects due to the heavy iodine substituent attached
Figure 2. IR spectra showing the photochemistry of azide 7d and nitrene
1d after 110 min irradiation with UV light (l=320 nm) in argon at 3.5 K.
a) Calculated IR spectrum of 1d. b) Calculated spectrum of 7d. c) Differ-
ence IR spectrum: Bands pointing downward disappear during irradia-
tion and belong to 1d and 7d; bands pointing upward appear and are as-
signed to 9 and are marked with ꢃ. These bands disappear upon anneal-
ing of the matrix. d) Calculated spectrum of azirinyl radical 9. All calcu-
lations are at the (U)B3LYP/6-311GACTHNUTRGNEUNG(d,p) level of theory.
9 at 4 K with visible light (l=420 nm) results in a decrease
of all bands assigned to these species and the rapid forma-
tion of a new species with intense bands at 1521.4, 1379.6,
1365, and 970.5 cmꢀ1. By comparison of these IR bands with
DFT calculations at the B3LYP/6-311GACTHNUTRGNEUNG(d,p) level of theory,
the new species is assigned to the bicyclic azirine 5d’
(Figure 3). A careful analysis of the experimental spectrum
4498
www.chemeurj.org
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 4496 – 4506

3,4,5,6-Tetrafluorophenylnitren-2-yl
FULL PAPER
to the aromatic ring.[3c] The
same conclusion can be drawn
to explain the relatively large E
value of 1d. The noticeably
weak signal intensity of 1d sug-
gests that the nitrene reacts to
give further products even
under the conditions of matrix
isolation.
EPR spectrum of nitreno radi-
cal 2d: In addition to the EPR
transitions discussed above, sev-
eral very weak signals at 1670,
1890, 5910, 6280, and 7300 G
Scheme 4. Photochemistry of aryl azide 7d.
can be identified. These signals
disappear upon annealing of
the matrix at 30 K. Similar behavior was also observed for
4b[3a,c] and is characteristic of radical pairs. In accordance
with the results from the simulation of the EPR spectrum
(see below), the weak EPR signals are assigned to nitreno
radical 2d in its quartet ground state (Figure 5). The iodine
atoms formed during photolysis could not be observed in
the EPR spectra. This was also found in other experiments
and is attributed to the orbital degeneracy and strong spin–
spin coupling in the iodine atoms, which result in very broad
signals. The EPR spectrum of matrix-isolated iodine atoms
could only be observed in solid xenon and other matrices
with strong interactions between the iodine atoms and the
matrix.[11]
Figure 3. IR spectra showing the 420 nm photochemistry of ketenimine
6d’ after 30 min irradiation in argon at 3.5 K. a) Calculated IR spectrum
of ketenimine 6d’. b) Difference IR spectrum: Bands pointing downward
disappear during irradiation and belong to 6d’; bands pointing upward
appear and are assigned to 5d and 5d’. c) Calculated IR spectrum of azir-
ine 5d. d) Calculated IR spectrum of azirine 5d’. All calculations at the
B3LYP/6-311GACHTUNGTRENNUNG(d,p) level of theory.
Figure 5. a) Experimental EPR spectrum showing the weak signals of
quartet nitreno radical 2d. b) Simulation of a quartet system (S=3/2)
using the ZFS parameters jD/hcj=0.357 cmꢀ1 and jE/hcj=0.0136 cmꢀ1
(g=2.003, n=9.5904 GHz). The experimental spectrum can nicely be re-
produced by the simulation.
Figure 4. EPR spectrum of 1d upon irradiation of 7d in solid argon at
4 K with UV light of an excimer laser (XeCl, 308 nm). At 7000 G the
transitions z1, x2, and y2 of a weak triplet signal can be identified. The
triplet spectrum can nicely be reproduced by simulation leading to zfs pa-
rameters jD/hcj=0.982 cmꢀ1 and jE/hcj=0.0071 cmꢀ1 (g=2.003, n=
9.5904 GHz).
Simulation of the EPR spectrum: The experimental EPR
spectrum of 2d could be simulated by assuming a quartet
state with the ZFS parameters jD/hcj=0.357 cmꢀ1 and jE/
hcj=0.0136 cmꢀ1 (g=2.003). Assignment of the transitions
(Table 2) is based on the simulated spectrum. The field de-
Chem. Eur. J. 2010, 16, 4496 – 4506
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
4499

W. Sander et al.
Table 2. Assignment of the EPR transitions of 2d.
weak signals in the simulated spectrum at 1135, 2130 and
4120 G can be assigned to transitions with the external mag-
netic field oriented parallel to the z axis. These signals
cannot be seen in the experimental spectrum, though. How-
ever, the calculated x and y transitions can be identified in
the simulated and experimental spectra.
The transitions z4 can be described as forbidden transi-
tions between the levels j2> and j4> and as off-axis transi-
tions. This transition becomes allowed if the external mag-
netic field is rotated only a few degrees away from the z
axis (Figure 7). The nature of the z4 transition was previous-
ly discussed in more detail.[3c] The signals at 5910 and
6280 G in the experimental spectrum cannot be attributed
to axial transitions. They appear at angles of about 258 in
the xz plane and 308 in the yz plane. These extra lines A
correspond to higher-order solutions of the spin Hamiltoni-
an and are characteristic of high-spin systems (S>1) with
large ZFS parameters.[13]
Signal
Transition
Assignment
simulation
calculation
experiment
1135
1670
1890
2130
3505
4120
5915
6280
6380
7300
1135
1670
1890
2115
3490
4110
–
–
6360
7260
–
z, j1>,j2>
y, j3>,j4>
x, j3>,j4>
z, j2>,j4>
z, j3>,j4>
z, j3>,j4>
–
z1
y1
x1
z4
z3’
z3’’
A
A
x3
y3
1675
1890
–
–
–
5910
6280
6365
7295
–
x, j1>,j2>
y, j1>,j2>
pendence of the magnetic energy levels was calculated for
the direction of the external magnetic field parallel to the
principal axes x, y, and z, by using the solution of the Hamil-
ton matrix of a quartet state[12] with the parameters derived
from the simulation (Figure 6). The analysis reveals that the
Figure 7. Angular dependence of the EPR signals of 2d when varying the
field orientation from parallel to x to parallel to y. The signals marked as
“A” can be assigned to off-axis transitions. The plot also shows the off-
axis behavior of transition z4, which becomes allowed when the magnetic
field is rotated in the zy plane by a few degrees.
Photochemistry of nitrene 1d: Since the work of Dunkin
and Thomson on the photochemistry of pentafluorophenyl-
nitrene in cryogenic matrices in 1982,[14] the rearrangements
of several phenyl nitrenes bearing fluorine atoms were in-
vestigated both by experimental and theoretical meth-
ods.[2,5,7c,8,15] It was found that fluorine atoms in ortho posi-
tion of the aromatic ring increase the stability of phenyl ni-
trene by steric and electronic effects.[7c]
Figure 6. Energy of the four quartet sublevels as a function of the exter-
nal field H parallel to z, x, and y, respectively, calculated for jD/hcj=
0.357 cmꢀ1, jE/hcj=0.0136 cmꢀ1 (g=2.003). The z1 transition at low field
is reproduced as a very weak band in the simulated spectrum.
4500
www.chemeurj.org
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 4496 – 4506

3,4,5,6-Tetrafluorophenylnitren-2-yl
FULL PAPER
A careful analysis of the photochemistry of nitrene 1d re-
veals high selectivity for the formation of keteneimine 6d’
on UV irradiation and for azirine 5d’ on visible-light irradia-
tion. In a recent study on the selectivity of rearrangements
of fluorinated 4-iodophenylnitrenes we found that visible-
light irradiation yields the azirines with the lower activation
barrier of the thermal singlet nitrene rearrangement, even if
they are thermodynamically less stable than their isomers.
In contrast, UV irradiation produces the thermodynamically
most stable ketenimine.[16] An example for this is the photo-
chemistry of nitrene 1e, where azirine 5e is formed upon
visible-light irradiation although it is thermodynamically less
stable than 5e’. UV irradiation produces the thermodynami-
cally more stable ketenimine 6e’ as the major product.[16]
product distribution can be explained by the kinetic and
thermodynamic preferences.
Azirinyl radical 9: As expected, the primary photoproduct
of the photolysis of azide 7d is nitrene 1d, which is formed
in significantly lower yield than the 4-iodo-2,3,5,6-tetrafluor-
phenylnitrene in our previous investigation.[3a,c] Nitrene 1d
was characterized by EPR and IR spectroscopy. Since split-
ꢀ
ting of the C I bond in 1d to produce the nitreno radical
2d is inefficient and competes with the rearrangement of 1d
to azirines and ketenimines, the intensity of the quartet
EPR signals of 2d is very low. With the less sensitive IR
spectroscopy, nitreno radical 2d could not be observed. In-
stead, azirino radical 9, which is formed during further irra-
diation of the matrix, was identified by IR spectroscopy.
This observation is supported by DFT calculations [B3LYP/
The calculated [(U)B3LYP/6-311GACHTUNTRGNEUNG(d,p)] barrier for for-
mation of azirine 5d’ from singlet nitrene S-1d is about
3.2 kcalmolꢀ1 lower than that of 5d (Figure 8). In this case
azirine 5d’ is also thermodynamically more stable than 5d.
The major product of the UV photochemistry, ketenimine
6d’, is 8 kcalmolꢀ1 more stable than its isomer 6d. Thus, al-
though the rearrangements are driven photochemically, the
6-311GACHTUNGTRNEUNG
(d,p)], which predict radical 9 to be 18.5 kcalmolꢀ1
lower in energy than nitreno radical 2d. The EPR spectra
also show relatively high yields of doublet radicals in the
spectral area around g=2 (Figure 4), as expected if radical 9
is formed.
Quantum chemical calculations reveal that the azirinyl
radical 9 has C2v symmetry, and
the spin density is delocalized
on
the
aromatic
system
(Figure 9). With a spin popula-
tion at the nitrogen atom of 0.6
but only 0.02 at the adjacent
carbon atoms, the nitrogen
atom bears the major part of
the spin density.
Azirinyl radical
9 is most
likely formed from nitreno radi-
cal 2d in an excited doublet
state by ring closure between
the nitrogen atom and the radi-
cal center in ortho position.
This mechanism is suggested by
the fact that the yield of 2d in
the matrix is very low and that
9 is formed under the same
conditions as the nitreno radi-
cal. A second possible mecha-
nism for the formation of 9 is
ꢀ
cleavage of the C I bond in bi-
cyclic azirine 5d. The major
part of the spin density then
moves from the carbon to the
nitrogen center, changing the
geometry to planar 9 with C2v
symmetry as the minimum
structure (Scheme 5). Upon an-
nealing of the matrix the iodine
atom recombines with 9, ob-
servable by a decrease of its IR
bands, but only the absorptions
of ketenimine 6d’ increase
Figure 8. a) Relative energies of the rearrangement products and transition states of 2,5-difluoro-4-iodophenyl-
nitrene (1e)[16] and b) of 1d [(U)B3LYP/6-311G
ACTHNUTRGNEUNG(d,p)]. Azirine 5e is the major product upon irradiation of the
matrix of 1e with visible light, while ketenimine 6e’ is the major isomer upon irradiation with UV light. The
rearrangement to the azirine is apparently kinetically controlled, while the rearrangement to the ketenimine is
thermodynamically controlled. The major azirine 5d’ formed upon irradiation of 1d is both the thermodynami-
cally and kinetically preferred azirine.
Chem. Eur. J. 2010, 16, 4496 – 4506
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
4501

W. Sander et al.
Figure 11. Three singly occupied molecular orbitals of 2d: spin-unrestrict-
ed natural orbitals (UNOs) in a localized representation; Pipek–Mezey[17]
localization; B3LYP/TZVPP.
Figure 9. Spin-density distribution of azirinyl radical 9 [UB3LYP/6-311G-
(d,p)]. 9 shows C2v symmetry, the spin density is delocalized into the aro-
ACHTUNGTRENNUNG
matic p system.
tion of 2d as a s,s,p triradical. The unpaired electron in the
in-plane nitrogen MO is well localized and only slightly po-
larizes the s system of the ring, whereas the unpaired elec-
tron located in the out-of-plane nitrogen MO is heavily de-
localized into the p system of the ring. Similarly, the carbon
s lone pair is fairly delocalized into the s system of the ring.
To determine the extent to which 2d has an isolated quar-
tet ground state, SORCI calculations were performed on
top of a state-averaged complete active space self-consistent
field (SA-CASSCF) wavefunction with three electrons in
three orbitals. This procedure has already proven to provide
essentially converged results for state-energy differences[18]
and has also been applied in studying 4b.[3c] Quasi-restricted
orbitals[19] from a BP86/TZVPP calculation served as start-
ing orbitals.
Scheme 5. Possible rearrangement to form 9.
slightly under these conditions. The changes in intensity are
very small, though.
Spectroscopic calculations: The quartet ground state of 2d,
calculated with the B3LYP functional and the TZVPP basis
set, shows
a fairly complex spin-density distribution
(Figure 10) involving s and p components. The spin-density
distribution is very similar to that observed in 2,3,5,6-tetra-
fluorophenylnitren-4-yl (4b).[3c]
The active orbitals of the 2d molecule transform into the
Cs point group under a’ (in-plane nitrogen MO and carbon
lone pair), and a’’ (out-of-plane nitrogen MO) irreducible
representations. Thus, the lowest quartet state of the system
is designated as 1-4A’. The SORCI calculation predicts
1-4A’’ to be the lowest state, followed by the first doublet
(1-2A’’) at about 0.31 eV (7 kcalmolꢀ1) and the second dou-
blet (2-2A’’) at about 0.81 eV (19 kcalmolꢀ1). The dominat-
ing electron configuration for both low-lying doublet states
ACTHNUTRGNEUNG
is in agreement with 4b the (a’)1(a’)1(a’’)1 configuration,
which corresponds to the two linearly independent spin-dou-
blet couplings.
The three SOMOs form a quasidegenerate set of MOs,
with differences in orbital energies of less than 2.3 eV. Fur-
thermore, since the SOMOs are spatially not well separated,
the exchange interactions between the unpaired electrons
must be large. Thus, from both the theoretical and experi-
mental results, it can be concluded that 2d has a well-isolat-
ed spin-quartet ground state.
Figure 10. Spin-density distribution of 2d. Positive values are contoured
in green, negative values in blue, both at the level of 0.003 eꢀ Bohrꢀ3
.
Calculations of the EPR parameters: Experimental and cal-
culated EPR properties of 2d are compared in Table 3. The
agreement between experiment and theory is reasonable. As
expected, the dominant contribution to the ZFS tensor is
the spin–spin (SS) interaction, which amounts to about 92%
of the total D value. The calculated SS contribution is essen-
tially local, with the largest contributions resulting from
one-center integrals. This was also found in the analysis of
The three singly occupied molecular orbitals (SOMOs) of
2d were identified by examining the exactly singly occupied
spin-unrestricted orbitals (UNOs) transformed into a local-
ized presentation (Figure 11). The conclusions that can be
drawn from inspection of the SOMOs confirm the descrip-
4502
www.chemeurj.org
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 4496 – 4506

3,4,5,6-Tetrafluorophenylnitren-2-yl
FULL PAPER
Table 3. Analysis of the calculated D tensor (B3LYP/TZVPP) of 2d.
D [cmꢀ1
]
E [cmꢀ1
]
total (calcd)
experimental
spin–spin
spin–orbit
spin–spin
0.284
0.357
0.262
0.022
0.023
0.014
0.022
0.001
1-center
2-center
3-center
4-center
Coulomb
exchange
0.315
ꢀ0.006
0.034
ꢀ0.047
ꢀ0.006
0.000
0.172
0.091
ꢀ0.005
0.000
0.025
ꢀ0.003
spin–orbit
M=0 (a!a)
M=0 (b!b)
M=+1 (b!a)
M=ꢀ1 (a!b)
0.006
0.010
0.021
ꢀ0.014
ꢀ0.003
ꢀ0.003
0.004
0.002
the ZFS contributions of the 2,3,5,6-tetrafluorophenylnitren-
4-yl radical (4b).[20]
The spin–orbit coupling (SOC) contribution to D is about
8%, which is slightly smaller than the corresponding value
in 4b. The SOC contribution to the D tensor is dominated
by spin-flip contributions. Unlike in 4b, the two spin-lower-
ing and spin-raising spin-flip contributions are of opposite
sign and play an important role.
The magnetic zy plane of the D tensor is located in the
molecular plane of 2d, while the x axis points out of the
plane. An analogous result was found for 4b. However,
ꢀ
while in 4b the easy axis is oriented along the C N bond, it
is rotated about 18 away from the axis in 2d, and the origin
of the D tensor is shifted from the molecular axis defined by
Figure 12. a) Orientation of the D tensor, b) orientation of the g tensor of
2d based on the B3LYP/TZVPP calculations.
ꢀ
the C N bond (Figure 12). Because the nitreno radical si-
multaneously exhibits nitrene and carbene character, the
large E value of 4b was interpreted by geometric considera-
tions. The dipolar field of the nitrene contribution in 4b
prevails over the carbene character, but unlike in 4b there is
ꢀ
no symmetry axis along the C N bond in 2b, and thus the
ꢀ
points along the C N bond, while the carbene contribution
easy axis of the spin system can be shifted by the influence
of the dipolar field of the carbene. Since the D value is
heavily influenced by the spin density at the spin carrying
centers, the higher D value of 2d can be explained by larger
spin density on its nitrene center compared to 4b. The ni-
trene centers in 2d and 4b bear spin populations of 1.598
and 1.577, respectively, calculated at the UB3LYP/6-311G-
points parallel to a hypothetical bond angle of 1808 of the
carbene moiety, and thus its dipolar field is oriented along
the y direction of the nitrene contribution. Thus, the nitrene
and carbene contributions are perpendicular to each other,
and since the nitrene character prevails over the carbene
character, the easy axis of the quartet D tensor is still orient-
ꢀ
ed parallel to the C N bond, and the carbene contribution
N
to the dipolar field points along the y direction, which re-
sults in significantly magnetically inequivalent y and x direc-
tions. The large E value can therefore be interpreted as re-
sulting from the carbene character of the 1-4A2 ground state.
The lower E value of 2d can then be explained from anal-
ogous geometric considerations. Since the carbene center in
the aromatic ring is located in ortho position to the nitrene
center, the dipolar field of the carbene contribution is not
located perpendicular to the magnetic z axis of the nitrene
center, and this results in a considerable contribution paral-
lel to the z direction of the nitrene dipolar field and there-
fore to a smaller E value of the quartet spin system com-
pared to 4b. Like for 4b, the nitrene character in 2d still
Chem. Eur. J. 2010, 16, 4496 – 4506
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
4503

W. Sander et al.
yield of these species depends on the irradiation conditions.
Of the two regioisomeric azirines 5d and 5d’ and keteni-
mines 6d and 6d’, in both cases the more stable isomers 5d’
and 6d’ are formed as major products.
Experimental Section
EPR measurements: X-band EPR spectra were recorded with a Bruker
Elexsys E500 EPR spectrometer with an ER077R magnet (75 mm gap
between pole faces), an ER047 XG-T microwave bridge, and an
ER4102ST resonator with a TE102 cavity. The matrices were deposited on
an oxygen-free high-conductivity copper rod (75 mm length, 3 mm diam-
eter) cooled by a Sumitomo SHI-4-5 closed-cycle 4.2 K cryostat. Since
saturation of the spectra might be a problem, especially at 4 K, most
Further photolysis of these product mixtures yields two
unusual products: nitreno radical 2d with a high-spin quar-
tet ground state and azirinyl radical 9. The yield of 2d is
quite low, and it could only be detected by the sensitive and
selective EPR spectroscopy. The EPR spectrum of 2d could
be simulated with the ZFS parameters jD/hcj=0.357 cmꢀ1
and jE/hcj=0.0136 cmꢀ1. The D value of 2d is slightly
larger than that of para nitreno radicals such as 4b, which
are in the range between 0.278 and 0.291 cmꢀ1, while the E
value is considerably smaller than that of the para nitreno
radicals, which lie between 0.040 and 0.043 cmꢀ1.[3c]
The classical interpretation of the ZFS parameters in
terms of spin–spin interactions is that D correlates with the
distance between the unpaired electrons and E describes de-
viations from cylindrical symmetry. This simplified interpre-
tation does not take into account spin–orbit contributions,
which indeed contribute less than 10% to the ZFS parame-
ters. The higher D value of 2d compared to 4d correlates
with a higher spin density at the nitrogen atom in 2d. The
higher E value in the more symmetrical para nitreno radical
4b compared to 2d is counterintuitive, but is in accordance
with our model of describing nitreno radicals as unifying
properties of both nitrenes and carbenes. In 4b the dipolar
field contribution of the carbene is perpendicular to that of
the nitrene moiety, whereas in 2d there is a much smaller
angle between the dipolar field vectors of the carbene and
nitrene units. Since the spin–spin interactions in nitrenes are
larger than in carbenes (the D values of nitrenes are much
larger than those of carbenes), the nitrene structure in 4b
determines the magnetic z axis, but the carbene structure re-
sults in a large contribution in the y direction and thus a
large E value. In 2d the symmetry is lower and the magnetic
axes are not restricted to the Cartesian axes as in C2v-sym-
metrical 4b. Thus, the less symmetrical 2d has a significantly
smaller E value than 4b. This description is in accordance
with the electronic structure of a s,s,p triradical.
spectra were recorded with
a relatively high microwave power of
20.1 mW at 4 K and in addition with lower microwave power. The spectra
were identical, and only the signal-to-noise ratio was much worse at
lower power. We therefore assume that saturation is not significant under
the experimental conditions used in our experiments.
The vacuum system consisted of a vacuum shroud equipped with a
sample inlet valve and a half-closed quartz tube (75 mm length, 10 mm
diameter) at the bottom and a vacuum pump system with a Pfeiffer
Vacuum TMU071P turbo pump backed by a Leybold two-stage, rotary-
vane pump. To avoid contamination of the high-vacuum segment by
pump oil from the backing pump, a catalytic oxidation filter was placed
between the rotary-vane pump and the turbo pump. During deposition,
the inlet port was positioned at the same height as the tip of the copper
rod. For irradiation, the copper rod was lowered into the quartz tube at
the bottom of the shroud, and for the measurement of the EPR spectra
the whole apparatus was moved downwards so that the quartz tube and
copper rod were positioned inside the EPR cavity.
Azide 7d was evaporated for 1 h at 08C and co-deposited with a large
excess of argon (Messer-Griesheim, 99.9999%) on the tip of the copper
rod at 4 K. The matrix-isolated sample was subsequently irradiated with
a Lambda Physik Lextra 200 Excimer Laser (XeCl, 308 nm), and spectra
were recorded at various irradiation times.
The computer simulation of the EPR spectrum was performed by using
the XSophe computer simulation software suite (version 1.0.4),[23] devel-
oped by the Centre for Magnetic Resonance and Department of Mathe-
matics, University of Queensland, Brisbane (Australia) and Bruker Ana-
lytik GmbH, Rheinstetten (Germany). The angular dependence of the
quartet transitions was calculated with EasySpin.[24]
Matrix IR measurements: All IR spectra were recorded with a Bruker
Ifs 66s spectrometer. For matrix isolation, Ar and Ne gases produced by
Messer Griessheim with a purity of 99.999% were used. For preparing
the matrices, the precursor 7d was condensed simultaneously with an
excess of the inert gas onto a CsI window cooled at 4 K by a helium
closed-cycle cryostat produced by Sumitomo. The deposition time was
between 60 and 70 min, while the precursor was evaporated at a temper-
ature of ꢀ3 to ꢀ58C. The matrix-isolated sample was subsequently
broadband-irradiated with different high-pressure mercury lamps pro-
duced by Ushio and equipped with quartz optics produced by L.O.T.
Oriel, which has an output of 500 W. The wavelength ranges were adjust-
ed by a combination of different dichroic reflectors produced by Oriel
and different range filters produced by Schott & Balzers. For an exact
wavelength of 254 nm a low-pressure mercury lamp developed and pro-
duced by Graentzel was used. Also the Excimer Compex 100 and Exci-
mer Compex 110 lasers produced by Lambda Physics were used for fast,
Azirinyl radical 9 is a novel type of radical. For the parent
azirinyl radical the heat of formation was estimated from
mass spectrometric data to 81 kcalmolꢀ1,[21] which is consid-
erably less than that estimated from ab initio studies.[22] Ac-
cording to these calculations the azirinyl radical shows Cs
pulsed irradiations of the systems at
a
wavelength of 248 nm
ꢀ
ꢀ
symmetry with a formal C N double bond and C N single
bond. Annelation of a benzene ring in 9 results in a C2v-sym-
(250 mJpulseꢀ1
10 pulsesꢀ1).
,
1 pulsesꢀ1
)
and at 308 nm (60–70 mJpulseꢀ1
,
1–
ꢀ
metrical radical with a short C C bond, similar to benzocy-
The computer simulation of the FTIR spectra was performed by using
Gaussian 03 and Gaussian 03 for Windows on PCs and workstations.[25]
In the calculations the B3LYP or the UB3LYP hybrid functional com-
clopropene, with most of the spin density localized at the ni-
trogen atom. The radical 9 is calculated to be 18.5 kcalmolꢀ1
more stable than nitreno radical 2d. However, since 2d has
a quartet and 9 a doublet ground state, direct cyclization is
spin-forbidden and requires thermal or photochemical exci-
tation of 2d to an excited doublet state. Thus, 2d is only
metastable with respect to 9, but kinetically stabilized by the
mismatch of the spin states of 2d and 9.
bined with 6-311GACTHNUTRGENNGU(d,p) and 6-311+GACHTUNGTRNE(NUGN d,p) basis sets was used, respec-
tively. Besides the FTIR spectra, quantum mechanical calculations of the
zero-point energies allowed comparison between the formed systems in
reality and in theory.
EPR calculations: All calculations were carried out with a development
version of the ORCA[26] program package. The investigated systems were
optimized by employing the BP86[27] GGA functional together with the
resolution of identity (RI)[28] technique. The basis set was of polarized
4504
www.chemeurj.org
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 4496 – 4506

3,4,5,6-Tetrafluorophenylnitren-2-yl
FULL PAPER
triple-z quality (TZVP[29]). Structures were verified to represent minima
through numerical frequency analysis.
kler, B. Cakir, D. Grote, H. F. Bettinger, J. Org. Chem. 2007, 72,
715–724; c) W. Sander, D. Grote, S. Kossmann, F. Neese, J. Am.
Chem. Soc. 2008, 130, 4396–4403.
The EPR parameters were calculated with the B3LYP[27a,30] hybrid func-
tional and a more extensively polarized triple-z basis set (TZVPP[29]).
[4] E. Leyva, M. S. Platz, G. Persy, J. Wirz, J. Am. Chem. Soc. 1986, 108,
3783–3790.
[5] C. Carra, R. Nussbaum, T. Bally, ChemPhysChem 2006, 7, 1268–
1275.
[6] N. P. Gritsan, M. S. Platz, Chem. Rev. 2006, 106, 3844–3867.
[7] a) R. Poe, K. Schnapp, M. J. T. Young, J. Grayzar, M. S. Platz, J. Am.
Chem. Soc. 1992, 114, 5054–5067; b) K. A. Schnapp, M. S. Platz,
Bioconjugate Chem. 1993, 4, 178–183; c) W. L. Karney, W. T.
Borden, J. Am. Chem. Soc. 1997, 119, 3347–3350.
[8] N. P. Gritsan, A. D. Gudmundsdottir, D. Tigelaar, Z. Zhu, W. L.
Karney, C. M. Hadad, M. S. Platz, J. Am. Chem. Soc. 2001, 123,
1951–1962.
The spin–spin (SS) contribution to the zero-field splitting[31] was treated
in the mean-field approximation by employing spin densities from the
spin-unrestricted natural orbital (UNO) determinant. The spin–orbit cou-
pling contributions to the zero-field splitting and to the g tensor[32] were
treated by a linear response approach, and the spin–orbit mean-field ap-
proach was used to represent the SOC operator.[33]
Additionally, correlated multireference ab initio calculations in the form
of the spectroscopy-oriented configuration interaction (SORCI) ap-
proach[34] were performed. The cutoffs were chosen to be Tsel =10ꢀ6 Eh,
T
pre =10ꢀ4, and Tnat =10ꢀ5
.
2-Iodo-3,4,5,6-tetrafluoroaniline: 2-Iodo-3,4,5,6-tetrafluoroaniline was
synthesized in analogy to a literature procedure.[35] A solution of ICl
(835 mg, 0.007 mol) in acetic acid (30 mL) was added slowly to a solution
of 3,4,5,6-tetrafluoroaniline (1 g, 0.006 mol) in acetic acid 25 mL. To this
rufous solution H2O (100 mL) was added, and the solution was heated to
reflux (808C). After 3 h the solution was cooled to room temperature
and an aqueous solution of NaOH was added until a pH value of 3 was
reached. By extraction with ethyl acetate, all organic components were
isolated, and by extraction of the organic phase with saturated solutions
of Na2S2O3 and NaCl, all water-soluble relics were removed. The organic
solution was dried over MgSO4 and evaporated completely. Column
chromatography (hexane/methyl tert-butyl ether (MTBE) 4/1) gave the
product (1.93 g, 77%). Rf =0.45 (hexane/MTBE 4/1); 1H NMR
(200 MHz, CDCl3, 258C, TMS): d=3.29 ppm (s, NH2); 13C NMR
(200 MHz, CDCl3, 258C, TMS): d=131.1 (C2), 136.9 (C6), 139.7 (C4),
142.3 (C5), 145.9 (C3), 148.3 ppm (C1); MS (EI): m/z (%): 291 [M+], 272
[M+ꢀF], 164 [M+ꢀI], 145 [M+ꢀIꢀF]; elemental analysis calcd (%) for
C6H2NF4I: C 25.0, H 0.7, N 4.8; found: C 24.96, H 0.78, N 4.75.
[9] J. C. Hayes, R. S. Sheridan, J. Am. Chem. Soc. 1990, 112, 5879–5881.
[10] E. Wasserman, Prog. Phys. Org. Chem. 1971, 8, 319–336.
[11] a) M. Iwasaki, K. Toriyama, H. Muto, J. Chem. Phys. 1979, 71,
2853–2859; b) U. S. Rai, M. C. R. Symons, J. L. Wyatt, W. R.
Bowman, J. Chem. Soc. Faraday Trans. 1993, 89, 1199–1201.
[12] a) N. M. Atherton, Principles of Electron Spin Resonance, Prentice
Hall, New York, 1993; b) P. Fleischhauer, S. Gehring, C. Saal, W.
Haase, Z. Tomkowicz, C. Zanchini, D. Gatteschi, D. Davidov, A. L.
Barra, J. Magn. Magn. Mater. 1996, 159, 166–174.
[13] a) Y. Teki, T. Takui, H. Yagi, K. Itoh, H. Iwamura, J. Chem. Phys.
1985, 83, 539–547; b) M. Matsushita, T. Momose, T. Shida, Y. Teki,
T. Takui, K. Itoh, J. Am. Chem. Soc. 1990, 112, 4700–4702.
[14] I. R. Dunkin, P. C. P. Thomson, J. Chem. Soc. Chem. Commun. 1982,
1192–1193.
[15] a) E. Leyva, D. Munoz, M. S. Platz, J. Org. Chem. 1989, 54, 5938–
5945; b) A. Marcinek, M. S. Platz, J. Phys. Chem. 1993, 97, 12674–
12677; c) K. A. Schnapp, R. Poe, E. Leyva, N. Soundararajan, M. S.
Platz, Bioconjugate Chem. 1993, 4, 172–177; d) B. A. Smith, C. J.
Cramer, J. Am. Chem. Soc. 1996, 118, 5490–5491; e) A. Admasu,
A. D. Gudmundsdottir, M. S. Platz, J. Phys. Chem. A 1997, 101,
3832–3840.
[16] D. Grote, W. Sander, J. Org. Chem. 2009, 74, 7370–7382.
[17] J. Pipek, P. G. Mezey, J. Chem. Phys. 1989, 90, 4916–4926.
[18] M. Wanko, M. Hoffmann, P. Strodet, A. Koslowski, W. Thiel, F.
Neese, T. Frauenheim, M. Elstner, J. Phys. Chem. B 2005, 109,
3606–3615.
2-Iodo-3,4,5,6-tetrafluorophenylazide (7d):
A
solution of (1.93 g,
6.539 mmol) 2-iodo-3,4,5,6-tetrafluoroaniline in CF3COOH (50 mL) was
cooled with ice/NaCl to ꢀ58C. Over a period of 15 min an aqueous solu-
tion of NaNO2 was added slowly to the cooled solution, whereby the
temperature did not exceed 58C. After stirring the mixture for 30 min at
about ꢀ58C, a solution of NaN3 (0.55 g, 8.5 mmol) was added dropwise
and the mixture was stirred for a further 30 min. For a short time the so-
lution was heated to 358C. After cooling the reaction mixture down to
room temperature, 250 g of ice was added, and the red aqueous solution
was extracted twice with MTBE. The resulting organic phase was extract-
ed twice with 250 mL of a saturated solution of NaHCO3 and dried over
MgSO4. The solvent was removed under vacuum and the red product
was purified by column chromatography (hexane) to yield 1.9 g (91.6%).
Rf =0.50 (hexane); 19F NMR (400 MHz, CDCl3, 258C, TMS): d=ꢀ114
(F6), ꢀ147 (F5), ꢀ152 (F)4, ꢀ157 ppm (F3); 13C NMR (400 MHz, CDCl3,
258C, TMS): d=136.3 (C2), 139.3 (C6), 140.1 (C4), 141.2 (C5), 142.7
(C3), 143.7 ppm (C1); MS (EI): m/z (%): 317 [M+], 291 [M+ꢀN3], 201
[M+ꢀN3ꢀ4F], 162 [M+ꢀIꢀN2], 127 [M+ꢀN3ꢀIꢀF]; elemental analysis
calcd (%) for C6N3F4I: C 22.74, H 0.0, N 13.25; found: C 22.78, H 0.01, N
13.26.
[19] F. Neese, J. Am. Chem. Soc. 2006, 128, 10213–10222.
[20] W. Sander, D. Grote, S. Kossmann, F. Neese, J. Am. Chem. Soc.
2008, 130, 4396–4403.
[21] J. L. Holmes, P. M. Mayer, J. Phys. Chem. 1995, 99, 1366–1370.
[22] P. M. Mayer, M. S. Taylor, M. W. Wong, L. Radom, J. Phys. Chem. A
1998, 102, 7074–7080.
[23] M. Griffin, A. Muys, C. Noble, D. Wang, C. Eldershaw, K. E. Gates,
K. Burrage, G. R. Hanson, Mol. Phys. Rep. 1999, 26, 60–84.
[24] S. Stoll, A. Schweiger, J. Magn. Reson. 2006, 178, 42–55.
[25] Gaussian 03, Revision C.02, M. J. Frisch, G. W. Trucks, H. B. Schle-
gel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, J. A. Montgom-
ery, Jr., T. Vreven, K. N. Kudin, J. C. Burant, J. M. Millam, S. S.
Iyengar, J. Tomasi, V. Barone, B. Mennucci, M. Cossi, G. Scalmani,
N. Rega, G. A. Petersson, H. Nakatsuji, M. Hada, M. Ehara, K.
Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda,
O. Kitao, H. Nakai, M. Klene, X. Li, J. E. Knox, H. P. Hratchian,
J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E.
Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W.
Ochterski, P. Y. Ayala, K. Morokuma, G. A. Voth, P. Salvador, J. J.
Dannenberg, V. G. Zakrzewski, S. Dapprich, A. D. Daniels, M. C.
Strain, O. Farkas, D. K. Malick, A. D. Rabuck, K. Raghavachari,
J. B. Foresman, J. V. Ortiz, Q. Cui, A. G. Baboul, S. Clifford, J. Cio-
slowski, B. B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaro-
mi, R. L. Martin, D. J. Fox, T. Keith, M. A. Al-Laham, C. Y. Peng,
A. Nanayakkara, M. Challacombe, P. M. W. Gill, B. Johnson, W.
Chen, M. W. Wong, C. Gonzalez, J. A. Pople, Gaussian, Inc., Wall-
ingford, CT, 2004.
Acknowledgements
This work was financially supported by the Deutsche Forschungsgemein-
schaft and the Fonds der Chemischen Industrie. F.N. and S.K. gratefully
acknowledge the SFB 624 (“Template—vom Design chemischer Schablo-
nen zur Reaktionssteuerung”, Universitꢁt Bonn) as well as the SFB 813
(“Chemistry at Spin-Centers”, Universitꢁt Bonn) for financial support.
[1] H. F. Bettinger, W. Sander, J. Am. Chem. Soc. 2003, 125, 9726–9733.
[2] J. Morawietz, W. Sander, J. Org. Chem. 1996, 61, 4351–4354.
[3] a) H. H. Wenk, W. Sander, Angew. Chem. 2002, 114, 2873–2876;
Angew. Chem. Int. Ed. 2002, 41, 2742–2745; b) W. Sander, M. Win-
Chem. Eur. J. 2010, 16, 4496 – 4506
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
4505

W. Sander et al.
[26] a) ORCA—An ab initio, Density Functional and Semiempirical Pro-
gram Package, F. Neese, University of Bonn, Bonn, 2008;
b) ORCA—An ab initio, Density Functional and Semiempirical
Program Package, F. Neese, U. Becker, D. Ganyushin, S. Koßmann,
A. Hansen, D. Liakos, T. Petrenko, C. Riplinger, F. Wennmohrs,
University of Bonn, Bonn, 2009.
[27] a) A. D. Becke, Phys. Rev. A 1988, 38, 3098; b) J. P. Perdew, Phys.
Rev. B 1986, 33, 8822.
[28] a) O. Vahtras, J. Almlçf, M. W. Feyereisen, Chem. Phys. Lett. 1993,
213, 514–518; b) F. Neese, J. Comput. Chem. 2003, 24, 1740–1747.
[29] A. Schꢁfer, C. Huber, R. Ahlrichs, J. Chem. Phys. 1994, 100, 5829–
5835.
[30] a) A. D. Becke, J. Chem. Phys. 1993, 98, 1372–1377; b) C. Lee, W.
Yang, R. G. Parr, Phys. Rev. B 1988, 37, 785.
[31] S. Sinnecker, F. Neese, J. Phys. Chem. A 2006, 110, 12267–12275.
[32] a) F. Neese, J. Chem. Phys. 2001, 115, 11080–11096; b) F. Neese, J.
Chem. Phys. 2005, 122, 034107/034101–13.
[33] B. A. Hess, C. M. Marian, U. Wahlgren, O. Gropen, Chem. Phys.
Lett. 1996, 251, 365.
[34] F. Neese, J. Chem. Phys. 2003, 119, 9428–9443.
[35] Y. Tobe, S.-i. Sasaki, M. Mizuno, K. Hirose, K. Naemura, J. Org.
Chem. 1998, 63, 7481–7489.
Received: November 30, 2009
Published online: March 15, 2010
4506
www.chemeurj.org
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 4496 – 4506